article review

388 Infantile Autism Reviewed: A Decade of Research

by Marian K. DeMyer, Joseph N. Hingtgen, and Roger K. Jackson

Abstract

Progress in the study of infantile autism over the past 10 years was reviewed in terms of classification and diagnosis, relation to other disorders, demographics, parental characteristics, genetics, intelli- gence, language, perceptual proc- esses, behavioral characteristics, neurobiological, biochemical and pharmacological aspects, behav- ioral/educational treatment methodologies, prognosis, and theoretical considerations. This decade's research led to the now generally accepted position that infantile autism is a type of devel- opmental disorder accompanied by severe and, to a large extent, permanent intellectual/behavioral deficits. However, its relationship to other forms of developmental disorders and to mental retarda- tion remains to be delineated. Perceptual/cognitive/language de- fects appear central to the autistic syndrome, but the specific under- lying mechanisms are unknown. Most studies indicate that autistic children have more signs of brain dysfunction than do normal chil- dren and about the same number as mentally retarded children. The overwhelming evidence suggests that the treatment of choice for maximal benefit to autistic chil- dren is a systematic, intrusive behavioral/educational approach. Yet, in spite of significant gains in almost all children treated, the typical prognostic picture is poor in terms of achieving self-sup- portive adulthood. The parents of autistic children have been found to be essentially similar to parents of children with organic brain dis- orders, and manifest no psycho- pathology which conceivably could induce the disorder. The

vast majority of theoretical arti- cles appearing in the 1970s pro- posed some form of neurobio- logical defect as the causative factor in autism. One of the major goals for future research is to un- dertake more extensive compara- tive studies of nonautistic brain dysfunctional children and autis- tic children that could yield clear- er differential behavioral profiles and testable neurobiological hypotheses.

It has been 9 years since a general survey of the most important de- velopments in the study of early childhood psychosis was pub- lished in Schizophrenia Bulletin (Hingtgen and Bryson 1972). That review covered the period from about 1964 to 1970 and was based on a perusal of over 400 articles pertinent to infantile autism, childhood schizophrenia, and re- lated disorders. Since over 1,100 additional publications appeared in the decade of the 1970s, includ- ing a new journal devoted to in- fantile autism, another review now seems appropriate to assess what

This review article is dedicated to the memory of two recently deceased pioneers in the study of infantile au- tism: Leo Kanner, the famed child psychiatrist, who in 1943 was the first to describe and name infantile autism as a specific syndrome; and Charles B. Ferster, the equally prominent re- search psychologist, who in 1961 col- laborated on a report that sub- sequently led to the current predominance of behavioral tech- niques in the treatment of autistic children.

Reprint requests to Dr. M.K. DeMyer at Institute of Psychiatric Re- search, Indiana University Medical Center, 1100 W. Michigan St., Indianapolis, IN 46223.

VOL. 7, NO. 3, 1981 389

progress has been made in under- standing the syndrome.

One of the major problems con- fronting the field has been the in- consistent use of terminology referring to the various disorders included under the general desig- nation of early childhood psycho- sis—for example, infantile autism, childhood schizophrenia, early childhood schizophrenia, early in- fantile psychosis, and symbiotic psychosis. Indeed, the term early childhood psychosis is no longer considered useful by most re- searchers and clinicians, since it may mislead one into assuming an association between this condition and adult psychotic disorders, a possibility that appears more and more remote. Reflecting this atti- t u d e is the n e w DSM-III: Diagnos- tic and Statistical Manual of Mental Disorders (American Psychiatric Association 1980), which uses the term pervasive developmental dis- order and includes infantile autism as one category under this general classification. Schizophrenia in childhood, on the other hand, in DSM-III is classified under the same subcategories used for the adult schizophrenias.

In our original review we did not want to exclude any articles pertinent to infantile autism. Thus, we dealt with studies of "psychotic" children manifesting most of the symptoms associated with infantile autism regardless of the terminology used in the diag- nosis. Now, 10 years later, this may be less of a problem since there are so few articles using the terms childhood psychosis or childhood schizophrenia. Indica- tive of this change in attitude is the change in name of the Journal of Autism and Childhood Schizo- phrenia to Journal of\Autism and De-

velopmental Disorders. Therefore, as much as possible, we will restrict the articles reviewed to those clearly dealing with infantile au- tism as described under pervasive developmental disorders in DSM- III. With a few exceptions, those - studies that are more obviously re- lated to what is now regarded as schizophrenia in children will not be considered here.

In our earlier review, we were able to cite almost all of the more than 400 articles reviewed. With over 1,100 papers published dur- ing the 1970s, we were required to be more selective in citing articles for the present review. (For read- ers who are interested, a total list of all references reviewed, includ- ing those not cited in this article, is available on request from the au- thors.) We have referred mainly to studies in which data were collect- ed with some consideration for ex- perimental control or were of spe- cial significance to the field. Case reports have been cited only when especially informative.

Classification and Diagnosis of Infantile Autism

How investigators classify and di- agnose infantile autism has de- pended heavily on the ideas of each investigator concerning etiol- ogy and the underlying meaning of various symptoms. There are nearly as many different systems as there are investigators. Never- theless, some important steps have been taken, and important knowledge has been gained. Be- fore giving these advances, it may be helpful to discuss the historical origin of controversies and issues in diagnosis and classification of infantile autism.

One of the chief problems has been how to handle the questions of mental retardation and organic brain disease, issues especially troublesome with regard to infan- tile autism. When Kanner (1943) first described the diagnostic fea- tures, he also remarked that the condition bore no resemblance to any known neurological condition and implied that autistic children had a basically normal intelli- gence. For over two decades after- wards, diagnosticians generally believed that the presence of men- tal retardation or neurologic signs ruled out the diagnosis of infantile autism in the Kannerian sense, even if the child met all behavioral criteria (Eisenberg 1966). Thus di- agnosis was frequently one dimen- sional; a child was labeled as af- flicted with either infantile autism or mental retardation, not both. On the other hand, Goldfarb (1961) early recognized two sub- classes of childhood schizophre- nia, the organic and nonorganic, a two category system.

Several problems were met in using these ideas. First, organic brain dysfunction is difficult to rule out, especially if the neuro- logical evaluation is not rigorous. Second, even if a patient has no positive neurological signs by ex- amination or by history, the brain may still be malfunctioning in an organic sense. An insult to the brain or a maldeveloped brain may leave certain areas such as the motor system intact and yet seri- ously compromise other areas such as the limbic lobe or language cen- ters without leaving a detectable trace in even a rigorous neurolog- ical evaluation. Insults to the brain may occur in a "silent" fashion without an overt gross neurologi- cal change at the time the insult

390 SCHIZOPHRENIA BULLETIN

occurred. Third, whether a child shows signs of organic dysfunc- tion may change over time. Followup studies have revealed that many preschool autistic chil- dren may have negative neurolog- ical examinations only to develop signs of brain malfunction later in life. For example, Rutter and Lockyer (1967) reported that 18 percent of their followup sample had developed seizure disorders by adolescence. Fourth, long-term studies of intelligence showed that most autistic children whose men- tal retardation was considered functional during their preschool years remained retarded in later years, often despite improvement in social functioning (Knobloch and Pasamanick 1975).

The fact that most studies of in- fantile psychoses have taken place in psychiatric/psychological clinics/hospitals has affected classi- fication and diagnosis. Children with psychotic symptoms have generally been prescreened, and those with obvious neurological signs tend not to be referred to such settings. As Knobloch and Pasamanick (1975) pointed out:

elimination of all patients with other specifically classified dis- orders might leave an apparent- ly discrete condition. Such a se- lection process could explain why debate continues about cri- teria for diagnosis and organic versus psycnogenic etiolo- gy [p. 183]

These two investigators worked in a children's hospital where they saw a wide spectrum of cases, most of whom were 2 years old or younger. Among 1,900 cases, 64 (3.4 percent) met Kanner's criteria, 14 of whom were phenylketonuric. These latter were removed from further comparison and 50 autis-

tics were compared with 50 chil- dren with some type of central nervous system (CNS) dysfunction but without autistic symptoms and 50 without neuropsychiatric disor- der. In both the autistic and devel- opmental disorder samples, there were equal signs of perinatal and neonatal complications and of other disorders implying neuro- logical dysfunction (e.g., develop- mental quotient below 50, convul- sive disorders, strabismus, and cerebral palsy). All 50 autistic chil- dren showed signs of CNS dys- function and differed from the nonpsychotic comparison group only in their autistic symptom complex. In a followup of 39 autis- tics whose mean age was 7 years, three-fourths had lost their autistic behavior but retained their mental retardation.

These kinds of findings led to changes in ideas of investigators who found it increasingly difficult to sort cases of autism that were "true" Kannerian cases from the ones that were "secondary" to brain dysfunction or mental retar- dation. They were beginning to suspect that more cases than here- tofore supposed were complicated by neurological dysfunction or by mental retardation or both. Classi- fication systems in the 1960s, and sometimes yet today, attempted to deal with these complex cases by providing such labels as "second- ary autism" or "organic brain damage with autistic features" or some roughly equivalent terminol- ogy-

Another attempt to deal with complex cases was the multicate- gory system as in DSM-II (Ameri- can Psychiatric Association 1968). For example, a schizophrenic child with an IQ of, say, 54 would have two diagnoses coded, namely

schizophrenia, childhood type, and mental retardation, mild. Still a third attempt was the multi-axial system which differs chiefly from the multicategory system in requiring that certain dimensions of the illness always be coded. In the multicategory system, the coding of several diagnoses is in a sense discretionary.

The pioneer multi-axial system of the World Health Organization (WHO)1 was found superior to a multicategory one by 24 child psy- chiatrists (Rutter, Shatter, and Shepherd 1975). Separation of eti- ological factors made classification easier and more uniform especially considering continuing disputes over theory and causation. If the system is used conscientiously, in- vestigators can protect themselves from bias and keep important di- mensions separate from each other to study their interrelationships.

The most important develop- ment of the 1970s for the immedi- ate future of classification and di- agnosis is the new DSM-HI (American Psychiatric Association 1980), which is multi-axial throughout, listing five separate axes (see table 1). While this was modeled on the WHO system, the intellectual diagnosis of mental re- tardation was eliminated as a sep- arate axis. Instead, coding of both a psychiatric disorder and mental retardation is done on Axis I (see table 2). In the only reliability study to date (Russell et al. 1979),

1 The WHO system had four axes: (1) clinical psychiatric syndrome; (2) intellectual level; (3) associated bio- logical factors; and (4) psychological/ social factors. DeMyer (1975b) also has used a three-dimensional system (behavioral, intellectual, and neuro- logical) since the early 1970s.

VOL. 7, NO. 3, 1981 391

Table 1. DSM-III axes of psychiatric diagnosis

Axis

Axis

I.

Axis Axis Axis

III IV V

Clinical psychiatric syndrome(s) and other conditions (multiple diagnoses are possible) Personality disorders (adults) Specific developmental disorders (children and adolescents) Physical disorders Severity of psychosocial stressors Highest level of adaptive functioning past year

diagnosticians had low rates of agreement when two diagnoses were needed in a single category. To us, it appears more appropriate to require an intellectual diagnosis as a separate axis for all children's psychiatric disorders. This coding seems justified, particularly be- cause of the importance of meas- ured IQ as a prognosticator in in- fantile autism2 (see the Prognosis section).

Age of onset has become an im- portant factor. Kolvin (1971) classi- fied child psychoses as early onset (before 30 months), middle onset (after 30 months up to 8 years), and late onset (prepubertal and pubertal). He found that early onset cases (which most would probably call infantile autism) were much more common than middle onset cases, and the former did not appear to him to be related to later developing schizophrenia. Prior et al. (1975) subjected data from RimlancTs Behavior Rating Checklist to a numerical taxonomic analysis and found that age of

2 The National Society for Autistic Children (NSAC) (Ritvo and Freeman 1978) has recommended a four-point diagnostic system. It will probably be superseded by DSM-III.

onset was one of the primary fea- tures that distinguished the two resulting classes of children (class 1: onset before 2 years of age, au- tistic type behavior and severely impaired interpersonal relation- ships; class 2: later onset, less se- vere relationship impairment, and greater symptom variability). The age break demonstrated by Prior et al. (1975) was 6 months younger than Kolvin's (1971).

DSM-III uses age of onset (i.e., 30 months) as the dividing line for diagnosing infantile autism versus childhood onset pervasive devel- opmental disorder (see table 2). The new manual, however, does not make use of Kolvin's (1971) findings concerning middle onset psychosis, but instead gives 12 years as the upper chronological age for diagnosing pervasive de- velopmental disorder. It appears that relationships of exact age of onset to other important variables such as symptom picture and out- come have not yet been firmly es- tablished (Chess 1977). Chess (1977), in her study of autism in congenital rubella, found that typical autism could develop after age 2V2 years. In any case, locating age of onset within an exact 6-month period is difficult to do, but no formal reliability studies on the subject have been conducted.

Names of Childhood Psychoses and Diagnostic Criteria. Perhaps no condition in the past has carried more different terms or more different names for subcate- gories than early childhood psy- chosis. Certainly many frequently used terms meant various things to various people. There was no consensus during the 1960s among investigators about features that distinguished between "infantile autism" and "childhood schizo- phrenia," two of the then most commonly used terms. Hingtgen and Bryson (1972) tabulated 28 separate behaviors mentioned by 13 groups of authors as pathogno- monic of infantile autism, child- hood schizophrenia, or childhood psychoses. The same problems ex- ist today for the reviewer in trying to determine whether an author of the 1970s was talking about a dif- ferent group of children when using the diagnostic label "infan- tile autism" than another author who used the now infrequent label "childhood schizophrenia." For this reason, it is difficult to com- pare the subcategories of early childhood psychosis. The use of these multiple terms—considering that no clear difference among them has been demonstrated— causes confusion, not only in literature reviews and among pro- fessionals but also in parents try- ing to understand the import of their child's diagnosis (DeMyer 1979).

Over the last decade some in- vestigators have changed their terms. Rimland (1971) in the E-l version of his checklist spoke of a "schizophrenia" score which he called in the later two versions a "nonautistic" score. DeMyer et al. (1971a) first called a higher func- tioning group of preschool psy-

II.

392 SCHIZOPHRENIA BULLETIN

chotic children "early schizophre- nia" because they thought it did not meet Kanner's criteria to the letter. Because these children had

developed serious symptoms in the first 2 years of life and seemed to fit what most other diagnos- ticians called "infantile autism,"

Table 2. DSM-III classification of disorders previously called by other terms (e.g., early child psychosis, early infantile autism, childhood schizophrenia, atypical child psychosis)

Axis I1

Generic term: Pervasive developmental disorders2

Subclass term

1. Infantile autism

Full syndrome Residual

299.00 299.01

2. Childhood onset perva- sive developmental disorder

Full syndrome Residual

299.90 299.91

3. Atypical pervasive de- velopmental disorder

299.8x

Diagnostic criteria

All of the following: a. Onset before age 30 months b. Serious lack of social response c. Language deficit—gross d. Speech peculiar, if present e. No delusions/hallucinations a. Severe disturbance of emotional rela-

tions b. Three or more of: excess anxiety; affect

disturbance; resist change; peculiar motility; dysprody; abnormal sensation; self-mutilation; onset between ages 21/2 and 12 years; and no delusions/ hallucinations

Distortion of social and language skills that cannot be classified as either 1 or 2 above

Generic term: Schizophrenic disorders3

Subclass term Diagnostic criteria

Schizophrenic disorders in childhood (5 subtypes as for adults) Full syndrome 295.1x through 295.6x Residual 295.6x

As for adult schizophrenia, all of the following: a. At least 1 of 10 delusional, hallucinatory,

associational, or affectional symptoms b. Impairment in two or more routine func-

tions c. Continuous illness for at least 6 months

11f an individual has mental retardation in addition to a mental illness, this diagnosis is also coded on Axis I and is described in DSM-III beginning on p. 36. Borderline mental retardation is described on p. 332. 2Described in DSM-III, pp. 8 6 - 9 2 , under the section "Disorders Usually First Evident in Infancy, Childhood or Adolescence." 3Described in DSM-III, pp. 181-193, in the section "Schizophrenic Disorders."

the authors dropped the term "schizophrenia" and called this group "high functioning infantile autism."

Piggott and Simson (1975) pre- sented some evidence that diagno- sis itself may have changed in the 1970s. Using DeMyer and Chur- chill's (DeMyer et al. 1971a) crite- ria and their chart information, they rediagnosed 33 children who had been originally diagnosed in the 1950s and 1960s. Childhood schizophrenia was the original di- agnosis in 58 percent of cases but constituted only 18 percent on rediagnosis. One third of children originally diagnosed as psychotic did not retain this category on rediagnosis.

We hope that reviewers at the end of the 1980s will be able to re- port separately on the subcate- gories of major mental illness of childhood if investigators have fol- lowed the new DSM-III diagnostic system. We will have to get used to new terms. "Childhood psycho- ses" does not appear at all. Instead of the old generic term "childhood schizophrenia," DSM-III gives us "pervasive developmental disor- ders" (see table 2). In fact, child- hood schizophrenia is nowhere mentioned, but we are advised to speak of "schizophrenia in child- hood." DSM-III also resurrected an old term that has appeared sel- dom in the literature of the 1970s—namely, "atypical"—and applied it to pervasive develop- mental disorder whose relation- ship to "schizophrenia in child- hood," the authors say, is disputed. But what is its relation to infantile autism, with which it shares a generic heading? (see ta- ble 2). The answer awaits future research.

VOL. 7, NO. 3, 1981 393

Studies of Diagnostic Reliability and Differential Diagnosis. In ad- dition to the dimensions of intelli- gence and neurological signs, the behavioral features of infantile au- tism create many thorny reliability problems. Kanner's behavioral cri- teria for infantile autism—namely, severe social distance, desire for preservation of sameness, and fail- ure to use language for communi- cation—have been used frequent- ly. This situation would seem to make for good diagnostic agree- ment, especially since the criteria seem to be relatively simple. Actu- ally they are complex constructs , that can be and are defined by many different behaviors. For ex- ample, various investigators differ in what and how many social avoidance behaviors need to be present to define autistic social distance. All of the rating scales, while having overlapping items, vary considerably in the number and types of items. For example, Rimland's (1971) scale has 80 items while Clancy, Dugdale, and Rendle-Short (1969) used 14. The categorical systems also vary in the features that must be present. The requisites for a useful diagnostic system are that different diagnos- ticians using it make the same di- agnosis, that it differentiates relat- ed but separate conditions (see table 2 for differential diagnosis), and that it be related to other im- portant aspects of the illness such as cause or type of therapy re- quired. We reviewed the literature to see how well these conditions had been met by any of the diag- nostic systems.

Rimland's scale, which has undergone three revisions (E-l, E-2, E-3), has received the most research attention. DeMyer et al. (1971a) found that it differentiated

only partially between infantile autism and mental subnormality; that his E-l version more closely coincided with their definition of primary autism than did his later E-2 version, but that only a small fraction of children diagnosed au- tistic by their criteria met Rim- land's +20 cutoff score. They also compared Rimland's system with those of Polan and Spencer (1959), Lotter (1966), and the British Working Party (Creak 1964). While there were significant correlations among scores of all these scales, there was an overlap no greater than 35 percent, meaning that sev- eral children receiving high scores on some scales would receive low scores on others. None of the sys- tems differentiated the subcategories of psychosis as de- fined by DeMyer and Churchill (DeMyer et al. 1971a) who achieved 96 percent agreement with each other. The authors con- cluded that investigators who work closely together can achieve good diagnostic agreement about subgroups of infantile psychoses. However, agreement lessens con- siderably when diagnosticians without constant feedback from one another compare diagnoses. Davids (1975) reported that Rimland's E-l checklist differenti- ated chart diagnoses of autistic, other psychotic, and nonpsychotic groups. However, the checklist did not differentiate children with the full symptom complex of in- fantile autism from those ap- pearing "autistic-like."

Questions have been raised about the reliability of Rimland's (1971) checklist because it relies solely on the parents scoring the items by themselves. Albert and Davis (1971) found high agreement between 31 sets of parents who

rated their normal preschool chil- dren. In the usual rating situation of Rimland's scale, however, par- ents are asked to rate how the child was behaviorally in the first 5 years of life. If the child is, say, 8 years old, parents may remember differently than they would have when the child was 5 years old. Thus children of different ages might be rated differently because of differences in parents' memory factor. Interestingly, the normal children received mean autism scores of 21.2 compared to 20.9 for autistic children as reported by Douglas and Sanders (1968) and 23.6 reported by DeMyer et al. (1971a).

When parent ratings were com- pared with teacher ratings by Prior and Bence (1975), there were a total of 63 disagreements for nine autistic children. Generally the teacher assessed the children at a lower level than the parents in es- timates of ability, social respon- siveness, and bizarre behavior. In only one case, however, would the diagnosis have been changed. The authors criticized the weight given to speech items. Children with higher E-2 scores had speech pat- terns that added considerably to their scores so that they were not strictly comparable with the scores of the nonspeakers for severity of handicap. Masters and Miller (1970) have pointed out also that the instrument does not signifi- cantly improve upon base rate pre- diction.

Rimland (1971) himself subtracts the "nonautistic score" from the "autistic score" and calls only those with +20 scores "autistic," 10 to 20 "equivocal," and below 10 "definitely not autistic." Of 2,218 suspected autistics whose E-2 forms he analyzed, only 9.7 per-

394 SCHIZOPHRENIA BULLETIN

cent met his cutoff +20 score. However, because these children were diagnosed by many different clinicians who used no uniform criteria, no effective diagnostic comparison groups were available to him.

Rimland's checklist was com- pared with Ruttenberg's Behavior Rating Instrument for Autistic and Atypical Children (BRIAAC) by Cohen et al. (1978) and with clini- cal diagnoses using DSM-III crite- ria. On clinical grounds 13 chil- dren received a diagnosis of "primary autism" and the re- maining 14 other diagnoses. All were rank ordered by degree of se- verity. Only 2 of 13 children with a clinical diagnosis of primary au- tism met Rimland's +20 score. Clinical assessment of severity re- vealed "relatively good agree- ment" on overall severity as meas- ured by BRIAAC. The latter test, however, did not differentiate among clinical diagnostic groups. Thus Cohen et al. (1978), like other investigators, found that a large proportion of the children with equivocal and nonautistic Rimland scores appeared clinically quite similar to the children with scores above +20. In contrast to Rim- land's system, which may result in many false negatives, the 14-point system of Clancy, Dugdale, and Rendle-Short (1969) may locate many false positives (Capute et al. 1975).

An important source of diagnos- tic unreliability comes from the re- lation of various symptoms to both mental age and chronological age. Freeman et al. (1978) compared au- tistic, normal, and retarded chil- dren, ages 2 to 5V2 years, using 67 "objectively defined behaviors" under controlled observer condi- tions. An important strategy was

that the two abnormal groups were matched for mental age and chronological age. A discriminant analysis revealed that only two be- haviors ("repeats sounds" and "communicates speech") defined the discriminant function and yielded only 63 percent overall correct classification of the three groups. While 78 percent of the normal children were correctly classified and only 12 percent of autistic and retarded children were misclassified as normal, it was not possible to separate the autistic from the retarded in the presence of the normal sample. There was some evidence that many of the items used on the rating scale were strongly related to both chronolog- ical and mental age. Such a situa- tion would mean, for example, that some behavior held to be es- sential to the diagnosis of autism by one investigator might be strongly present at one chronolog- ical or mental age, but not at an- other and lead to a difference in diagnosis in the same child at dif- ferent chronological ages. The au- thors concluded that several years of work were necessary to develop a rating scale which could distin- guish reliably between young au- tistic children and other diagnostic groups.

The continuous nature of autis- tic symptoms across various diag- nostic groups such as blindness, deafness, mental retardation, learning disorders (Wing 1969), disintegrative psychosis (Corbett et al. 1977), and even normality has posed many problems in achieving reliability in the differ- ential diagnosis. There are very few symptoms, for example, that are found nearly exclusively in in- fantile autism. Even though the various types of childhood psy-

chosis are supposed to be defined by a group of symptoms in each child so diagnosed, factor-analytic, cluster-analytic, and numerical taxonomic analyses, by and large, have not verified this assumption, no matter whose diagnostic system has been studied. In the largest study (DeMyer, Bryson, and Chur- chill 1973) examining these issues (normals, n=30; nonautistic neuro- logical cases, n=36; autistics, n = 70), the following findings emerged: (1) The psychiatric diag- nostic groups lay on a severity continuum, with nonautistics hav- ing the fewest symptoms, high- functioning autistics next fewest, and lower functioning autistics the most. (2) The only symptom items differentiating the three psychiat- ric diagnostic groups were degree of social withdrawal and degree of communicativeness of speech. (3) A cluster analysis showed much symptom overlap, thus precluding any definition of cluster groups. However, this analysis did point out unequivocally the continuous nature of symptom distribution across normal, neurological, and autistic groups and confirmed the severity continuum of these groups. The latter feature was pointed out by Anthony in 1958.

Capute et al. (1975) have dem- onstrated how frequently autistic symptoms are found in develop- mentally disabled children by using the 14-point autistic checklist of Clancy, Dugdale, and Rendle-Short (1969). They found that as the severity of mental re- tardation increased, so did the fre- quency of autism as diagnosed by Clancy's scale. While the authors did not compare symptom fre- quency with mental age, their findings strongly suggested a pos- itive correlation and would tend to

VOL. 7, NO. 3, 1981 395

support the findings of Freeman et al. (1978).

Clearly a crucial question still re- mains about which behavioral functions critically define the dif- ferences between infantile autism and related conditions. Research of the 1970s strongly suggests that degree of social distance and de- gree of communication of speech are the most consistent differenti- ating rating items. However, the optimum rating instrument has yet to be devised. Any rating instru- ment that cannot differentiate in 100 percent of cases the normal child from the autistic must be used in conjunction with other aids. It seems clear also that re- searchers in rating behavioral symptoms must find ways to con- trol for both mental age and chron- ological age.

Rutter (1978) has suggested that the single most important factor for the diagnosis of infantile au- tism is that the social age be on a lower plane than the mental age, but much research must be done on such an assertion. For example, the mental age can be derived using several dimensions of intel- ligence. Which dimensions should be used in the comparison: verbal age, perceptual-motor age, gross motor age, or a general intelli- gence estimate? If the last dimen- sion is chosen, what test or group of tests should be used? This is a crucial issue because of the wide scatter of many psychotic children on psychological tests. Another important question is how to measure social age, which in itself can be defined by many different behaviors. And thus we return full circle to the ever present problems in achieving diagnostic reliability.

The most recent classification system has been introduced by

Schopler et al. (1980), who used their Childhood Autism Rating Scale (CARS) to compare their def- inition of autism with that of Ritvo and Freeman (1978) and of Rutter (1978). While there is agreement among all three systems about some behavioral factors that are important for diagnosis, there are important differences. Schopler et al., with 15 rating items, use the broadest concept of autism, and they include the specific items about which Ritvo and Freeman and Rutter disagree: "sensory pe- culiarities" (Ritvo and Freeman) and insistence on sameness and stereotypies (Rutter). Of the 266 children identified as autistic by the CARS, only 20.6 percent met Rutter's criteria for autism, while 46.6 percent met Ritvo and Free- man's criteria—a significant differ- ence. On this basis, Schopler et al. argue that sensory peculiarities should be included as a primary diagnostic feature of autism.

Validity of Diagnoses. The ulti- mate test of usefulness of a diag- nostic system is whether it is relat- ed to some other important features of the condition such as cause, outcome, or treatment. Un- fortunately, none of the present systems can completely achieve this goal. No matter what system is used, a diagnosis, for example, of infantile autism, early child- hood schizophrenia, or early child- hood psychosis predicts a less than normal outcome in general (DeMyer 1979). With regard to eti- ological and therapeutic consider- ations, none of the checklists or categorical systems which have been proposed in the last decade appear unequivocally superior. Thus, continued work toward devising a more useful diagnostic

system is required. To us, it ap- pears that multi-axial systems have the greatest potential, espe- cially if the influences of mental and chronological age on behavi- oral symptoms can be better delin- eated.

Relation to Other Disorders

The major conditions that infantile autism must be differentiated from according to DSM-III are listed in table 3 along with major differentiating features. As was discussed in the diagnosis and classification section, the various formal checklist diagnostic sys- tems lacked diagnostic specificity and sometimes could not differen- tiate autistic from young normal children let alone from related pathologic groups. It is clear we still need a clinical diagnostic his- tory and observation to use the various proposed rating scales in- telligently. To recapitulate: The two features that most consistently separated nonautistic from autistic populations were degree of com- municative skills and social relat- edness. These differentiating fea- tures applied both to the mental retardation and expressive lan- guage disorder populations (see table 3). (For relationship to dys- phasia, see the Language section.)

Mental Retardation. Bartak and Rutter (1976) found some symp- tom differences between autistics above and below 70 IQ (non- verbal). The more intellectually able autistics had better relation- ships with adults, higher language skills, better developmental mile- stones, and fewer ritualistic and compulsive behaviors. Using the BRIAAC, Wolf, Wenar, and

396 SCHIZOPHRENIA BULLETIN

Table 3. DSM-III differential diagnosis of pervasive developmental disorders: Major differential features listed in DSM-III

Axis I

Infantile autism vs. childhood onset pervasive developmental disorder Former onset before 30 months of age and latter after 30 months of

age—chief differential criterion

Pervasive developmental disorders vs: 1. Mental retardation

a. Full syndrome PDD rarely present b. When both disorders present, both diagnoses should be made

2. Schizophrenia in childhood a. Hallucinations, delusions, incoherence present

3. Hearing impairment a. History of responding only to loud noises b. Audiogram indicates deafness or hearing impairment

4. Developmental language disorder, receptive type a. Eye and social contact made b. Appropriate gestures used

Ruttenberg (1972) found that re- tarded children, particularly those with Down's syndrome, were less emotionally disturbed and better integrated than autistic children who were unable to generalize "between key areas of function/7

(See also the Intelligence section.)

Borderline or Schizoid States. Ac- cording to Wolff and Barlow (1979), about 3 to 4 percent of new referrals to child psychiatry could be labeled as "schizoid" or resem- bling Asperger's (Van Krevelen 1971) autistic psychopaths. They were described as verbal, solitary, unadaptable, overly sensitive, and paranoid with "odd ideation." Schizoid children were intermedi- ate in their overall functioning be- tween autistics and normals and did poorly on cognitive and mem- ory tasks. Autistics did relatively well on rote memory and visual- spatial tasks and poorly on lan-

guage comprehension. Thus, there was a continuum effect, as well as an indication that the schizoid children had a somewhat different disability pattern. It is important to note that both experimental groups were not typical of the larger groups from which they were drawn; that is, the experi- mental autistic children were higher functioning than autistic children as a whole and the exper- imental schizoid were lower func- tioning than their larger group. This observation points out strongly that schizoid children are not so severely handicapped intel- lectually as autistic children and further substantiates the continu- um of this handicap across diag- nostic groups.

Later Developing Schizophrenia. Followup studies (Rutter and Lockyer 1967; DeMyer et al. 1973) have emphasized that lower func-

tioning infantile autistic children grown into adulthood generally do not resemble adult schizophrenics but more often appear mentally re- tarded. Higher functioning autistic adults also are not typically diag- nosed as schizophrenic but gener- ally have improved their social symptoms which, while stopping short of normal, are not typically those of an adult schizophrenic. Those studies that have looked into the childhood symptoms of typical adult schizophrenia have failed to find many examples of typical cases of infantile autism or childhood schizophrenia (Roff, Knight, and Wertheim 1976; Watt 1978). Many cases of infantile au- tism as adults have been cared for both in institutions for the mental- ly retarded and mentally ill where, according to Kanner (1973, p. 185), many "settled down to a life not too remote from a Nirvana-like state," presumably meaning a con- tinuation of the early social isola- tion.

However, Bender and Faretra (1973, p. 57), who included infan- tile autism under the rubric of childhood schizophrenia, stated that "child schizophrenia invaria- bly leads to adult schizophrenia al- though it may be expressed in a wide variety of clinical pictures." Also Fish (1977) indicated that adult and child schizophrenics are part of the same spectrum on the basis of common perceptual-motor integration problems. However, Piggott and Gottlieb (1973) found childhood schizophrenics did not resemble adult schizophrenics on several biological measures. Chess (1979) presented evidence that children with congenital rubella who developed autistic symptoms in the postinfancy period looked the same as those becoming autis-

VOL 7, NO. 3, 1981 397

tic during infancy. To her this was evidence that autism is a separate syndrome. Whether infantile au- tism is the earliest expression of schizophrenia needs yet to be verified or refuted through sys- tematic longitudinal study and comparison of adult autistics with adult schizophrenics, a study that has never been done.

After comparing results of all developmental studies on child- hood psychosis, we can state that there is a strong continuum effect in general intellectual ability of the four diagnostic groups, namely: infantile autism is lowest, later de- veloping psychosis (e.g., schizo- phrenia in childhood) intermedi- ate, and borderline states and nonpsychotic learning disability children highest. One possible connection between all the psy- chotic and borderline states in all chronological age groups is the presence of cognitive disability. Here again, there is a possibility of continuum effect, with autistics and child schizophrenics having the most severe degree of cogni- tive defect and schizoid and adult schizophrenia populations having less severe degrees. Trunnell (1965), using a Piagetian frame of reference, found that adult schizo- phrenic patients were like 7- to 11-year-old children in forming logical concepts. But all mentally retarded individuals without psy- chosis are also retarded in this re- spect, so the explanation for psy- chotic symptoms would have to lie at least partially in some other sphere or in a particular and unique cognitive disability.

Demography

Wing et al. (1976) compared all prevalence studies in England and

Wales, the U.S.A., and Denmark. Those studies in which a total population or a wide range of handicapped children were sur- veyed with case-note inspection and interviews all revealed the estimated prevalence of autism to be between 4 and 5 children per 10,000 aged under 15 years. In contrast, those studies employing diagnosed cases only or using ad- ministrative records only produced lower prevalence rates. Ornitz and Ritvo (1976, p. 615) consider that all the published estimates are low because "autistic children younger than 4 and older than 6 are often misdiagnosed." Chess (1977) has shown that the prevalence rate in children with congenital rubella is 741 per 10,000 which is roughly comparable to one estimate of the rate in severely subnormal chil- dren (see Lotter below). In Denmark, Haracopos and Kelstrup (1978) estimated that about 25 per- cent of mentally retarded children displayed "psychotic behavior."

Lotter (1978) examined 1,300 mentally retarded children in Africa and found 2.3 percent had some autistic-like behavior and only 0.6 percent were autistic according to Western standards. While it was not a true prevalence study, the numbers of autistics found were far fewer than ex- pected; e.g., in England, 5 to 8 percent of all severely subnormal children would have some marked autistic behavior. In comparison to Western autistics, rocking, head banging, and complex object use were rare. In sex ratio, occurrence of epilepsy, and social status, the African group "was broadly com- parable" to the Western group.

In New South Wales, 21.9 per- cent of all subjects diagnosed as autistic over a 20-year period had

at least one foreign-born, non- English-speaking parent (Harper and Williams 1976). Greeks and Germans were overrepresented in this sample. Parker (1978) from Australia reported that the season of birth did not differ between autistics and the normal popula- tion.

In a study of autistic, cerebral palsied, mongoloid, and normal children, Ando and Tsuda (1975) from Japan found that autistics did not differ from the normal control group or from general population statistics in maternal age, birth or- der, or age interval between the subjects and their closest sibling. Autistics had a 4.95:1 male/female ratio in contrast to cerebral palsied 1.75, mongoloids 1.18, and nor- mals 1.1. The data on maternal age, birth order, and male/female ratio were consistent with those of most other investigators in the 1970s from the United States and England (Kolvin et al. 1971; Ritvo et al. 1971a; Spence et al. 1973). In contrast, Allen et al. (1971) re- ported that mothers of autistics were significantly older than a normal control group matched for age of child, birth order of child, and position of child in the family. Also O'Moore (1972) found the age of mothers of autistics in Dublin, Ireland, to be higher than the national average (32 years vs. 28).

Previously, most studies agreed that parents of autistics come from higher socioeconomic groups than the general population or from control populations (Treffert 1970; Allen et al. 1971; Kolvin et al. 1971; O'Moore 1972; Cox et al. 1975). Types of indicators of higher soci- oeconomic status (SES) in the au- tistic groups were higher incomes, more professional and managerial jobs, home and car ownership,

398 SCHIZOPHRENIA BULLETIN

and spacious housing. Ritvo et al. (1971fl), however, found similar SES characteristics between autis- tic children's families and those of children with another psychiatric diagnosis. Schopler et al. (1980) reported that in their sample of 264 autistics from North Carolina, 74 percent came from lower SES families.

L. Wing's (1980) epidemiological survey of autistic, other psychotic, and mentally retarded children of southeast London suggests the reason that so many study popula- tions of autistics appear to come from high SES families. She found no SES differences among the three groups studied but did find that the higher SES fathers of au- tistics were more likely than the other fathers to apply for service from clinics with a special interest in autism. Membership in the Na- tional Association for Autistic Children was also related to higher SES.

All measures of SES also have implications for the IQ of the heads of household, with those from higher SES groups having higher IQs than lower SES groups. Findings from the United States and Japan (Allen et al. 1971; Ki- kuchi et al. 1971; Florsheim and Peterfreund 1974) indicated that only fathers of autistic children may have greater than average in- tellectual abilities (mean IQs about 116) while the mothers' mean IQs were somewhat lower. In contrast, one U.S. study (Wolff and Morris 1971) of only five sets of parents of autistics found the fathers' mean IQ to be 107.8 and the mothers' 109.8. In view of L. Wing's find- ings, it seems necessary that all such studies of parental intelli- gence must be suspect unless care

is taken to match comparison groups for SES.

Parents and Families

Parents of autistic children have been portrayed as cold, aloof, lack- ing in real warmth, and unstimu- lating to their children. Kanner (1949) described them as obses- sive, perfectionistic, humorless in- dividuals who used set rules as substitutes for life's enjoyments. Severe early stresses to the infants such as parental rejection, separa- tion, or maternal depression have been linked to autism.

Cox et al. (1975) investigated early stress events and parental warmth, responsiveness, and soci- ability in parents of autistic chil- dren with normal nonverbal intelligence and no evidence of neurological disorders with par- ents of matched dysphasic chil- dren. The parents' groups did not differ in incidence of psychiatric conditions, housing, illness, fi- nances, or interpersonal relation- ships. Ratings of emotional warmth and sociability were simi- lar except that parents of autistics spent more time with friends. Thus the parents of autistic chil- dren were as sociable, demonstra- tive, and emotionally responsive as the parents of dysphasic chil- dren.

McAdoo and DeMyer (1978fl) compared the Minnesota Multi- phasic Personality Inventories (MMPIs) of parents of autistics with a random sample of parents being treated in an adult outpa- tient psychiatric clinic. If the par- ents of autistics had severe psy- chopathology, then their MMPI profiles should have been similar to those of psychiatric patients. In- stead, significantly more psycho-

pathology was found in the identi- fied patients. MMPI profiles of parents of autistics were similar to those of a random sample of child guidance clinic parents. The fail- ure to find differences between the two parent groups might be ex- pected if there were two different groups of autistic children (Gold- farb 1970). If most of the parents of organic, autistic children were like parents of normal children, and parents of nonorganic children had significant psychopathology, then combining these two groups might minimize even dramatic differences. McAdoo and DeMyer (1978a) found that MMPI profiles were no different in parents of "organic vs. nonorganic" autistic children. Goldfarb, Spitzer, and Endicott (1976) also failed to demonstrate parent group differ- ences in psychopathology and functioning.

Infant care practices have been hypothesized to reflect deviant parent personalities, and autistic children have been viewed as "vic- tims" of extreme parent proce- dures. DeMyer et al. (1972b) found through interviews that parents of autistics and extensively matched normal children did not differ in infant acceptance, warmth, nur- turing, feeding, and tactile or gen- eral stimulation. A control parent group of nonautistic handicapped children of lower SES was judged less warm and stimulating. Nor- mal infants were recalled as more alert and responsive than autistic or handicapped infants who were similar to each other. In contrast, Massie (1978), from home movies taken before age 6 months, judged that mothers of "mixed" type child psychotics showed less adequate eye gaze and infant touching than normal controls. Feeding measures

VOL 7, NO. 3, 1981 399

for the two infant groups were judged not to differ. On the face of it, these findings offer support for the nurture causation theory of child psychosis, but low rater cor- relations (.39 to .54) and important variations in film segments cast doubt on the adequacy of the method. Nevertheless, interview methods have inherent limitations and the study bears replication with better standardized film.

Using the Ferreira and Winter Unrevealed Differences Task, Byassee and Murrell (1975) found similar family interactions in nor- mal and autistic groups, while par- ents of emotionally disturbed chil- dren had fewer spontaneous agreements than either group. Cantwell, Baker, and Rutter (1978b) demonstrated that parent- child interactions of families with autistic and dysphasic children were similar in quality and intensi- ty except that the autistic children received more interactions. Using Ittleson Center Scales, the authors found group similarities in sponta- neity, decisiveness, anticipation, control, and meeting children's demands. The mothers of autistics had the same linguistic clarity, complexity, and grammaticality to their speech. In contrast, Goldfarb, Yudkovitz, and Gold- farb (1973) judged mothers of schizophrenics inferior to normal controls in labeling and describing objects. The differences between the two studies may be the result of the criterion groups (autistics, childhood schizophrenics), the control groups (dysphasics, nor- mals) or the differences in interac- tional periods (90 minutes, 10 min- utes). King (1975), rating chart descriptions, reported more double-bind interactions in mothers of autistics than in con-

trols. He proposed that because of the mother's double-bind attitude, the autistic child "wants to es- cape."

After reviewing recent family re- search, McAdoo and DeMyer (1978b) concluded that, as a group, parents of autistic children (1) dis- play no more signs of mental or emotional illness than parents whose children have "organic" disorders with or without psycho- sis; (2) do not have extreme personality traits such as coldness, obsessiveness, social anxiety, or rage; and (3) do not possess specif- ic deficits in infant and child care.

Parent Reactions to Symptoms. An autistic child's social with- drawal, unusual object use, and lack of communication is also like- ly to produce parental uncertainty about how to respond. The long- term persistence of these behav- iors produces stress, which even- tually may result in personality changes or psychiatric symptoms. McAdoo and DeMyer (1978a) found no support for this hypoth- esis after comparing MMPI re- sponses of parents of autistics and parents of nonpsychotic child guidance clinic children. Nevertheless, in both England and the United States (Cox et al. 1975; DeMyer 1979) evidence was found that, while severe depression was rare, rearing an autistic child was stressful and that depressive symptoms were common in mothers after months to years of unsuccessful struggle to socialize an autistic child. Before pregnancy with the index child, mental illness was no more common in parents of autistics than in a well-matched normal control group (DeMyer 1979). In the Cox et al. study, fa- thers of autistics felt more often

than fathers of dysphasics that child-rearing difficulties had al- tered their wives' general behav- ior. Nevertheless, mothers of au- tistics were judged as warm to their children as comparison mothers.

DeMyer (1979) collated from extensive interviews the maternal reactions to the difficulties of rearing an autistic child. All mothers described greater than usual tension and most expressed guilt over what their role might be in producing the child's symp- toms. About one third said they had become unsure of their moth- ering capacity. Other reactions were anger over irritating behav- iors, emotional hurt over the child's lack of affection, feeling old before their time, demoralization, frustration, and curtailment of life's enjoyment. Fathers, as a group, were less expressive of their emotions but nevertheless were deeply affected, saying that the mothers' emotional pain in turn depressed them and they worried over the ultimate effect the continuing stress would have on their wives' health.

About half the normal compari- son children had a generally posi- tive effect on the marriages, while the autistic child's more irritating symptoms created "nervous ten- sion" which precipitated more quarreling and thoughts of di- vorce. About 70 percent of the mothers said they needed more "moral support" from the hus- bands in attempting to fulfill their difficult maternal role. In many cases (45 percent), the parents' sexual relations were rated as diminished chiefly because of the mother's worry about the autistic child. When either marriage part- ner expressed strong bitterness to-

400 SCHIZOPHRENIA BULLETIN

ward the mate, divorce was likely to follow. Some parents said that the decisions they made together about the autistic child ultimately strengthened their ties to each oth- er. When fathers shared the mothers' concerns about the child and gave adequate "moral sup- port" (about 33 percent of cases), then marital ties were not dimin- ished. The following were evi- dence of fathers' lack of support: critical attitudes, failure to praise, not taking the family on outings, never helping with the physical care, unreasonable or too harsh child discipline, and not talking with wives about problems or be- littling their efforts to obtain help.

An excellent summary of effects of living with and seeking help for an autistic child was given by Marcus (1977, p. 398):

parents are worried and con- fused by their child's condition; management problems have exhausted them; although vaguely aware that their child is suffering a developmental delay . . . the possibility of retardation has seemed remote; . . . [many symptoms] are annoying and at times maddening; . . . unre- sponsiveness to affection has caused personal pain; parents are fatigued and tense and the family . . . [in] considerable con- flict; in some instances, the par- ents have not been supported or provided with . . . advice they need.

Marcus advised professionals to engage the family in a collabora- tive effort while giving the parents ''the main facts of psychosis" and a "frank appraisal" of the child's cognitive impairment. "Although usually upsetting to the parent . . . [this discussion] begins the process of reducing expectations to more realistic proportions." Also parents need direction and hope,

qualities that derive from in- creased control . . . over their dilemma. The clinician should have practical solutions to real problems, be willing to work di- rectly with the child, work for community involvement such as appropriate school placement, and, in general, take the side of the family in its struggle with a life-long disorder, [p. 398].

Despite confusion and anxiety, parents have been found by Schopler and Reichler (1972) to be able to assess their child's level of development with reasonable accuracy, a result confirmed by Wing and Gould (1978). Schopler and Reichler (1972, p. 398) inter- preted the finding as suggesting that:

parents can be trusted in their evaluation of their child and in the information . . . they pro- vide. . . . Professionals, howev- er, because of their distrust of parental judgment, frequently superficially reassure, or even worse, humiliate parents by labeling them overconcerned and overanxious, or blame the child's problems on parental management. Such criticism without appropriate evaluation and specific instructions in im- proved management can . . . leave parents on their own to cope with extremely difficult de- velopmental problems, discour- aging their seeking of further help, and harmfully delaying appropriate interventions.

Parental Cognitive Characteris- tics. Evidence is mixed as to whether parents of autistic chil- dren have different cognitive characteristics than controls. Schopler and Loftin (1969) showed that parents of autistic children who were interviewed about one of their normal children obtained the same Object Sorting Test (OST) scores as control parents, while those in psychotherapy at-

tained higher scores. The authors concluded that anxious confusion about the autistic child could lead to atypical thinking. Netley, Lock- yer, and Greenbaum (1975) found that more parents of autistic child- ren were cognitively disordered than parents of nonautistics on the basis of the Grid Test of Thought Disorder scores. In contrast Lennox, Callias, and Rutter (1977), who used both the OST and Grid Test, found parents of autistics similar to parents of normals. The two tests failed to agree with each other, and there was no consistent association between thought dis- order and anxiety. SES member- ship affected the scores. Clearly more work is needed to explain the contradictions (Netley and Lockyer 1978).

Genetics

In an excellent review of possible genetic components of infantile autism, Spence (1976) concluded (1) that comparisons of identical and fraternal twins suggested a possible genetic mechanism which would not involve a single gene or the chromosomes but would have to be polygenic or multifactorial; (2) these hypotheses would be complicated to test because of in- sufficient family data and the dif- ferent aggregation of symptoms in various autistics; (3) while weak evidence from twin studies sup- ported a role for genes, no evi- dence pointed to a specific mecha- nism. Among those theories Spence hypothecated was that cer- tain aspects, or symptoms, of au- tism were inherited while others were not.

Folstein and Rutter (1977) gave evidence tending to support this

VOL 7, NO. 3, 1981 401

hypothesis. They located 21 same- sexed twins, at least one of whom in each pair was diagnosed as autistic without associated diag- nosable neurological disorder. Four of 11 sets of monozygotic (MZ) twins were both autistic, while none of the 10 dizygotic (DZ) co-twins were both autistic (p = .055). Of most interest was that all 25 autistics met at least two of four criteria for cognitive and linguistic impairment in contrast to only six nonautistics. Nine of 11 MZ twins (in contrast to 1 of 10 DZ twins) were concordant for cogni- tive or language disability (concor- dance rates 82 percent MZ and 10 percent DZ, p = .0015). This differ- ence more than met Spence's (1976) ratio of 80:20 as a require- ment for genetic involvement. Most importantly, the basic dis- ability in autism was linked to a cognitive-language disorder, an idea convincingly demonstrated by Churchill (1972). (See also Lan- guage and Intelligence sections.)

In the 17 discordant pairs, the autistic twin suffered a biological hazard, usually perinatal, which did not affect the nonautistic co- twin. According to Rutter (1977), this latter finding is consistent with a multidetermined model of autism in which at least one of the factors is biological.

Kolvin et al. (1971) found that early onset cases differed from late onset cases in symptoms, out- come, and in familial rates of schizophrenia. Higher schizo- phrenia rates were found in fami- lies of late onset cases. The au- thors concluded that late onset child psychosis resembled adult schizophrenia genetically while infantile psychosis was unrelated to either adult schizophrenia or late onset child psychoses. Hanson

and Gottesman (1976) agreed and gave some evidence that early on- set psychosis closely resembled or- ganic pathology of childhood. However, Lobascher, Kingerlee, and Gubbay (1970) found that alcoholism, psychiatric illness, and mental retardation occurred more often in families of autistics than in normal controls.

A family history of speech delay was reported in about 25 percent of autistic families by Bartak, Rut- ter, and Cox (1975) and by Rutter, Bartak, and Newman (1971). De- Myer (1979) found that siblings of autistics and of normal controls were not different in incidence of learning plus speech problems (autistics 20 percent, normals 15 percent). When learning, speech, and physical defects were com- bined, however, sibs of autistics tended to have a greater incidence of such difficulties (autistic sibs 36 percent; normal sibs 21 percent). About 69 percent of autistic ex- tended families had a member with a defect compared to 30 per- cent of normal families (p < .01).

The low incidence of autism among the siblings of autistics ap- pears to negate a genetic effect. When the estimated sibling inci- dence of about 2 percent (Rutter 1968) is compared with the inci- dence in the general population of 4.5 per 10,000 (Lotter 1966), how- ever, the rate of autism in the sib- lings is clearly much higher.

In conclusion, twin and other sibling studies lend weak support to genetic inheritance which may not involve the complete autistic syndrome, but only its language- cognitive components. As a cautionary note, Handford (1975) and Campion and Tucker (1973) have seriously challenged one key basis for the genetic hypothesis of

schizophrenia: the increased concordance in MZ twins over DZ twins. The prenatal environments of MZ twins, contrary to common belief, may not be identical. About 70 percent of MZ twins (in contrast to 30 percent of DZ twins) share a single chorion and therefore are subject to periodic reduction of blood supply to one or the other twin, thereby causing cortical infarction in near-term twins (Towbin 1970). Also MZ twins have a greater incidence of neurological dysfunction than DZ twins (Campion and Tucker 1973). The near similarity in concordance rates for single chorions and cog- nitive disabilities in MZ twins (70 and 80 percent) and the lower but similar rates in DZ twins (30 and 10 percent) make it imperative that the incidence rate of single chorion be determined in any twin study. As Handford (1975, p. 193) pointed out, "without corrobora- tions of the twin studies, the case for a genetic basis of schizophrenia is greatly weakened/' The same can be said about infantile autism.

Measured Intelligence

In the largest followup study to date of intelligence estimates from standardized tests, DeMyer et al. (1974) tested 135 autistic children (mean age 5.32 years) at initial evaluation and 70 of these same children 6 years later (mean age 11.45 years). In the preschool years, about 74 percent had a gen- eral IQ below 52 and only 2.6 per- cent had an IQ above 85. While "performance" IQs were higher than verbal IQs, nevertheless 78 percent of autistics had perform- ance IQs below 68. Initial and followup IQs correlated .70

402 SCHIZOPHRENIA BULLETIN

(p < .001); and most of the child- ren, even those who had shown considerable social improvement, remained in the retarded range. The mean IQs of those 25 autistic children who received several years of treatment and special ed- ucation were the same at followup as the 30 untreated cases. This study confirms Hingtgen and Bryson's (1972) summation of the 1960s: (1) Most psychotic children score in the mentally retarded ranges; (2) IQs are remarkably sta- ble over time; (3) splinter skills do not reduce the validity of IQ scores; and (4) IQ scores are pre- dictive of outcome. Two separate investigations (Rutter and Lockyer 1967; DeMyer et al. 1973) have found that an IQ below 40 is pre- dictive of a poor outcome. While not all children with IQs above 60 or 70 have a good outcome, the ones who make accelerated intel- lectual progress and make a better social adjustment come from this higher IQ group.

Early onset cases may have low- er mean IQs than later onset cases. Kolvin, Humphrey, and McNay (1971) reported that 51 percent of 47 early onset cases were either "untestable" or had IQs under 50 in contrast to about 3 percent of 30 late onset cases. Pollack et al. (1970) found the mean IQ of a mixed group of autistic and schiz- ophrenic children (n = 76) to be 70.5, with 72 percent of scores be- low 90. The 113 siblings' mean IQ was 111.6

In a search for an intelligence testing instrument that would meet the testing needs (i.e., wide scatter and infancy basal mental ages) of the young autistic child, DeMyer, Barton, and Norton (1972) culled items from 13 standardized tests for infants and

children. The "easiest" items in each of five classes could be suc- cessfully performed by infants, and the "most difficult" items could only be performed by chil- dren with a mental age of 6 years and higher.

Beginning with the easiest items and reinforcing for attempts to perform, the authors found that nearly every preschool autistic child was testable and that a pro- file of mental ages could be de- rived for language, motor, and perceptual-motor modalities. Of 97 consecutive referrals, nonautistic subnormal children tested higher than autistics in all modalities. The autistics did relatively better in perceptual-motor and motor tasks than in intellectual and verbal tasks. Lowest scores were verbal abstract reasoning and highest scores in fitting and assembly tasks—that is, Seguin formboard and Wechsler Intelligence Scale for Children (WISC) object assembly and block design subtests. In followup testing, mean age IIV2 years, the entire WISC profile could be successfully used with only about 28 percent of autistics, while the remainder had to be test- ed with a battery suitable for chil- dren with mental ages below 6 years. Again at followup, autistic children's highest scores were in object assembly and block design and the lowest in verbal compre- hension.

Lockyer and Rutter (1970) first demonstrated in autistics, whose mental age was high enough to be testable using the WISC, that autistics achieved lower compre- hension and higher block design and digit span scores than a con- trol group of nonpsychotic behav- ior disturbed controls. The autis- tics were inferior to the controls in

the Vineland Test of Social Com- petence. Lockyer and Rutter (1970, p. 152) hypothesized that autism represents "a central disorder of language and of perception of sounds." Walker and Birch (1974) found in 120 schizophrenic boys, ages 10 to 15, that performance IQ was superior to verbal IQ in those with WISC IQ over 75. In children below IQ 75, verbal scores were relatively better than performance scores. Tymchuk, Simmons, and Neafsey (1977) compared WISC subtest scores between 20 adolesc- ent autistics and 20 other exten- sively matched children of mixed nonpsychotic diagnoses (mean IQ: autistics 88, contrast group 89). The autistics achieved lower com- prehension scores and higher block design scores than nonpsychotics. The autistics showed significant subtest varia- bility with comprehension, coding, and picture arrangement being lower and digit span and block de- sign being higher.

Profiles of intelligence test re- sults meet objections that a single IQ score in face of the wide scatter found in many autistic children is not useful. While the general IQ is a powerful statistic in predicting the overall illness outcome, the more useful statistics in planning education programs for an indi- vidual child are the individual points of his profile of test scores.

In conclusion, the intelligence of each autistic child should be esti- mated using tests that are within his capability. The goal should be not to diagnose mental retardation versus infantile autism but to judge the degree of intellectual re- tardation or adequacy in the various facets of intelligence. As Wing (1979) pointed out, one of the biggest advances of the last

VOL. 7, NO. 3, 1981 403

decade has been an increasing awareness of how autistic children compare with developmental norms and with other handicapped groups. Standard- ized intelligence tests have provid- ed one good yardstick.

Language Gross disturbances in language development are a generally ac- cepted major characteristic of in- fantile autism. Indeed, some re- searchers regard disorders in language coding and speech to be the primary defect in autism (e.g., Rutter, Bartak, and Newman 1971; Churchill 1972, 1978). In an exten- sive study of 47 boys, aged 5 to 10, who all possessed a severe de- velopmental language disorder with no demonstrable neurological dysfunction or hearing loss, the British group of Rutter and his col- leagues (Bartak, Rutter, and Cox 1975) initially reported that within this group, those boys diagnosed as autistic had more deviant lan- guage development, more severe comprehension defects, and also showed more defects in social us- age of the limited language that they did have. Since few differ- ences were seen in the groups with regard to patterns of nonlin- guistic skills, they concluded that a language disability is probably necessary for the development of the behavioral syndrome of infan- tile autism.

This group continued their re- ports on the same population of boys over the next few years (Cantwell, Baker, and Rutter 1978fl), and the data have recently been reviewed by Rutter (1979). He formally broadens the concept of a central language deficit to be included under a cognitive deficit manifested by impaired language,

sequencing, abstraction, and coding functions. When compared to a group of dysphasic children, the autistic children did not differ significantly in visual-spatial per- ception, articulation, or syntactical skills. Marked differences in the two groups, however, were ob- served in five areas: Autistic boys scored lower on understanding of language and language expres- sion, lower on imaginative play and understanding of gesture, lower on spontaneous speech and speech for social communication, and higher on echolalia (especially self-echolalia and delayed echola- lia); they also exhibited a highly variable pattern in verbal scale scores on the WISC. Rutter con- cludes that a particular kind of cognitive/linguistic deficit funda- mental to autism is as much asso- ciated with language deviance as it is with language delay.

The lack of spontaneity in speech, even in those autistic chil- dren with some communicative speech, has been noted by many investigators. The Illinois Test of Psychological Abilities was used to assess and compare language- related functioning in 20 autistic and 20 retarded children (Prior 1977b). In this Australian popula- tion, it was found that higher functioning groups of both diag- noses differed from lower func- tioning groups overall; but for both high- and low-functioning autistic children, verbal and man- ual expressive performance was particularly impoverished and in- dicated a severe deficit in sponta- neous communicative ability. The absence of spontaneous protode- clarative gestures, a preverbal form of intentional communication occurring around the first year of life for normal infants, in a group

of 12 mute autistic children sug- gested a qualitatively distinct pat- tern of prelinguistic development (Curcio 1978). This severe paucity of gestural usage by autistic chil- dren also had been pointed out by others (Wing 1971; Bartak, Rutter, and Cox 1975).

Shapiro's work (Shapiro, Chiarandini, and Fish 1974; Shapiro and Huebner 1976) led him to conclude that various groups of autistic children can be categorized according to their lan- guage deficits, which he regards as central to the developmental dis- order. In a 1976 followup study of five autistic children during ado- lescence, speech samples of the children between the ages of 12 and 19 were analyzed and com- pared to earlier speech patterns (8-12 years). The speech patterns of the children when older showed the expected varying combinations of constraint in length of utter- ance, echoing, syntactic disturb- ance, semantic concreteness, con- text inappropriateness, and disorders of prosody. The most dramatic finding was that early de- scriptions of these children were reliably matched to later speech samples by 10 psychiatrists (using blind matching techniques), indi- cating a continuity in style of speech deviance. Simmons and Baltaxe (1975) have also hypothe- sized the existence of distinct lin- guistic subgroups of autistic chil- dren, perhaps related to disa- bilities of rhythmic and prosodic perception.

Restricting their investigations to verbal autistic children, Barto- lucci and his colleagues have made a comprehensive study of the lan- guage of these English-speaking Canadian children compared to that of mentally retarded and nor-

404 SCHIZOPHRENIA BULLETIN

mal children, all matched for nonlinguistic mental age (approxi- mately 6 years: Leiter Performance Scale). Autistic children used cor- rect verb tense only 8 percent of the time during a structured pic- ture and toy test session vs. 80 percent and 60 percent for normal and retarded children, respectively (Bartolucci and Albers 1974). When the free speech syntax of autistic children was compared to retarded and normal groups, the former had lower developmental sentence and grammar complexity scores and higher error rates (Pierce and Bartolucci 1977). The grammatical system of the autistic group did appear rule governed, but simpler in construction than that observed in the other two groups. Although autistic children showed a more extreme and global delay in language development, they appeared similar to retarded children with regard to simple phonological perception or pro- duction, as well as in their fre- quency of omission of 14 normally occurring grammatical morphemes (Bartolucci and Pierce 1977; Bartolucci, Pierce, and Streiner 1980; Bartolucci et al. 1976). Baltaxe and Simmons (1975) found some of the same deviancies in their sample of verbal autistic ado- lescents. When compared to Down's syndrome adolescents, the deficits in the retarded group cen- tered primarily in the area of artic- ulation and syntax, whereas the autistic group's deficits related pri- marily to prosodic features of lan- guage, semantics, an inability to switch linguistic codes, and a manifestation of echolalia. Regard- ing articulation, Boucher (1976b) also found her sample of verbal autistic children to have superior ability to both retarded and

dysphasic control groups. Because of the prominence of

echolalia in the speech of autistic children, this topic has fascinated many investigators. In a series of studies on echoing in autistic chil- dren, Shapiro and colleagues (Shapiro, Roberts, and Fish 1970; Shapiro 1977; Shapiro and Lucy 1978) have been able to make the following observations: (1) The to- tal number of responses of a child who echoes is greater than that of a comparable nonechoing child at a similar developmental stage; (2) the mean length of utterance of echoes is longer than the mean length of other responses because echoes bypass intrinsic coding de- vices; (3) echoing occurs when a question taxes the child's receptive understanding; (4) echoes tend to cluster once an echo chain is set in motion; (5) there is a difference in reaction time for echoing and nonechoing (spontaneous) re- sponses in autistic children, whereas this is not seen in normal children.

An experiment by Carr, Schreib- man, and Lovaas (1975) demon- strated that echolalic children were most likely to echo questions and commands to which they had not yet learned an appropriate re- sponse, but rarely echoed ques- tions and commands to which they had learned an appropriate re- sponse. Further, once an echolalic was taught an appropriate re- sponse to a previously echoed question, the child no longer echoed the question on future oc- casions. Although it appeared that echolalia could be reduced by this method, it was judged not feasible to teach a child a response to every verbal stimulus that might be en- countered. Thus, in a subsequent study (Schreibman and Carr 1978),

two echolalic children were taught to respond to a set of previously echoed questions with the sen- tence, "I don't know." At the same time, their nonecholalic re- sponding to those few questions to which they already had an appro- priate response was maintained. Then the children were tested on a different set of previously echoed questions to see if the treatment intervention generalized to the new questions. It was found that each child discriminated correctly between those questions that had previously been echoed and those that had not. Followup probes showed that treatment gains were maintained at least 1 month later.

Other investigators have viewed echolalia in autistic children as a manifestation of late-onset imita- tive speech related to the central language disorder in autism (Phil- ips and Dyer 1977). If echoic mem- ory for words is regarded as a peak ability in the more able autistic child (Boucher 1978), similar to digit-span ability, this near normal facility, in contrast to most other behaviors at levels far below nor- mal functioning, could account for the fixation at this stage of normal language development. In an in- teresting study comparing a group of 12 autistic and 12 dysphasic boys of matched ages (the same population as used in the Bartak study; see above), the data con- firmed that echoing is prevalent in both these types of children and that the two groups cannot be dis- tinguished by the total amount of echoing (Cantwell and Baker 1978a). The only significant differ- ences between the two groups were in the amount of inappropri- ate repetitions of self and the amount of delayed echoing. For both of these types of echoes, the

VOL. 7, NO. 3, 1981 405

incidence was higher in the autis- tic group, but there were no differ- ences between the groups in the amount of exact, reduced, expand- ed, mitigated, or congruent echoes.

A number of attempts have been made to teach sign language to au- tistic children, especially to that group which shows little progress in expressive or receptive verbali- zations even after intensive behavioral therapeutic interven- tion (Webster et al. 1973; DeVill- iers and Naughton 1974; Creedon 1975; Eulwiler and Fouts 1976; Benaroya et al. 1977; Carr et al. 1978). The majority of these inves- tigations suggest that while the re- sults are highly variable from child to child, children who have some imitative verbal skills before sign language training show gains after this training; whereas mute chil- dren, who usually do learn some signing, rarely learn to talk (Miller and Miller 1973; Salvin et al. 1977; Carr 1979). There is also some in- dication that abstract concepts, syntax, and generative skills can be taught through sign language, and that some children do show increases in spontaneous commu- nication, decreases in self- stimulatory behavior, and some improvement in socialization after sign training (Casey 1978; Carr 1979). Despite the promise indi- cated by some of the earlier re- ports, however, sign language has, by and large, not produced the dramatic gains in language skills in autistic children that some researchers at first suggested. On the other hand, sign language or other types of nonspeech language training (McLean and McLean 1974; LaVigna 1977; Churchill 1978), as well as automated meth- ods (Colby and Kraemer 1975;

Hargrave and Swisher 1975), are all successful in increasing the to- tal behavioral output of the autistic child, and to this extent should be considered for incorporation into the treatment program designed for each autistic child. In fact, re- cent evidence strongly suggests that a multisensory approach, combining both verbalization and sign techniques, is more effective than either approach by itself (Brady and Smouse 1978; Benaroya et al. 1979; Konstantareas, Web- ster, and Oxman 1979; Barrera, Lobato-Barrera, and Sulzer- Azaroff 1980).

Two widely accepted clinical im- pressions of the language of autis- tic children can no longer be con- sidered tenable as a result of research conducted in the 1970s. It has been frequently stated that the autistic child's language is atypical because of the manner in which his parents respond to his at- tempts to communicate. While many parents do develop strate- gies of communication in an at- tempt to circumvent the language disabilities of their autistic chil- dren (Ricks and Wing 1975), nu- merous studies have yielded no evidence that the speech of par- ents had a detrimental effect on that of their autistic children (Frank et al. 1976; Cantwell, Baker, and Rutter 1977; Cantwell and Baker 1978b). An interesting sidelight to this issue is the obser- vation that the child's own verbal behavior provides a significant source of sensory reinforcement in autistic as well as in normal chil- dren (Lovaas et al. 1977). The sec- ond clinical view was that autistic children refuse to use the pronoun " I " because of a disturbance in perception of self. Bartak and Rut- ter (1974) showed that autistic chil-

dren imitated pronouns at the ends of sentences, including " I , " no matter what pronouns were used. Silberg (1978) carried this re- search a step further and was able to demonstrate that in her sample of autistic children, those pos- sessing some of the least verbal behavior used the pronoun " I " more frequently than the second or third person pronouns.

The extreme deficiencies in abili- ties for concept formation (Noach 1974) and the overwhelming de- pendence on the use of imitation as a major strategy in linguistic coding (Shapiro and Kapit 1978) in those autistic children who do possess some language skills pre- sent a rather pessimistic outlook for the eventual development of true spontaneous speech in the majority of autistic children, even with intensive behavioral proce- dures (Lovaas 1977; Carr 1979; Goetz, Schuler, and Sailor 1979; Prior 1979). In spite of this gener- ally poor prognosis, the extensive studies of language and language training in autistic children cer- tainly are justified by the improve- ments in communication that have been achieved (see Treatment sec- tion) and by the possibility for a better understanding of language development in both autistic and nonautistic disorders.

Perceptual Process

The extreme difficulty that most autistic children experience with even the most rudimentary intel- lectual and language develop- mental tasks has often been attrib- uted to underlying perceptual disturbances. Indeed, studies in the 1960s strongly suggested this as a possible source for the autistic child's behavioral deficits. Has this

406 SCHIZOPHRENIA BULLETIN

early suggestion been supported by the past 10 years of research? Some of the major experiments will be reviewed in order to an- swer this question.

One of the new areas of percep- tual research with autistic children since 1970 has been the concept of stimulus overselectivity as pro- posed by Lovaas and his col- leagues. This phenomenon refers to the fact that autistic children tend to respond only to a few cues from a larger range of available cues in a learning task. In an initial study (Lovaas et al. 1971), three groups of children (autistic, re- tarded, and normal) were re- warded for responding to a com- plex auditory-visual-tactile set of simultaneous stimulus cues. A dis- crimination having been estab- lished, elements of the complex stimulus were presented to assess which aspects of the complex had acquired control over the child's behavior. The autistic children re- sponded to only one of the cues, whereas the normals responded uniformly to all three, with the retardates functioning at a level in between. When the stimulus com- plex was simplified to only two types (white noise and red flood- light), the autistics still demon- strated stimulus overselectivity be- cause seven out of nine subjects responded to only one of the com- ponents (Lovaas and Schreibman 1971). In both studies, however, the previously nonfunctional stim- ulus was made functional when trained separately.

Although the early studies indi- cated that reinforcement during successive training trials was in- strumental in reducing stimulus overselectivity, a later study (Schreibman, Koegel, and Craig 1977) demonstrated that training

per se, rather than reinforcement, was an important factor. Nineteen autistic children were trained on a discrimination task with a cue complex composed of two visual cues. When they reached criterion, they were presented a testing phase involving probe trials in which the cue components were presented singly. The following results were obtained: 16 children initially showed some overselec- tivity, but 13 of these children de- creased their level of over- selectivity within less than 48 probe trials. Since there was no re- inforcement for responses to the single component during probe trials, and yet a reduction in overselectivity occurred, the re- sults suggested that repeated ex- posure to testing, rather than rein- forcement, was the important factor in reducing overselective at- tention to cues. This finding is es- pecially interesting because overselectivity was also signifi- cantly reduced when partial rein- forcement (variable ratio 3) was used during part of the training, compared to continuous reinforce- ment used throughout the training (Koegel et al. 1979).

Additional studies have shown that stimulus overselectivity is clearly related to a lowered mental age (Wilhelm and Lovaas 1976) and thus is probably closely asso- ciated with behavioral retardation across different types of children rather than being a characteristic response tendency of only the au- tistic child (Koegel and Lovaas 1978). However, it may still reflect a critical perceptual deficit in in- fantile autism, even though its ini- tial cause may differ from that in nonautistic retarded children.

It should be noted that at least one negative study has been re-

ported (Litrownik et al. 1978) which suggested that when autis- tic children are compared to nor- mal children of the same approxi- mate mental age, they perform more similarly to the normal chil- dren on matching-to-sample tasks than do Down's syndrome chil- dren. These researchers claimed that their results indicated that overselective attention is more characteristic of retardate than au- tistic responding, with the autistic response deficits attributable to a possible stimulus control or lack- of-retention factor that develops over a series of training trials. Since this study had many proce- dural differences from the previ- ous experiments, the data are sub- ject to multiple interpretations, but they do point out the need for fur- ther controlled studies of the overselectivity phenomenon in au- tistic children.

Perhaps related to the stimulus overselectivity response in autistic children is their apparent inability to use prompt fading techniques to facilitate learning. Prompt fading (the introduction of an additional cue to aid discrimination, with the gradual removal of that cue when correct responding occurs with high frequency) has been used with some success for teaching simple and complex discrimina- tions to retarded children. When used with autistic children, how- ever, it appears to slow learning rather than improve it. Schreib- man (1975) reported that whereas autistic children usually failed to learn stimulus discriminations without a prompt, they always failed to learn with an extra- stimulus prompt (the type of prompt fading described above). When within-stimulus prompts were used (an exaggeration of the

VOL 7, NO. 3, 1981 407

relevant component of the training stimulus, thus not requiring the child to respond to multiple cues), the children usually learned the discrimination task (similar to the study of Sherman and Webster 1974). Those results were con- firmed by later studies (Koegel and Rincover 1976; Arick and Krug 1978) and indicated that extra- stimuli prompt fading was actually detrimental to learning in autistic children, and even had a negative influence on transfer of therapeu- tic training effects from one setting to another (Rincover and Koegel 1975). Extensive studies of the im- portant variables in prompt fading that might reduce the attentional requirements for discrimination learning in autistic children, such as the recent experiments of Rincover (1978b), are needed to understand better the bizarre na- ture of multiple stimulus organiza- tion in these children and to facili- tate the design of more effective training programs. Errorless criterion-related stimulus shaping (rather than stimulus fading) should also be considered as a po- tentially useful procedure with au- tistic children (Etzel and LeBlanc 1979).

There is some evidence that au- tistic children impose their own perceptual patterns on incoming sets of stimuli. In two studies of pattern detection (Frith 1970a, 1970b), one dealing with immedi- ate recall of auditory sequences and-the other dealing with repro- duction of color sequences, when normal and autistic children of comparable performance levels were tested, the latter subjects tended to show a marked response bias, imposing their own inde- pendent patterns on input stimuli rather than using the structures

present in the stimulus patterns. Subsequent studies have support- ed these conclusions (Frith 1972; Hermelin 1972; Hermelin and O'Connor 1971). Also indicative of the unusual nature of perceptual processing in autistic children is the fact that simple photic stimula- tion can serve as a powerful rein- forcer for lever-pressing in relation to the frequency of stimulation, which was not the case for re- tarded children (Frankel et al. 1976). Similar results were also re- ported for vestibular stimulation (Freeman, Frankel, and Ritvo 1976).

Multiple and complex cues pose special problems for autistic chil- dren. Although autistic subjects were able to successfully acquire a simple object discrimination learn- ing set (Prior and Chen 1975), they failed to solve a conditional matching learning set problem requiring the use of an abstract symbol as a cue for the correct re- sponse (Prior 1977a). This failure was in contrast to the success of their normal and retarded counter- parts, all three groups being matched for mental age (about 5 years), and demonstrated the diffi- culties that autistic children have in dealing with complex symbolic input. When presented double cue visual discrimination problems (using varying stimuli of color, form, and size) on a modified Wisconsin General Test Appara- tus, nonverbal autistic children performed significantly more poorly than either verbal autistic or Down's syndrome children (Kovattana and Kraemer 1974). Using a stimulus generalization paradigm, autistic children per- formed as well as normal children in a test with single stimuli, but they overgeneralized when tested

with complex stimuli by re- sponding to fewer features of the stimulus (Fein, Tinder, and Waterhouse 1979).

Abnormalities in audition have been found in some groups of au- tistic children (Chess 1977; Hayes and Gordon 1977). In a group of dysfunctional autistic-like chil- dren, Condon (1975) reported mul- tiple responding to auditory stimuli as if the children were perceiving the sound as being presented more than once. He also saw dyssynchronous or awkward overall behavioral response organ- ization in these children. This lat- ter finding was supported by the research of Oxman, Webster, and Konstantareas (1978), but they could not replicate the multiple re- sponding observation. Further in- direct evidence for atypical audito- ry perception was provided by Fassler and Bryant (1971), who found that reduction in auditory input through the use of ear pro- tectors improved classroom atten- tion and task performance in autis- tic children.

Disturbances in perceptual input mechanisms could be related to cross-modal association and/or deficits that some studies find evi- dence for in autistic children. In addition to severe problems with learning auditory-visual associa- tions, Bryson (1972) found that most of the autistic children tested also had substantial visual-visual short-term memory deficits, an ob- servation made earlier by Her- melin and O'Connor (1970). How- ever, when autistic, retarded, and normal children were matched for mental age (approximately 5 years), there were no differences in visual short-term memory among the three groups, although recall deteriorated with increase in

408 SCHIZOPHRENIA BULLETIN

the delay interval (30 to 120 sec- onds) for all groups (Prior and Chen 1976). The children also did equally well on a visual serial memory task. Further evidence for auditory-visual association defects was provided by Morton-Evans and Hensley (1978) in their study of four groups of children (autistic, aphasic, retarded, and normal) matched for nonverbal mental age (4.4 years). The autistic and recep- tive aphasic children were both slower in associating sounds with their visual counterparts, but the aphasic children overcame this deficit at a significantly faster rate than the autistic children. No sig- nificant differences were found among the groups of children in learning visual-visual* associations. Abnormalities in developing spe- cific visual-motor associations by autistic children have also been described (DeMyer 1975fl; Fulker- son and Freeman 1980).

In a recent study, Koegel, Dunlap, and Dyer (1980) found intertrial interval duration to be di- rectly related to frequency of cor- rect responding. Intervals of 1 sec- ond led to higher levels of correct responding and improving trends in performance, whereas intervals of 4 seconds showed large num- bers of errors with no sign of im- provement even over a long series of trials. It should be noted that re- inforcement was always presented immediately after the correct re- sponse; only the beginning of the succeeding trial was delayed. The activities that occurred during the intertrial intervals, such as self- stimulatory behaviors (Lovaas, Litrownik, and Mann 1971), could have played an important role in interfering with normal memory

processes, or the data could indi- cate a deficit in the memory proc- ess itself.

Lack of eye-to-eye contact, fail- ure to recognize family or self, and lack of interest in or withdrawal from interpersonal relationships are commonly described features of autistic behavior and have been hypothesized to be related to per- ceptual deficits. Although autistic children are clearly deficient in body imitation (DeMyer et al. 1972a), which strongly suggests a self-perception disturbance, the evidence for low frequency of eye- to-eye contact and visual self- recognition is negative. Churchill and Bryson (1972) failed to find any differences in 14 autistic and schizophrenic children when com- pared to matched normal controls in terms of looking at and ap- proach toward a strange adult. Both groups also responded more to an attentive adult than to a nonattentive adult. Another study (Neuman and Hill 1978) found that autistic children above the age of 5 were capable of self-recognition and displayed an intense interest in their own images and attempted to manipulate and control them.

An overview of the evidence for severe perceptual deficits in autis- tic children appears to make a sol- id case for the presence of a signif- icant disturbance in perceptual processing. Whether the underly- ing mechanism is related to an attentional factor, a complex per- ceptual processing mechanism, a memory factor, or some other vari- able, however, has not yet been clearly established. In addition, the relationship of these defects to mental retardation needs to be fur- ther delineated.

Behavioral Characteristics

Hingtgen and Bryson (1972) sum- marized behavioral studies of au- tistic children in the 1960s as fol- lows: (1) Behavior remains stable and relatively unaffected by events or fatigue except that avoidance and stereotyped behavior may be related to success/failure ratio in structured tasks. (2) Repetitive, nonfunctional behavior with ob- jects and their own bodies consti- tutes a large portion of the autistic child's repertoire. (3) While there was conflicting evidence about whether autistics avoid eye-to-eye contact, it appeared that they look less than nonautistics at all envi- ronmental objects. When looking at adults they lack other behaviors considered indicative of attention in normal children. The behavioral studies of the 1970s were few in number, but the findings were im- portant because the authors took care to look into the developmen- tal nature of the symptoms they studied.

Play. The poverty of symbolic play seen regularly in infantile autism was postulated by Wing et al. (1977) to come from an inability to abstract concepts from experience and to store these abstracts in symbolic form. From structured parent interviews and direct child observations of 108 autistic and re- tarded children, there emerged three types of play groups: (1) 43 with flexible and varied symbolic play, occurring only in those with language ages above 19 months and in no autistic child; (2) 23 with stereotyped, repetitive copying play; and (3) 42 with no symbolic play, all of whom had a language age below 20 months. Most chil-

VOL 7, NO. 3, 1981 409

dren in groups 2 and 3 were autis- tic or had "marked autistic fea- tures." Thus the conditions of stereotyped play (found also in normal 2-year-olds) was related both to the diagnosis of autism and to developmental age.

Observing five autistic boys in four different environments, Black, Freeman, and Montgomery (1975) confirmed the repetitive, nonfunctional object use and the paucity of peer interaction. In a confined space, the boys "mod- eled and imitated" at low develop- mental- levels, but only two chil- dren engaged in any mutual play which nevertheless was parallel in nature and thus not higher than a mental age of 2 years.

Pantomime, Hand Use, and Imita- tion. Normal children of 3 years pantomime nonabstractly, and at 6 years they begin to represent an object abstractly. In a group of 24 "psychotic children" who met Creak's (1964) definition of child- hood schizophrenia, Curcio and Piserchia (1978) found that sub- jects with echolalia could produce almost no abstract forms of panto- mime while those with higher lev- els of speech produced some. Introducing a model increased the pantomimic scores. Those who discriminated a teacher from a child in a drawing task attained higher pantomime scores. All chil- dren were below the pantomimic norms for their ages.

Abnormal hand gestures, including fixed posturing and movement gestures such as flapping, clapping, rubbing, and pinching, were displayed by 16 of 18 autistic children during uncon- trolled environmental conditions (Masagatani 1973). Mental retardates often hold their thumbs

in continuous flexion (Peiper 1963), a condition also found in 9 of 18 autistics. The other half dis- played conditions resembling a partial primitive grasp. Those chil- dren with fewer abnormal hand movements had more spontaneous speech. Masagatani stated that the reflexive, clumsy, and fixed pat- terns observed in the autistic chil- dren (ages 2V2 to 13 years) should have been replaced by more skill- ful voluntary movements by 12 months of age. The author advised that the term "primitive reflex pat- terns" with physiological implica- tions should be applied to the au- tistics' abnormal hand gestures rather than the term "bizarre."

Imitation is an important activity in normal children because much learning comes from reproducing parental vocalizations and body actions. Douglas and Sanders (1968) found that failure to imitate before 3 years of age differentiated retardates from autistics. A more severe motor imitative dyspraxia was found in 12 autistic children in comparison to 9 subnormal nonautistic children (DeMyer et al. 1972a). This result suggested to the authors that most autistic chil- dren may have not only a severe abstract language defect but also a dyspraxia, both defects contrib- uting to their inability to learn the meaning of and reproduce body language which forms a large part of the infant and the young child's communication mode with others. The authors also suggested that the basis of the praxia defect might be in poor visual memories or an inability to transfer a visual memo- ry to the motor system (agnosia) and that such deficiencies might be one basis for failures of autistic children to think abstractly and to symbolize.

The seeing of common and dif- ferent properties in the "mind's eye" of various objects is essen- tial for abstract and symbolic mental operations. It is no acci- dent that the colloquial term "I see" means "I understand." [DeMyer et al. 1972a, p. 284]

Maintenance of Sameness and Negativism. Among 32 psychotic children, 10 children with > 20 Rimland E-2 scores exhibited higher "sameness" scores which reflected a greater amount of ster- eotypy in general behavior in con- trast to those children with < 20 Rimland scores (Prior and Macmillan 1973). However, two children with minus E-2 scores also had very high sameness scores. Those children with ability to create and maintain more com- plex sameness patterns appeared to have higher cognitive abilities as measured by presence of speech and school attendance than those children with less complex stereotypies.

Negativism, defined as the con- sistent avoidance of a correct re- sponse, was not demonstrated in any of 27 autistic children tested by Clark and Rutter (1977) contra- ry to findings of Cowan, Hodinott, and Wright (1965). Of all possible explanations for the difference, the authors felt that a lack of school exposure might be responsible for the negativism of Cowan's group since all of Clark and Rutter's sub- jects were well used to various kinds of school work. Successful task performance was related to language competence.

Motor and Perceptual-Motor De- velopment. The gross motor de- velopment of many individual au- tistic infants and children has been described as within normal limits.

410 SCHIZOPHRENIA BULLETIN

However, when groups of autistic children were compared with nor- mal children, the autistic children showed significantly lower levels of motor skills. In 65 autistic chil- dren of preschool and primary school age, gross motor abilities, as tested by adequacy of stair climbing, were found to be below developmental norms (mean de- velopmental quotient 52; DeMyer, Barton, and Norton 1972). When perceptual motor skills were tested by such standardized tasks as drawing geometric figures and ball play, autistic children were behind their normal counterparts (mean developmental quotients 40 and 42). Thus gross motor skills seem relatively more intact than com- plex eye-hand-body coordination skills. In a study of infants born to schizophrenic mothers in a state hospital, Fish and Hagin (1973) found that failure in hand-to-hand performances at 4 to 9 months of age was related to subsequent emotional illness at followup, age 10-11 years.

Nevertheless, superior drawing skills have been reported in indi- vidual cases (DeMyer, Barton, and Norton 1972). DeMyer (1975a) re- ported graphic skills as a splinter ability in about 3 percent of autis- tics. However, the commonest perceptual-motor splinter skills appear to be in fitting and assem- bly tasks, such as the Seguin form board, WISC object assembly and block design subtests (see Intelli- gence section). These tasks can be accomplished by recognition of similar form and outlines and by the picking up and release of ob- jects, an infant motor skill. When- ever a more advanced motor skill or cognitive judgment, as in the coding subtest of the WISC or draw-a-person test, was called for,

autistic children did relatively worse.

Parental Reports of Develop- mental Milestones. Two separate studies investigated develop- mental milestones of autistic chil- dren in comparison to matched normal control groups. Using a pa- rental paper-pencil questionnaire, Ornitz, Guthrie, and Farley (1977) foutid in 74 autistic children, mean age 45.2 months, that nearly all as- pects of development were de- layed. Receptive and expressive speech development was delayed beginning at age 2 months. By age 6 months, continuing through the second year, significant motor de- lays were reported—for example, holding head erect, first rolling onto back, sitting without support, crawling, and walking. An inter- esting finding of this carefully done study was that autistic chil- dren without a concomitant neuro- logical condition did not differ from those who had such overt pa- thology, illustrating the difficulties of basing presence or absence of brain dysfunction solely on out- ward signs of such pathology.

Using face-to-face interview methods and all available records such as baby books and photo- graphs, DeMyer (1979) also found 33 autistic children as a group to be delayed in passing develop- mental milestones. In addition to delay in language and the better known motor milestones, autistics were behind agemates in use of eating utensils, drawing geometric figures, and dancing and rhythmic performances. Music skills in both listening and performance aspects were also behind the normal com- parison group.

Even the autistic children with exceptional splinter skills such as

"perfect" music pitch lacked other abilities such as good rhythm or hand coordination or cognitive abilities which thus made impossi- ble an age-appropriate music per- formance. Most autistic children (90 percent) were interested in lis- tening to music, but even in this regard they were less advanced than the matched normals. Autis- tic children generally preferred simple music while many normal children listened to a wide variety of music, including classical mu- sic. Only the autistic children with advanced tonal skills, about 9 per- cent of the sample, listened to classical music.

In conclusion, autistic children, like normal and retarded children, obey the laws of child develop- ment in that specific behavioral features such as the ability to pan- tomime, imitate, and engage in symbolic play and successful task performance are directly related to the mental age of the child. Some evidence has accrued, however, that autistic children, in contrast to nonautistic mental retardates, may have relatively more serious prob- lems in (1) symbolic representation necessary to the development of symbolic play and abstract thought; and also (2) visual-motor dyspraxia which precludes the learning of pantomime and body imitation at levels sufficient to par- ticipate in everyday nonverbal communication. The extremely low developmental levels of both verbal and nonverbal communica- tion in turn may be related to the phenomenon of psychotic social distance in the autistic population. The failure to find the trait of neg- ativism in autistic children is of utmost importance. It indicates that the scores of autistic children on various developmental and in-

VOL. 7, NO. 3, 1981 411

tellectual tests can be trusted in most cases to reflect what the au- tistic child "can't do" and not what he "won't do." Equally im- portant are the two separate stud- ies in which parents reported later than normal development, even in the first year of life, not only of language milestones but also various other milestones studied, such as motor, music, drawing, and rhythmic skills.

Neurobiological Studies

From the 1960s, Hingtgen and Bryson (1972) concluded that neurobiological factors may be im- plicated in some cases of child- hood psychosis. Investigators of the 1970s using extensive neuro- logical examinations, including electroencephalograms, have re- ported that signs of neurological dysfunction appear more frequent- ly in autistic children than in nor- mal children and that autistic chil- dren are much like nonpsychotic subnormal children in this regard. In Kolvin, Ounsted, and Roth's (1971) series from England, 54 per- cent of infantile psychosis and 15 percent of late onset psychosis cases had some evidence of brain dysfunction. Gubbay, Lobascher, and Kingerlee (1970), from West- ern Australia, found 56 percent of 25 autistic children had "unequiv- ocal evidence of brain disease," 28 percent had probable brain dis- ease, leaving only 16 percent with no such evidence. The multiple types of neurological findings sug- gested to the authors that autism is not "produced by the malfunc- tion of a specific anatomical region of the brain" and also "that only the occurrence of the characteristic types of behavior patterns seemed to justify [the] separation diagnos-

tically [of autistic children] from any heterogeneous group of re- tarded children" (p. 427). Harper and Williams (1974) reported that 47 percent of 131 autistics had evi- dence of organicity.

In studies from the United States, DeMyer et al. (1973) using a weighted scoring system showed the mean brain dysfunction index of 48 normal children to be 11.6, of 155 autistic children 45.9, and of 26 subnormal children 52.5. Only 14.2 percent of the autistics' scores fell within one standard deviation of the normals. Knobloch and Pasamanick (1975) reported find- ings with similar implications and stated unequivocally that all young autistics they examined showed signs of neurological dysfunction (see Classification and Diagnosis of Childhood Psychoses section). In a comparison of 76 schizophre- nic and autistic children with 113 of their biological sibs, 70 percent of index cases had soft signs and 6 percent hard signs while only 15 percent of sibs showed soft signs and 2 percent hard signs (Pollack et al. 1970).

Other authors using more cir- cumscribed evaluation methods have reported increased incidence of "minor" physical anomalies (e.g., low set ears, electric hair, high arched palate) among psy- chotic groups vs. normal groups (Steg and Rapoport 1975; Walker 1977b). In 52 autistic children, Campbell et al. (1978b) demon- strated significantly more anoma- lies than in either a sibling or nonrelated normal control group while the latter two groups did not differ. A presumptive connection exists between presence of minor physical anomalies and first tri- mester pregnancy complications. Anomaly scores tended to be

higher in 17 patients with a history of such complications than in 35 without them. A later study by Campbell et al. (1980) reported that young autistic children were significantly shorter than a control population.

Abnormal finger print patterns, formed by the 18th week of gesta- tion would signify that something had gone awry in the first or early second trimester. Dermatoglyphic patterns are largely under genetic control, but other factors such as rubella virus and thalidomide can cause abnormal patterns. The fin- gerprint patterns of 20 schizophre- nic children were described by Hilbun (1970) as not abnormal in and of themselves, but the distri- bution of the types of patterns was significantly different than that of a randomly selected normal popu- lation (Alter and Schulenberg 1966). Walker (1977a) reported that 73 of 78 autistic cases showed one or more dermatoglyphic abnormalities, most of which were extensive, in contrast to 43 of 78 normal patterns in a normal group. Also, ridge pattern fre- quencies differed markedly be- tween Walker's autistic group and other diagnostic groups.

Handedness and Right-Left Ori- entation. Normal children develop an adult type hand preference at 3 to 4 years, while subnormal chil- dren usually have mixed handed- ness (ambidextrousness) until later ages (Annett 1967). Apparently autistic children resemble subnor- mal children in this regard. Only 35 percent of preschool autistics had developed a right-hand domi- nance in comparison to 88 percent of normal controls (Colby and Parkison 1977). In children ages 6 to 19 years, right-hand dominance

412 SCHIZOPHRENIA BULLETIN

was displayed by 59 percent of au- tistics and 74 percent of normals (Barry and James 1978). Dominant hand usage (whether right or left) was related significantly to chron- ological age in all children, show- ing its developmental nature.

Of a population of 80 schizo- phrenic boys (8 to 12 years old and IQs above 70), 32 percent always used their right hand, 12 percent always used their left hand, and 55 percent were mixed (Walker and Birch 1970). In contrast, only 8 percent of normal boys had mixed dominance (Belmont and Birch 1963). Mixed eye/foot dominance was also more common in the pa- tients. In tests of right-left awareness, 97 percent of normals and 62 percent of schizophrenics achieved all items correct. Among the schizophrenics, right-left awareness scores correlated posi- tively with chronological age. The range of IQ scores was so narrow that no positive correlation be- tween awareness and IQ could be expected to develop. The more in- formative test would have been correlation of mental age with right awareness scores. Wing (1971) also found problems with right-left orientation in 75 percent of speaking autistic children. Boucher (1977) showed slight but insignificant tendencies for sinis- trality in autistics and dextrality in their parents.

In summary, right or left domi- nance and right-left awareness de- velops later in psychotic children than in normals but at rates similar to those seen in mental retardates. In one study in which perfect scores were required as evidence for presence of dominance, schizo- phrenic boys remained inferior to their normal counterparts through age 12. Thus we do not know

whether psychotic children always remain behind normals in domi- nance and right-left awareness or whether they may catch up in later adolescence. Comparison of lateral dominance and awareness scores with mental ages would demon- strate more adequately whether incompetence in lateral dominance is related more to developmental level than to psychiatric diagnosis.

Electroencephalographic (EEG) Studies. The EEG has been used as a tool for studying the intactness of the central nervous system of autistic subjects.3 Like many such tools, the EEG must be used properly if results are to be trusted. Small (in press) assessed the technical adequacy and the findings of 13 studies reported in the 1960s and 1970s. Incidence of EEG abnormalities ranged from 10 to 83 percent, with an average of 52 percent. Only 3 of 13 studies met at least three of the four re- quirements for a technically ade- quate study, plus including control groups. In these three studies, au- tistics had as many or more EEG abnormalities than children with other psychiatric diagnoses. In all six studies in which a normal con- trol group was provided, autistics had significantly more EEG abnor- malities. Small concluded that "such data . . . have added signif- icantly to the growing body of evi- dence that infantile autism is an organic disorder characterized by severe central nervous system im- pairment." There is much disa-

3 Because the neurophysiology of autism has been reviewed in detail by James and Barry (1980) for the Schizo- phrenia Bulletin, only a brief review is given here.

greement about the kinds of EEG abnormalities and clinical correla- tions probably due "in part to technical difficulties in obtaining adequate records during waking, activation procedures, and sleep in extremely disturbed uncooperative children." The differences might also be explained as reflecting dif- fering amounts, types, and sites of brain dysfunction in various autis- tic children.

Cerebral Asymmetry, Pneumoen- cephalograms, and Computerized Tomographic (CT) Scans. Black- stock (1978) demonstrated that au- tistic children listened to music in preference to spoken material and generally preferred their left ears while normal controls had variable preferences. A second matched control group of nonautistic sub- normal children would have given added information about the spec- ificity of this phenomenon to au- tistic children. Blackstock believed his results supported the idea that autistics are "predominantly right hemisphere processors, a concept also supported in Prior's (1979) re- view."

Thorough neurological studies by Hauser, DeLong, and Rosman (1975) of 17 autistic children re- vealed consistent abnormalities only in the pneumoencephalo- graphic (PEG) examinations. Fif- teen cases showed "some enlarge- ment of the left lateral ventricle (with respect to the right) and, particularly, enlargement of the left temporal horn . . . [which] re- flected primarily flattening and at- rophy of the hippocampal con- tours . . ." (p. 683). A variety of other PEG abnormalities appeared in 10 cases. Degree of temporal horn abnormalities, however, was not associated with severity of Ian-

VOL 7, NO. 3, 1981 413

guage or behavioral symptoms, with presumed etiology, or with severity of intellectual deficit. The authors felt this finding supported medial temporal lobe disease being responsible for autistic symptoms. The lesion was thought to be ''asymmetrically bilateral'' but mainly located in the left hemi- sphere.

Using computerized tomogra- phy (CT) scans, Hier, Lemay, and Rosenberger (1978) found that 7 of 9 autistics had a right parieto- occipital region larger than the left in contrast to 11 of 41 nonautistic retardates (p < .02). This reversed pattern of asymmetry suggested to the authors a left hemisphere dys- function in autistics. Mild ventric- ular abnormalities such as abnor- mal bifrontal dimensions in 5 of 17 of an autistic group, reversal of ventricular asymmetry in 6 of 17, greater than normal asymmetry in 6 of 17, and "clipping" of a frontal horn in 4 of 17 were reported by Damasio et al. (1980). Three pa- tients had localizable lesions and five showed clear evidence of bi- lateral ventricular enlargement. There was no single abnormal pat- tern and those that were present were probably "consequent to a variety of (CNS) disease(s)."

Cerebral Spinal Fluid Studies. Cerebral spinal fluid (CSF) analy- ses by Young et al. (1977) for glu- cose, protein, cells, and folate re- vealed no abnormalities in 11 autistic children, 3.8 to 9.11 years old. The possibility that autism might be caused by a slow virus led these authors to measure CSF immunoglobulins. No abnormal levels were found, thus not sup- porting the slow virus hypothesis. As the authors pointed out, how- ever, additional studies are needed

since in many slow virus infec- tions, CSF immunoglobulins are generally normal.

Antigen Studies. To test the viral etiology theory of autism, Stubbs and Magenis (1980) typed 20 autis- tic children and their parents for histocompatibility blood antigens types HLA-A and HLA-B and compared them with antigens of 757 controls. While no significant differences were found, the au- thors felt the autistic group was too small to come to any conclu- sions and that other new measures of the immune response should be used.

Neuropathological Studies. Darby (1976) located 29 literature cases of autism and four unpublished cases, all with documentation of neuropathological changes. Cere- bral lipoidoses were found in 21 percent, tuberous sclerosis were found in 6 percent, and a variety of neuropathological findings occurred in other abnormal cases (54 percent). Darby believed that both degenerative child psychosis, formerly called Heller's disease, and other child psychoses "consti- tute the overt behavioral expres- sion of a number of disorders . . . analogous to the 'final common pathway' concept espoused by Bellak in 1958 for adult schizo- phrenics" (p. 49). In contrast, Hauser and Williams (1979) re- ported all three autistic cases stud- ied showed no postmortem evi- dence of neuropathologic changes.

Diagnosable Neurological Condi- tions. It often occurs that many children have complaints that are originally diagnosed as due to a psychiatric disorder only to be rediagnosed later as neurological.

Of 12 such cases presenting in a year's time to a hospital pediatric neurology service, two cases origi- nally called "child psychosis" had the following ultimate neurological diagnoses: (1) sex-linked diffuse cortical sclerosis with Addison's disease and (2) subacute sclerosing panencephalitis following measles (Rivinus, Jamison, and Graham 1975).

In a "neuro-epileptic group" of 34 children (ages 5 to 14 years) with overt neurological disorders "above the brain stem," Rutter (1977) reported an increased inci- dence of hyperkinesis and psycho- sis over nonbrain-damaged chil- dren with psychiatric disorder (17 vs. 2 percent). In this group, psy- chosis was found more often in low IQ than high IQ cases.

Coleman (1979) has tabulated 149 cases of infantile autism ac- companied by a diagnosable neurologic condition encom- passing metabolic, infectious, chromosomal, structural, and nonspecific types. The most com- mon individual illnesses of the 29 listed were phenylketonuria in 50 autistic cases and rubella in 29 cases, the latter clearly docu- mented by Chess (1977). Autism also has been found in histi- dinemia, celiac disease, tuberous sclerosis, purine disorder, toxo- plasmosis, congenital syphilis, en- cephalitis, infantile spasms, Cornelia deLange Syndrome, and cytomegalic inclusion disease, among others.

In the future, study of the neu- ropathology of phenylketonuria should reveal much about the con- nections among autistic behavioral symptoms, biochemistry, and ana- tomical sites of the brain lesions. Recent studies reveal that the buildup of phenylalanine is proba-

414 SCHIZOPHRENIA BULLETIN

bly toxic to the brain (Rosenberger and Nigam 1979). In view of the high incidence of autistic behavior in phenylketonuric subjects, the anatomical substrate of autism might be made clearer through biochemical-anatomical autopsy studies of the phenylketonuric brain. Comparison should be made between brains of phenylketonuric people with and without autistic symptoms. A sim- ilar neuropathological study of congenital rubella brains with and without autistic symptoms to lo- cate differences in the sites of le- sions also might be productive.

Biological Problems During Ges- tation, Birth, and Infancy. A high incidence of postmaturity (65 per- cent) in a Western Australian sample of autistics was reported by Lobascher, Kingerlee, and Gubbay (1970). Also there were more complications of labor and more frequent neonatal abnormalities over normal con- trols. The latter finding was con- firmed by DeMyer (1979) in the United States. Others reporting excess of perinatal risk factors were Bender and Faretra (1973), Pollack et al. (1970), and Ruttenberg (1970) while Bomberg, Szurek, and Etemad (1973) and Treffert (1970) did not. O'Moore (1972), in Dublin, Ireland, found a high incidence of pregnancies terminating unsuccessfully in mothers of autistics (8/25), some degree of fetal anoxia during par- turition (21/25), and forceps deliv- eries (13/25).

In a detailed study of a large sample (105) of autistics admitted to Bellevue Hospital in New York City, Campbell, Hardesty, and Burdock (1978) found bleeding to be the most frequent complication

of pregnancy (about 13 percent) which occurred more often with autistics than with their siblings. Using a weighted scoring system, the authors found that autistic children (n = 46) had more prena- tal and perinatal complications than their normal siblings (n = 20) but fewer than the siblings with minimal brain damage (n = 12).

DeMyer (1979) identified a pos- sible biological etiological factor in about 88 percent of cases in the first 3 years of life (troubled preg- nancy 48 percent, difficult labor 37 percent, first year illness followed by behavioral change 15 percent, second year illness 18 percent, and third year illness 6 percent). About a quarter of the 33 children had more than one possible cause. Chronic constipation or diarrhea in infancy or early childhood was more common in autistics (75 per- cent) than in normal controls (33 percent).

Implications of Neurobiological Findings. A wide variety of neurobiological studies revealed that autistic children in compari- son to normal children had more physical signs carrying the impli- cation that something had gone wrong or could have gone wrong with the central nervous system. Studies encompassed such tech- niques as standard and circum- scribed neurological evaluations, fingerprint patterns, hand domi- nance, right-left orientation, EEG, PEG, CT scans, neuropathology, early-life medical events, and analyses of body tissues and cere- bral spinal fluids. In contrast, au- tistics and retardates had nearly equal signs of neurological import. Nevertheless, an anatomical and functional explanation must be found as to why some

neurologically damaged children have the autistic syndrome and others do not. Logically we could assume that the lesion(s) in autis- tic children with normal perform- ance IQs may be more subtle than in lower IQ children. Attempts to localize the lesions within any pre- cise area of the brain of autistics have not been successful, although it does appear that the left side of the brain may be more heavily in- volved than the right side. Rutter's (1978) conclusion that the lesion "must be bilateral" is given some credence. Localization studies are in their infancy and so far nonautistic retardate controls have been included in only one CT scan study. Such controls are a requi- site for specifying neuroanatomical differences between autistics and retardates.

Those theorists who invoke per- ceptual inconstancy (see James and Barry 1980 for a review) as the basic neurological dysfunction have hypothesized that the lesion is located in the reticular activating system, vestibular and brainstem area. Data from most localizing studies suggest that theorists should begin to think about more widespread lesions in many cases that have affected not only think- ing and language but have in- volved vaso-vegetative centers as well. The perceptual inconstancy theory is further weakened by ob- servations that individual autistic children are quite consistent in the items they pass or fail on various developmental tests. Also while autistic children vary one from an- other on which neurological mo- dalities are most dysfunctional, each child is quite consistent over time in his own dysfunctional mo- dality (see especially the language studies of Churchill 1978).

VOL 7, NO. 3, 1981 415

One brain center that routinely seems fairly well spared in many autistic children is the motor sys- tem controlling large muscle activ- ity. Most autistic children walk ad- equately if not completely normally. (See section on Behavioral Characteristics.) Most retarded children whose verbal IQs are as low as those of most au- tistics (profound and severely re- tarded) might be found either not to walk at all or to begin walking after 3 or 4 years of age. The autis- tic child's motor abnormalities are more consistently found in activi- ties involving complex visual- motor modalities such as motor imitation and ball play. However, the better motor function in autis- tic children may itself be an artifact of diagnosis. If the motor system deficit is prominent, the child is likely to be labeled as cerebral palsy or profound retardation even though he has typically autistic behavioral features. Ordinarily such children are excluded from studies of autistic children.

To make further progress in locating the basic neurological dysfunction crucial for producing the autistic syndrome requires that investigators study multiple func- tions at the same time in popula- tions of well-described autistics and mental-age-matched normals and subnormals. Only then will we begin to sort out which of the many disabilities of the autistic child differentiate this category from the nonautistic neurologically handicapped child. The site of the lesion(s) will be better known as our noninvasive brain lesion localizing techniques improve and as our neuropathological studies take advantage of phenylketon- uric, rubella, and other neurolog- ical disease populations in which

the incidence of autism is relative- ly high.

Biochemistry

In the last 10 years, many investi- gators have tried to elucidate pos- sible biochemical etiologies of in- fantile autism. While various approaches have been used, including examination of metabo- lism of biogenic amines, amino acids, hormones, and trace ele- ments, no specific biochemical markers have been found. Efforts to find such markers have been confounded by the lack of diag- nostic specificity and by the nu- merous other factors such as age, sex, intellectual status, and activi- ty level that also affect the various body biochemicals. While it is pos- sible that biochemically distinct subgroups of psychotic children do exist, studies to date have not demonstrated them with any cer- tainty.

The serotonergic system has re- ceived the most widespread atten- tion by measurement of serum and platelet levels, platelet uptake and efflux studies, and assaying CSF and urine for serotonin and indoleamine metabolites. Studies of peripheral serotonin (5-HT) lev- els done by independent investi- gators suggest that 5-HT levels are higher than normal in autistic chil- dren. It is not possible to conclude that this increase is a specific path- ogenetic factor in autistic illnesses because of the high individual var- iability of levels within diagnostic categories as well as the observa- tion of elevated 5-HT levels in nonautistic individuals. The most compelling evidence supports the idea that blood serotonin levels are more strongly related to intellectu- al status than to psychiatric diag-

nosis. Higher levels of serotonin were found consistently in lower IQ groups. Thus it is not surpris- ing that in autistic children who customarily also have low IQs serotonin would be increased over groups of normal or mildly sub- normal children. This finding makes it especially important that IQ be controlled in all future stud- ies of the serotonergic system in autistic and control populations.

Serotonin was measured by Campbell et al. (1974) in small groups of emotionally disturbed and neurologically impaired chil- dren matched for age, sex, and verbal IQ. No significant differ- ence in group mean 5-HT level was found, but low IQ was related to elevations of serotonin. In a followup study by Campbell et al. (1975), childhood schizophrenics (n =23) and normal controls (n = 16) were compared. Serotonin lev- els were not different, age and 5-HT levels were not correlated, and neuroleptic drugs did not seem to alter 5-HT levels. Interest- ingly, comparison of low (IQ = 3-54, mean = 30.6, n = 16) to high (IQ = 70-135, mean = 89.9, n = 7) intellectual function yielded 5-HT levels of .324 ± .214 //g/ml and .139 ± .102A«g/ml, respectively (p < .05). A similar relationship of low intel- lectual function to elevated 5-HT was shown by Hanley, Stahl, and Freedman (1977) in a study of au- tistics (n = 27), severe retardates (n = 25), mild retardates (n = 23), and normals (n = 6). Mean 5-HT levels for the autistics and severe retardates were significantly higher than those of the mild retardates and normals. Neither autistics and severe retardates nor mild retardates and normals were significantly different in 5-HT lev- els. "Hyperserotonemia" was then

416 SCHIZOPHRENIA BULLETIN

defined as a 5-HT level more than 1.67 SD above the mean for the mildly retarded group. Using this definition, 30 percent of autistics, 52 percent of severe retardates, 9 percent of mild retardates, and 0 percent of normals were defined as "hyperserotonemic." Another study correlating low intellectual function to alterations in 5-HT was contributed by Sankar (1977). He reported that autistics and retardates with psychosis had low- er platelet 5-HT uptake than nonautistic schizophrenics and children with behavior disorders. Low IQ was correlated with low platelet 5-HT uptake in children with various diagnoses.

Ritvo et al. (1970) measured 5-HT and platelets in autistics and normal controls and found an in- verse relationship between age and these variables. The group mean 5-HT level for autistics was significantly higher than that of age-matched controls, but the 5-HT platelet ratio was not signifi- cantly different. In a related study, Yuwiler et al. (1971) looked for circadian rhythmicity in 5-HT and platelets in autistics and nonautistics. The groups did not differ significantly with respect to 5-HT and platelets and no rhythmicity was demonstrated.

Several studies have been done on the nature of the relationship of platelets to serotonin metabolism because platelets are thought to be a model of serotonergic neurons in the central nervous system. Boullin, Coleman, and O'Brien (1970) reported that autistics (n = 6) had (1) significantly higher platelet counts (4.38 ± 0.36 x 10 vs. 3.51 ± 0.22 x 10*, p < 0.05); (2) platelets that accumulated serotonin to a slightly greater de- gree; and (3) platelets with a de-

creased ability to retain 5-HT (as evidenced by a two-fold increase in efflux) when compared to nor- mals (n = 6). In a followup study, Boullin et al. (1971) studied dis- turbed children (n = 10) and nor- mals (n = 8). Using a three-fold increase in the normal rate of 5-HT efflux as a criterion for autism, they found a biochemical correlate that agreed with the Rimland E-2 score prediction for autism in 9 of 10 cases. The discordant case was a child rated as autistic by the Rimland score but who was not observed to have an elevated rate of efflux.

Groups of autistics, hospitalized comparison patients, and normal controls were studied by Yuwiler et al. (1975). While autistics were found to have significantly higher whole blood 5-HT levels and 5-HT platelet ratios when compared to normals, no significant differences in uptake or efflux of 5-HT were shown among the three groups.

Higher platelet serotonin levels were confirmed by Takahashi, Kanai, and Miyamoto (1976) in a study of autistics (n = 30), nonautistics (n = 45), and normal controls (n = 30). Autistics had higher levels than controls (980 ± 357 vs. 807 ± 202 ng 5-HT/mg pro- tein, p < 0.025). Nonautistics were not significantly different than ei- ther of the other groups. Hyperactivity was correlated to in- creased platelet 5-HT in the nonautistic group. Eleven of 45 nonautistics were rated as hyperactive, and this subgroup had a higher 5-HT level compared to the remainder of the group (1,012 ± 420 vs. 745 ± 288 ng 5-HT/mg protein, p < 0.05). They suggested that hyperactivity might be a factor in elevated serotonin

levels seen in some disturbed chil- dren.

Studies of platelet monoamine oxidase (MAO) have generally agreed that there is no significant difference in the activity of this en- zyme in disturbed children (Belmaker, Hattab, and Ebstein 1978; Boullin et al. 1975; Campbell et al. 1976a; Cohen, Young, and Roth 1977; Lake, Ziegler, and Mur- phy 1977; Roth, Young, and Cohen 1976; Takahashi, Kanai, and Miyamoto 1977). Roth, Young, and Cohen (1976) did find that females had higher MAO activity than males (29.96 ± 1.34, n = 42 vs. 22.05 ± 1.39, n = 37, nmoles/mg protein/hour, p < 0.001). Also, they noted that among females, younger individuals had higher MAO activity.

Another subject of interest in the biochemistry of autistic chil- dren has been the catecholamines. Young et al. (1978) studied urinary free catecholamines in normal (n = 9) and autistic (n = 5) boys and found autistics had markedly low- er levels of urinary amines (10.8 ± 2.3 vs. 32.6 ± 18.6 /ug/24 hours, p < 0.01). In a study of platelet uptake and efflux of dopamine, Boullin and O'Brien (1972) found no dif- ference between normals and con- trols.

The research history of dopamine-/3-hydroxylase (DBH), a synthesizing catecholamine en- zyme catalyzing the conversion of dopamine to norepinephrine, epit- omizes the need for multiple bio- logical controls if biochemical re- search is to be meaningful. Coleman et al. (1974) found DBH to be slightly but nonsignificantly lower in autistic children than in normal controls, but these findings were reversed in a later report (Goldstein 1976). Lake, Ziegler,

VOL. 7, NO. 3, 1981 417

and Murphy (1977) measured plas- ma norepinepherine (NE) and dopamine-^-hydroxylase (DBH) ac- tivity in autistics (n = 11), family members of autistics (n = 39), and controls (n = 12). Significantly higher NE levels were observed in autistics vs normals both in basal (462 ± 42 vs. 233 ± 42 pg/ml, p < .001) and standing (676 ± 41 vs. 480 ± 70 pg/ml, p < .025) posi- tions. Significantly lower levels of DBH activity were observed both in autistics and their family mem- bers when compared to controls. Plasma DBH was found by Belmaker, Hattab, and Ebstein (1978) to be elevated in children with functional psychosis when compared to those with organic psychosis. No differences were ob- served in catechol-O-methyl- transferase activity.

In 1980 Young et al. found that 22 autistic children did not have DBH levels different from other psychiatric patients or normal con- trols. However, there was increasing DBH activity as chrono- logical age increased among the controls but not in the autistics. The authors felt this finding might be due to the smaller age range of autistics, or alternatively, might be related to a decrease in urinary free catecholamine and MHPG (Young et al. 1978). They also found a strong genetic effect of DBH activity.

Levels of biogenic amine metab- olites in CSF have been studied as a result of the suspected relation- ships of these levels to the turn- over rates of the parent com- pounds in the central nervous system. Cohen et al. (1974) meas- ured CSF levels of homovanillic acid (HVA), and 5-hydroxyindo- leacetic acid (5-HIAA) in psychotic children, epileptics, and children

with movement disorders. Eleva- ted levels of these metabolites were seen in psychotics when compared to epileptics. The rela- tionship between levels of metabo- lites and probenecid suggested the need to measure 5-HIAA, HVA, and probenecid simultaneously. In a related study of children with autism, nonautistic psychosis, central processing disturbances, and aphasia and pediatric controls, Cohen et al. (1977) found autistics had lower CSF 5-HIAA levels than nonautistic psychotics.

Studies of abnormal metabolites of biogenic amines in childhood psychoses have been prompted by observations of possible abnor- malities in adult schizophrenics. Widelitz and Feldman (1969) stud- ied childhood schizophrenics for the presence of "pink spot" in urine. The incidence of "pink spot" was nearly equal in normals and schizophrenics but ". . . the intensity of the 'spot' was greater in schizophrenics" (p. 29). The qualitative assay for "pink spot" has not been useful as a marker for childhood psychoses.

Bufotenin and other N,N- dimethyl derivatives of 5-HT have been measured in psychotic chil- dren. Himwich et al. (1972) found that some autistics excreted these compounds while no normals were observed to do so. Narasimhachari and Himwich (1975) found that some psychotic children excreted bufotenin. Moreover, they ob- served that many first-degree rela- tives of the index cases excreted bufotenin despite being behaviorally normal. The inci- dence of positive bufotenin excre- tion in relatives of patients was much higher than that of the gen- eral population, but the implica-

tions of this observation remain unclear.

Investigations of other biochem- ical parameters also have not identified specific markers for childhood psychoses. Johnson, Wiersema, and Kraft (1974) found no difference in hair amino acids between autistics and normals. Jackson and Garrod (1978) found no differences in plasma zinc, cop- per, and amino acids in autistics. Gittelman and Cleeman (1969, 1971) have studied serum magne- sium in psychotic children and found no differences when com- pared to normals. Perry, Hansen, and Christie (1978) found a signifi- cant difference in the level of CSF ethanolamine between controls and psychotics, but it is not certain that elevated ethanolamine is unique to psychotic illness.

Several studies examining the endocrine system of psychiatri- cally disturbed children have been done. Brambilla, Viani, and Rossotti (1969) studied the adre- nal, pituitary, and thyroid status of 16 children with psychotic ill- nesses, and found the following: (1) Corticotropin hyposecretion was seen in 62 percent of all pa- tients; (2) gonadotropin hypose- cretion was seen in 75 percent; and (3) thyrotropin hyposecretion was seen in 50 percent. In general, children with more debilitating psychoses had higher rates of en- docrinopathy. Of the patients studied, 87 percent had demon- strable endocrinopathies thought to be related to problems in the pi- tuitary or hypothalamus and not in the peripheral target organs, al- though this result was not well documented.

Yamazaki et al. (1975) studied the stress response and circadian rhythmicity of 11-hydroxycortico-

418 SCHIZOPHRENIA BULLETIN

steroids in autistics. They found the stress response to be preserved, but the development of normal circadian rhythmicity was dis- turbed in autistics, suggesting an abnormality in the development of the cerebrohypothalamic system. Mahler et al. (1975) studied an in- sulin stress response and found controls recovered normal glucose levels more quickly, autistics had more prolonged cortisol elevation in response to stress, and no dif- ference in growth hormone levels was seen between the groups. From these results, the authors suggested a functional abnormality in the endocrine stress-coping mechanism in autistics.

Diurnal cortisol and temperature variations were measured in autis- tics and normals by Hill et al. (1977). Normal children (n = 3) were similar to adults in cortisol and temperature variations, while four of six autistics showed abnormalities. One autistic child on antipsychotic medication had diurnal rhythms similar to adults.

Free fatty acid (FFA) metabolism was studied in autistic, schizo- phrenic, and emotionally dis- turbed children by DeMyer et al. (1971b). The following were ob- served: (1) Plasma FFA levels were more variable in psychotics than other groups; (2) 66.6 percent of autistics had normal lowering of FFA in response to exogenous in- sulin, which was contrary to re- ports on adult schizophrenics; (3) no difference was observed in the FFA response to endogenous insu- lin as tested by glucose stimula- tion. A poorly functioning regulatory feedback mechanism was suggested to explain these data.

Reviews of biochemical studies in psychiatrically disturbed chil-

dren have been done by Guthrie and Wyatt (1975), Cohen and Young (1977), and Ritvo (1977). They found that cross-comparison of biochemical and endocrine ex- periments was difficult because of differences in diagnostic categories considered and experimental methodology. In addition, most experimenters have used rather small groups of patients because of the difficulty of any one center's accumulating a relatively large au- tistic population. The aggregate of results strongly implies that fac- tors such as age, sex, activity level, intellectual function, organic neu- rological or medical disease, treat- ment modalities, physical environ- ment, and severity of illness need to be taken into account as any of these variables may cause altera- tions in the parameters studied. Higher than normal serotonin levels are found to be more strong- ly related to low intelligence than to the diagnosis of infantile au- tism. A collaborative effort controlling for these variables in larger populations would be able to answer better the question of which, if any, of the many inter- esting biochemical deviations re- ported in autistic children are spe- cific to the syndrome.

Pharmacotherapy

Drug treatment of autistic children has proved difficult because of un- certain efficacy of drugs used and numerous side effects. Controlled drug studies for the most part have involved small, nonhomo- geneous populations, making in- terpretation of the results difficult. Agents studied have included antipsychotics, biogenic amine precursors, stimulants, psyche- delics, antidepressants, vitamins,

thyroid analogs, and lithium. While these agents have been as- sociated with improvement in some individuals, no group of drugs is clearly superior in treating the complex symptoms of this syn- drome, and improvement may be seen in some symptoms while oth- ers are aggravated.

Antipsychotics have been asso- ciated with improvement in some children. Rock (1974) reported that trifluoperazine, when used early, decreased autism, improved the ability to relate to adults and other children, decreased hyperactivity, improved the ability to perceive surroundings, and arrested the de- velopment of secondary retarda- tion as measured by WISC scores. Further, serious side effects were not observed with trifluoperazine.

Faretra, Dooher, and Dowling (1970) found haloperidol and fluphenazine to be equally effec- tive in decreasing anxiety and au- tism with little change noted in assaultiveness and regressive be- havior after 8 weeks of treatment. Haloperidol exerted its effects somewhat more rapidly than did fluphenazine. Engelhardt et al. (1973) found haloperidol and fluphenazine to be equally effec- tive in improving problems with increased psychomotor activity, stereotypic behavior, responsive- ness, sleep disturbance, and feed- ing. They did observe a higher in- cidence of extrapyramidal symptoms with fluphenazine than haloperidol.

A study by Rimland4 using par- ent replies to questionnaires

4 Information from July 1980 bulle- tin of the Institute for Child Behavior Research, San Diego, California; Bernard Rimland, director.

VOL. 7, NO. 3, 1981 419

showed that of 15 psychopharma- cologic agents, haloperidol was perceived by parents as resulting in some improvement in the greatest percentage (58 percent) of cases in which drug therapy was tried. In comparison, all drugs av- eraged were perceived as helpful in 28 percent of cases, psychother- apy in 34 percent, high-dose vita- mins in 67 percent, day school in 70 percent, and operant condi- tioning in 82 percent.

Claghorn (1972) compared haloperidol and thioridazine in an outpatient group and found both drugs to be beneficial. Haloperidol was better in controlling behavi- oral symptoms and more rapid in onset of action. Haloperidol and behavior therapy were studied by Campbell et al. (1978a). Haloperidol was clearly superior to placebo in controlling withdrawal and stereotypic behavior. It was also found to increase appropriate orienting reactions (Cohen et al. 1980). Concurrent behavior thera- py and haloperidol seemed to pro- duce good results, although it was not clear if the drug facilitated learning in a direct or indirect fashion.

Thiothixene and trifluperidol were found to be significantly bet- ter than chlorpromazine in less- ening withdrawal, excitability, psychotic speech, and stereotypic behavior by Campbell et al. (1970). The sedative effects of chlorpro- mazine were suggested as the cause of the observed differences. Waizer et al. (1972) reported thiothixene to be effective in controlling hyperactivity, stereo- typic behavior, and sleep distur- bances in an outpatient popula- tion. All patients studied showed some improvement with a low in- cidence of untoward side effects.

Trifluperidol, chlorpromazine, chloral hydrate, and amphetamine were compared to placebo by Campbell et al. (1972c). They found trifluperidol to be the only agent to be significantly better than placebo in controlling social, language, affective, and motor symptoms.

Molindone, an indole derivative with antipsychotic and stimulating properties, was studied by Camp- bell et al. (1971a). Eight of 10 pa- tients treated with molindone showed improvement as evi- denced by decreased withdrawal and negativism and increased af- fective responsiveness. In the studies reviewed, antipsychotics were effective, while sedating agents were associated with in- creased withdrawal and apathy.

Side effects of antipsychotics are an important consideration when they are used in pediatric popula- tions. Antipsychotic withdrawal was studied in 34 psychotic chil- dren who had been on drugs 6-15 months by Polizos et al. (1973). Drugs studied were fluphenazine, haloperidol, thioridazine, trifluoperazine, and thiothixene. Clinical relapse was seen in 32 pa- tients within 2 weeks of termina- tion. Involuntary movements and ataxia involving the extremities and trunk were observed in 42 percent of all cases after with- drawal.

In a retrospective study of 125 children, McAndrew, Case, and Treffert (1972) reported weight gain, tardive dyskinesia, impaired learning, and ocular changes as significant complications of antipsychotic treatment. Using pa- tients as their own controls to as- sess the need for further drug treatment was suggested as a means of reducing cumulative

dose, which is thought to affect the rate of side effect occurrence.

The biogenic amine precursors, dihydroxyphenylalanine (Lrdopa) and 5-hydroxytryptophan (5-HTP), have been shown to have little effect on psychotic children. Zarcone et al. (1973) reported in- creased rapid eye movement sleep and no behavioral change in two schizophrenic boys treated with 5-HTP. Sverd et al. (1978) also found no significant change in be- havior with 5-HTP treatment. Ritvo et al. (1971b) found no sig- nificant alteration in symptomatol- ogy despite significant lowering of serum 5-HT levels with chronic L-dopa administration. L-Dopa and amphetamine were compared in a crossover study by Campbell et al. (1976b). Amphetamine had only transient beneficial effects which were outweighed by wors- ening of preexisting symptoms. LrDopa treatment resulted in stim- ulation of motor behavior, lan- guage production, and affective responsiveness in 4 of 11 subjects studied.

Although megavitamin therapy has been suggested as a possible treatment for schizophrenia, few studies involving children have been done. Niacinamide was eval- uated by Greenbaum (1970) and found to have no significant thera- peutic effect after 6 months of treatment. Rimland, Callaway, and Dreyfus (1978) studied vita- min B6 (pyridoxine) in a popula- tion of children who had previous- ly responded favorably to B6 treatment and reported "behavior . . . deteriorating significantly during B6 withdrawal" (p. 472). The authors discussed several problems with the experimental design, problems which raise

420 SCHIZOPHRENIA BULLETIN

questions about the validity of the observations.

Trials of psychedelic drugs have been limited. In a review of seven studies involving 91 patients, Mogar and Aldrich (1969) sug- gested further studies based on observation of increased speech production, affective responsive- ness, and decreased ritualistic be- haviors in patients treated with psychedelics. Simmons, Benor, and Daniel (1972) reported in- creased affective response and contact with adults and decreased responsiveness to external stimuli and motor activity with LSD-25. These changes were of short dura- tion, and the effects of chronic LSD administration are unknown.

Treatment with thyroid hor- mone initially was promising, but later results failed to uphold these findings. Campbell et al. (1972a) compared triiodothyronine (T3) and amphetamine and found T3 to have stimulating and antipsychotic effects leading to significant im- provement in a heterogeneous population. T3 was found to be ef- fective in patients classified as euthyroid, leading the investiga- tors to suggest a greater "need" for thyroid hormone for adequate CNS function in this population. Campbell et al. (1973) reported similar results with liothyronine treatment. As a followup, Camp- bell et al. (197Sd) conducted a placebo-controlled crossover study of T3 in 30 autistics and found no significant difference between T3 and placebo.

Abassi, Linscheid, and Coleman (1978) measured T3, T4, and TSH in 13 autistics and found no evi- dence for hypothyroidism. Two patients developed clinical hyper- thyroidism with administration of exogenous T3. The hypothesis of

thyroid dysfunction in autistics was not supported.

Results after trials with other drugs have generally been unrewarding. In a study by Campbell et al. (1971b) imipramine was not useful generally for treating psychotic children, al- though further study of its appli- cation in retarded, mute, anergic, and borderline psychotic children was suggested. Results of lithium and chlorpromazine administra- tion suggested lithium may be use- ful in treatment of aggression, ex- plosive affect, and hyperactivity (Campbell et al. 1972b).

In summary, of the drugs stud- ied, nonsedating antipsychotics seemed to offer the most promise, whereas stimulants were usually found to cause greater disorgani- zation. Mixed results have been obtained with vitamins, thyroid hormone, and psychedelics. Biogenic amine precursors have not been found to be therapeutic. Antidepressants and lithium may be useful in treatment of specific symptoms; when any of these agents are used, the benefits must be carefully weighed against po- tential side effects, especially in the pediatric population with its potential for prolonged exposure.

Behavioral/Educational Treatment Programs

During the last 10 years, the most significant development in the treatment of infantile autism has been the continued growth in the use of behavioral/educational ap- proaches. This development has been due to the generally accepted position that traditional forms of psychotherapy have not been ef- fective in reducing the severe behavioral deficits associated with

this syndrome. Following the first direct use of behavior modification methods with early childhood psy- chotics, as reported by Ferster and DeMyer in 1961, over 100 papers in this area had appeared by 1970. The next decade has seen the pub- lication of over 200 more papers using various modifications of these techniques. Although behavioral methods are by no means a cure for infantile autism, there is a consensus that the use of systematic-intrusive approaches can lead to significant expansions in the behavioral repertoires of the children, allowing them to reach a higher level of functioning than if they went untreated or were treated with traditional therapeutic procedures. In this section some of the more extensive studies report- ing the use of behavioral/educa- tional programs with autistic chil- dren will be reviewed.

Verbal Behavior. One of the most important applications of behavioral methods has been to develop language skills in autistic children, since severe deficits in verbal behavior are often regarded as the major characteristic of this syndrome. Lovaas and co-workers (1973) have continued their early work in developing verbal skills with autistic children. After an ini- tial period of reinforcing verbal imitation, they developed more spontaneous speech using a shaping process involving labeling objects (at first with prompts which are gradually faded) and then using simple statements about objects in an appropriate context. While this procedure is ef- fective with either autistic children who are completely mute or with those exhibiting some spontane- ous speech, therapeutic progress is

VOL. 7, NO. 3, 1981 421

more rapid and achieves a higher level of functional speech with children who begin the treatment with some verbal skills. Of 20 chil- dren treated, all showed gains in verbal behavior, with spontaneous language being developed in the children with some speech when treatment was initiated. However, elementary vocabularies and object identification characterized the older and mute autistic children. One- to 4-year followups on these children indicated that those who were institutionalized tended to regress in verbal behavior as well as in all other behaviors while the children receiving further behavior therapy from their parents contin- ued to show improvement. A brief reinstatement of behavioral meth- ods temporarily reestablished some of the original verbal gains in the institutionalized children. These findings point out the im- portance of intense and continued work with the individual autistic child if behavioral procedures are to be effective (Lovaas, Schreib- man, and Koegel 1974). Much ground is lost rather quickly once the reinforcement contingencies and/or the concentrated training procedures are no longer in force. It also suggests that spontaneous developmental processes neces- sary for language acquisition are not being triggered or stimulated (if, indeed, they are even poten- tially available in these children) by behavioral methods in the case of most autistic children (see the Language section).

Many other workers have had success with expanding verbal skills in autistic children which is similar to that reported by Lovaas' group. Halpern (1970), for exam- ple, studied 15 children over a 4-year speech treatment period.

When the children started the pro- gram, nine were completely nonverbal and six showed some minimal verbal output. By the end of the treatment, 13 of the children demonstrated some useful speech to the degree that 11 of these were subsequently enrolled in special public school classes. Using prompting and fading procedures, Marshall and Hegrenes (1970) worked with four previously nonverbal autistic children and successfully developed object name and description skills. Work- ing with children possessing some basic verbal skills, other research- ers have been able to develop more complex language behaviors, such as simple and compound sen- tence usage (Stevens-Long and Rasmussen 1974), prepositional usage (Sailor and Taman 1972), and simple reading comprehen- sion (Rosenbaum and Breiling 1976). In the Sailor and Taman study, it was found that the use of nonambiguous stimulus condi- tions was far superior as a training procedure, thus demonstrating the continued need for concrete refer- ences in language even for autistic children with some verbal skills. Also, short verbal commands (one to four words) were more effective in eliciting correct responses from autistic children than longer com- mands (Browning 1974).

In spite of the substantial gains in verbal output which the behavioral methods yield, it is ob- vious that only in a small number of cases have these procedures led to the development of true lan- guage in autistic children. Even Lovaas7 group no longer expresses the optimism regarding long-range goals for speech development that they once did (Lovaas et al. 1977). As more and more researchers

come to the conclusion that perceptual/language disorders play a critical role in the autistic syn- drome, it appears that more realis- tic goals for verbal behavior in these children are being defined. However, any speech, even if it is a primitive rote speech, is better than complete mutism, and behavioral methods appear to be some of the most effective for working with the mute autistic child. The autistic child already possessing some verbal skills ap- parently can benefit from a num- ber of intensive behavioral/edu- cational programs and has some chance to develop a more func- tionally useful language than his mute autistic counterpart.

Special Behavioral/Educational Approaches. The one-to-one therapist-child relationship is usu- ally considered the ideal, if not the essential, situation if significant progress is to be achieved in work- ing with autistic children. Howev- er, there is some evidence that a 1:1 ratio is not always superior to a group approach (Frankel 1976; Frankel and Graham 1976). In ad- dition, it is not practical or finan- cially feasible to maintain this schedule for the many hours per day, and for the many years, re- quired to expand the behaviors of autistic children. Thus, a number of researchers have attempted to adapt behavioral/educational methods for classroom use over long-term treatment programs. Koegel and Rincover (1974), start- ing with a 1:1 student-teacher ra- tio, gradually phased in larger classroom ratios and found that multiple baseline behaviors gener- ally increased for both verbal and nonverbal (e.g., attending to

422 SCHIZOPHRENIA BULLETIN

teacher upon command) re- sponding.

In a subsequent study (Rincover and Koegel 1977) the effectiveness of unsupervised responding in a classroom setting as a variable influencing behavioral progress was investigated. Initially each child in a four-child class was giv- en individual instruction on a rotating basis. Typically one child stopped responding when the teacher moved on to the next child. However, when prompt fad- ing, chaining, and programmed instructional material were intro- duced to teach each student to continue working for an extended period of time after the individual instruction, significant academic progress was observed. By the end of the training sessions one teacher was able to teach (various tasks related to tracing and draw- ing letters) all four children simul- taneously with each child learning at his own rate with minimal su- pervision.

Although substantial gains can be achieved with intensive behavioral/educational treatment, learning deficits will likely contin- ue to interfere with the acquisition of new responses, thus requiring strict adherence to structured home or institutional programs to maintain treatment gains (Brown- ing 1971). The enormity of the task of training these children should be faced realistically. In a 1-year study of 14 autistic children receiving programmed instruction with a student-teacher ratio of 10:1, 11 children completed only an average of 10 percent of the course material with three children making much faster progress (Fischer and Glanville 1970). If the slower children progressed at the same rate for successive years of

schooling, the authors estimated that it would take about 10 years before they were ready to enter a normal school program. However, it should be pointed out that other workers (e.g., Koegel, Egel, and Dunlap 1980) take a more optimis- tic view and stress the fact that some progress can be made even with the slowest children.

In a comprehensive study of 50 autistic children treated in three different special educational treat- ment settings with a 3V2- to 4-year followup (Bartak and Rutter 1971, 1973; Rutter and Bartak 1973), all children made considerable prog- ress in educational, cognitive, lin- guistic, social, and behavioral sta- tus sufficient to justify the effort and expense involved. The most important finding appeared to be that in comparing the three types of treatment settings (general re- gressive therapeutic approach with little attention to development of skills vs. permissive classroom en- vironment combined with regres- sive therapeutic methods vs. struc- tured classroom training emphasizing specific perceptual, cognitive, and motor skills), the latter experience was the most ef- fective in fostering the greatest scholastic progress, as well as yielding as much progress in gen- eral social and behavioral develop- ment as the other two methods. The advantage of organized edu- cational programs in combination with intrusive behavioral methods was also supported by Rutter and Sussenwein (1971), Schopler et al. (1971), Graziano (1970), and Ward (1972, 1978). They can effectively be taught to teachers and a variety of nonprofessionals, especially parents (Schopler and Reichler 1971; Kozloff 1973; Marcus et al. 1978), enabling more "therapist"-

child interactions in varied set- tings, one factor which seems criti- cal for significant progress to be attained and maintained (Nordquist and Wahler 1973; Gallagher and Wiegerink 1976; Schopler 1976; Zifferblatt et al. 1977; Singh 1978).

Related to this, the importance of systematic teacher education should not be overlooked. Empir- ical assessment of teacher training in behavior therapy is possible. Not surprisingly, systematic im- provement in the autistic child's behavior was found to be related to the training level of the child worker (Koegel, Russo, and Rincover 1977). In addition, the importance of the teacher-child re- lationship should not be underestimated. Although auto- mated instruction may be of some aid in training autistic children, at least one study has indicated that teaching machines are effective with autistic children only when the teacher is present, whereas the teacher did not require the ma- chine to teach the task successfully (matching-to-sample) to the child (Russo, Koegel, and Lovaas 1978).

In spite of their emphasis on mastery of skills, the structured behavioral approaches have not penalized the autistic child in gen- eral behavioral-social develop- ment, when compared to unstructured methods. One study did fail to find differences in thera- peutic effectiveness among various approaches: behavior modifica- tion, educational programs, rela- tionship therapy, and activity therapy (Wenar and Ruttenberg 1976). However, the weight of evi- dence from the greatest number of investigations indicates that the method of choice for maximal behavioral expansion in autistic

VOL. 7, NO. 3, 1981 423

children is a systematic behavioral/educational program involving as many child contact hours as possible and using thera- pists who have been trained in the use of behavioral methods.

Other Aspects of Behavioral/Edu- cational Programs. In the last 10 years some additional studies have been published describing impor- tant variables that may influence the learning that should occur dur- ing behavioral/educational pro- grams with autistic children. These include research on time-out stimuli, responding to multiple cues, maintenance and generaliza- tion of training, sensory stimula- tion as reinforcers, and the reduc- tion of self-stimulatory behavior.

Time-out procedures might nor- mally be expected to serve as pow- erful aversive stimuli for autistic children. On the other hand, if they provide the opportunity for preferred, nonproductive self- stimulatory responses, they might actually increase the undesired re- sponses, rather than reduce them. This appeared to be the case with tantrums; however, when self- stimulation was prevented during time-out, these methods did re- duce the tantruming (Solnick, Rincover, and Peterson 1977). Nu- merous other studies have ques- tioned the routine use of time-out as a training procedure with autis- tic and/or retarded children. Gen- erally, they have demonstrated that it is of extremely limited train- ing value with this population and is a poor use of valuable training time (Husted, Hall, and Agin 1971; Martin 1975; Plummer, Baer, and LeBlanc 1977).

The inability to respond to mul- tiple cues has been suggested as one of the underlying deficits in

autistic behavior (see the Percep- tual Processes section). Koegel and Schreibman (1977) demonstrated that four autistic children were able to learn to discriminate a multiple-cue complex from each of its two component cues. However, the autistic children did not learn the discrimination in the same manner as normal children, re- sponding at a higher level to one of the cues in the early trials. Only after a long series of successive discrimination problems were they able to develop a set for approach- ing new discriminations by re- sponding equally on the basis of both component cues.

Maintenance and generalization of the learned behavior are crucial considerations for any training program involving autistic chil- dren. Desired behaviors that are emitted only in the restricted training setting during the actual training period are of limited prac- tical value compared to those that are taken up as part of the general repertoire of the child. Thus, the delineation of variables that influ- ence generalization and/or mainte- nance of the newly acquired re- sponses is essential for improving the success rate of training pro- grams. In two experiments (Rincover and Koegel 1975; Koegel and Rincover 1977) it was clearly demonstrated that training which failed to generalize could be made to generalize when the antecedent stimuli that were functional during training were identified and intro- duced into the extratherapy set- ting. Even when generalization occurred, however, maintenance was not usually noted. The dual use of partial reinforcement sched- ules in the original treatment envi- ronment, and noncontingent rein- forcers in the extratherapy

environment significantly in- creased generalization. The au- thors concluded that there were two distinct parameters of extratherapy responding: generali- zation and maintenance, both requiring separate assessment for developing a behavioral program that can be carried out effectively outside of the actual therapy ses- sions. This may also be the case with regard to developing appro- priate social behaviors in autistic children, since these responses are resistant to generalization (Romanczyk et al. 1975; Strain, Kerr, and Ragland 1979).

Since food reinforcers, typically used in training autistic children, have the disadvantage of a low sa- tiation threshold, other reinforcers are always being sought out. Sen- sory stimulation (music, a visual flickering stimulation, or a visual movement) was successfully used with four autistic children to main- tain high rates of responding (bar pressing on an FR 5 schedule) over a long period of time (Rincover et al. 1977). The reinforcement strength of the sensory events varied from child to child, and when a child appeared to satiate for a particular sensory event, a relatively small change in the event was sufficient to produce a high rate of responding once again. Similar results were found with strobe light stimulation used as a reinforcer (Frankel et al. 1976). Of special importance is the recent quantitative evidence that success- ful completion of tasks by autistic children can itself increase motiva- tion to continue to respond effec- tively to those tasks (Koegel and Egel 1979).

The frequency of excessive self- stimulatory behavior in the autistic child can interfere with efficient

424 SCHIZOPHRENIA BULLETIN

training of more normal re- sponding (Lovaas, Litrownik, and Mann 1971). These repetitive re- sponses are apparently affected by some types of environmental stim- ulation and may be related to the mental age of the autistic child (Colman et al. 1976; Frankel et al. 1978). Methods that can eliminate or reduce these behaviors are of great importance. It has been shown that simple removal of the sensory reinforcement of various self-stimulation responses in three autistic children was effective in extinguishing the behavior (Rincover 1978a)—for example, removing auditory feedback of plate spinning by installing carpet- ing on the table where the child would normally spin plates.

An overcorrection method was used to eliminate self-stimulatory hand clapping in an autistic 7-year-old boy (Foxx and Azrin 1973). The procedures involved 5-minute periods of functional movement training contingent upon hand clapping. The child was instructed and manually as- sisted in moving his hands to one of five different positions. Eventu- ally the training required only a verbal cue. Within a few days hand clapping was reduced to near zero levels, thus permitting more appropriate and functional use of the hands. In general, the authors found overcorrection methods to be more effective than alternative procedures (e.g., physical punish- ment) in eliminating self- stimulatory behavior. Another group (Wells, Forehand, and Hick- ey 1977; Wells et al. 1977) also used overcorrection as a method for decreasing self-stimulatory re- sponding (e.g., mouthing and twirling objects, flapping and shaking hands) in autistic chil-

dren. In addition, they often found increased spontaneous toy play accompanied significant re- ductions in self-stimulatory re- sponding. This latter observation was supported by Harris and Wolchik (1979) and had been earlier reported (Koegel et al. 1974) under conditions in which physi- cal suppression of self-stimulation was used to reduce the inappro- priate behaviors. In contrast to the usual attempts to reduce self- stimulation, Hung (1978) showed that the opportunity to engage in self-stimulatory behavior, judi- ciously used, could serve as a rein- forcer for spontaneous sentences of two autistic boys.

Self-stimulatory behavior may take the form of self-injurious be- havior in some autistic children. In addition to the obvious danger to the child that this responding presents, it is extremely disruptive to training sessions. Although some researchers clearly feel that self-mutilation is reinforceable, just as any other operant (Lovaas, Schreibman, and Koegel 1974), others hypothesize that self- injurious responding is simultane- ously maintained by both intrinsic and extrinsic reinforcement (Carr 1977). High demand situations (such as usually found in training sessions) also could lead to in- creased rates of self-mutilation (Carr, Newsom, and Binkoff 1976). Frankel and Simmons (1976) pro- pose both operant and respondent paradigms as the best model to deal with data from both human and animal studies of self-mutila- tion.

Lovaas' group has been predom- inant in the use of response con- tingent electric shock to suppress the self-abusive responses. Con- sidering the risks to the child from

injuries that they can self-inflict, Lovaas feels that this danger justifies the use of drastic meas- ures such as shock. In fact, he feels that it is one of the most ef- fective and reliable of all the behavioral methods used with au- tistic children (Lovaas, Schreib- man, and Koegel 1974). Although most therapists hesitate at using such techniques, especially in view of possible unforeseen side effects, a recent review of studies em- ploying electric shock with autistic children actually reports more pos- itive side effects (response gener- alization, increases in social be- havior, and positive emotional behavior) than negative side ef- fects as the result of its use (Lichstein and Schreibman 1976). While not ruling out a role for punishment in suppressing self- injurious responding, Frankel and Simmons (1976) felt that not enough emphasis has been placed on programs that train adaptive and appropriate behaviors which are strengthened by social rein- forcement.

The widespread use of behavioral/educational methods with autistic children in the last 10 years has led to significant im- provements in the overall behavioral repertoire of almost all the children given advantage of these programs (Treffert, McAndrew, and Dreifuerst 1973; Margolies 1977; Egel, Koegel, and Schreibman 1980; Koegel, Egel, and Dunlap 1980). Except in the case of those children who already have a more extensive behavioral background, however, these ther- apeutic advances usually do not result in a change in the overall poor prognostic outlook for most autistic children (Prior 1979). In spite of this fact, significant ad-

VOL. 7, NO. 3, 1981 425

Table 4. Percent distribution

Outcome variable

Overall outcome Good and very good Fair Poor and very poor

Speech Normal Useful (communicative) Develop useful speech after

age 5 Worsening None

Education Age appropriate (within 2

years) Regular school (2 or more

years) Other variables

Gainful employment Loss of autism Long-term institutionalization

n of autistic (psychotic) subjects Mean chronological age at

followup (years)

of autistic

DeMyer et al. (1973)

10 16 74

6 49

11 11 32

10

70

0 3 5

42 126

12

(psychotic)

Rutter and Lockyer (1967)

14 25 61

16 46

11 11

2

2 4

14 44 63

15

population

Eisenberg (1956)

5 22 73

51

54 63

15

by study1

Kanner (1971)

18 27 55

9

18

54 11

35

Lotter (1974)

14 24 62

0

90

4

48 29

17

Etemad and Szurek (1973)

19 17 64

47

19

1 Sources which detail the diagnostic criteria and procedures are identified in the References. 2 Less than 50 percent. 3Over 15 years of age. 4 Over 16 years of age. 5Twenty to 75 percent depending on diagnostic group.

vantages associated with the behavioral/educational methods remain. They provide maximal op- portunities to those children with the greatest potential, while those children with the least potential receive at least some important behavioral gains.

Prognosis

The outcome for most autistic peo- ple is a life of complete or semidependence. Three studies (Rutter and Lockyer 1967; Eisenberg 1956; and DeMyer et al.

1973) showed fairly close agree- ment about the poor chance of the autistic child to become completely normal. Only 1 or 2 percent achieve such status. In table 4 (adapted from DeMyer et al. 1973) are the findings of six outcome studies of the mid-1950s to the mid-1970s that lend themselves to meaningful comparison. About 5 to 19 percent of autistics reached the borderline of normality, while 16 to 27 percent had a fair out- come, and 60 to 75 percent had a poor or very poor outcome.

Rees and Taylor (1975) found

that 36 percent of 54 mixed psy- chotics had a "successful out- come"; that is, were within 1 year of appropriate school grade or gainfully employed. Of Etemad and Szurek's (1973) sample of 84 mixed psychotic children, only about 29 percent at age 19 years were not hospitalized and were semi-independent. Eggers (1978) from Germany reported a 20 per- cent remission rate in 57 childhood schizophrenia cases. All patients with onset before age 10 years had a poor outcome. Bender (1973) in- dicated 33/100 child schizophrenics

426 SCHIZOPHRENIA BULLETIN

(25-50 years old) made a satisfac- tory community adjustment. Thus childhood psychosis developing at or near puberty may have a some- what better prognosis than earlier onset cases.

The most potent factor associ- ated with a relatively good out- come is a relatively good general ability as measured by IQ tests on initial evaluation. DeMyer et al. (1973) found the best predictors of outcome in children initially exam- ined at mean age 5V2 years were (listed in rank order) IQ, severity of illness, social rating, speech, and the number of neurological signs. Lotter (1974) found speech and IQ to be the best predictors (r = .89), with others being severity of illness, delay in passing mile- stones, sex, neurological findings, and years of school. In his series, no autistic girl had a good or fair outcome; however, one of the two normal outcome cases in the sample of DeMyer et al. was a girl. From Lotter7s reviews of other studies, he concluded that rela- tively more male autistics may have a good outcome than female autistics. Other authors reporting IQ tests and verbal skills as good predictive items were Campbell et al. (1978c) and Rees and Taylor (1975). Those important factors re- ported not to have prognostic im- plications were family history of neuropsychiatric disorder, family atmosphere (Eggers 1978), and ad- equacy of parental functioning (DeMyer et al. 1973).

If an autistic child has an initial IQ over 60 or 70, his chances for a relatively good outcome are better than for lower IQ children. In the followup sample of Lockyer and Rutter (1970), 42 percent of higher IQ autistics were working or still in school when reexamined 10 years later. All of DeMyer's better

outcome autistics had initial IQs over 60. However, not all higher IQ children will be successful—a finding Lotter (1974) attributed to lack of social intelligence. Adult autistics tend to be impractical and lacking in ability to understand and respond to feelings of others. Some demonstrate lack of emo- tional control and others a lack of ambition (DeMyer 1979).

While overall outcome is not good for independent living, many autistic children improve sympto-

matically. Table 5 adapted from DeMyer et al. (1973) shows changes in percentage distribution toward the more able rating alternatives for three important as- pects of behavior, namely conver- sational speech, social relation- ships, and work/school rating from mean age 5Vi years to 12 years. Unfortunately very few of the au- tistics were given normal or nearly normal ratings at followup and some autistic children (about 10 percent) lost ground. Those with

Evaluation period Initial % o f

subjects

0 5.9

21.0 28.6 44.5

Followup % o f

subjects

3.7 26.2 15.9 15.0 39.2

119 107

Table 5. Changes in percentage distribution of speech, social, and work/school ratings among autistic children over a mean 6-year followup1

Conversational speech rating alternatives

1. Normal for age 2. Expresses ideas but below chronological age 3. Expresses immediate needs + echolalia 4. Echolalia only or few words 5. Mute

Total n

Social behavior 1. Nonpsychotic

a. Normal b. Immature/hostile/shy

2. Psychotic c. Loner d. Severe withdrawal e. Oblivious

Total n

Work/school rating 1. Normal 2. Borderline 3. Educable retarded 4. Trainable retarded 5. Subtrainable retarded

Total n

0 5

21.1 49.1 25.4

114

0 2.5

12.4 30.6 54.5

8.1 29.3

38.4 26.3

6.0

99

6.6 3.3

16.5 24.0 49.6

121 121

1 Adapted from DeMyer (1973).

VOL 7, NO. 3, 1981 427

extreme hyperirritability including self-injurious behavior may re- gress from a previously higher functioning level (DeMyer 1979). Many parents whose autistic chil- dren remain difficult to care for during adolescence and show min- imal improvement in intelligence and self-care become worn out psychologically and physically, the burn-out syndrome. In one large followup study, 59 percent of au- tistics over 14 years old were residing in chronic care institu- tions (DeMyer et al. 1973).

Theories of Etiology and Pathogenesis At the beginning of the 1970s, Hingtgen and Bryson (1972) identified three basic types of the- ories related to infantile autism: nonorganic, organic-experiential, and experiential. Most nonorganic theories then, as now, were psy- chodynamic, assumed the infant was normal at birth, and attributed the development of symptoms to defective ''nurture" from the par- ents, mainly the mother (see table 6 for examples). Many versions of deficient parental nurture have been proposed, but these can be grouped into two chief types. In the first variety, the parents of au- tistics were considered to have greater amounts of psychopathol- ogy than other parent groups. In the second one, they were viewed as having extreme personality types such as being cold, full of rage, or without sense of self.

In either version the infant was viewed as basically normal biolog- ically, but normal identification was blocked or the parents failed to give optimum warmth and stim- ulation. We use the word "infant" to emphasize that most investiga- tors thought the seeds of infantile autism were sown during infancy

Table 6. Examples of nurture theories of infantile autism

Fraknoi and Rutten- berg (1971)

Szurek (1973)

Williams and Harper (1973)

Ekstein and Friedman (1974)

King (1975)

Massie (1978)

Inadequate stimulation from mother leads to unre- lieved tension and then to a state of frozen bal- ance between aggressive and libidinal drives

Ungratified infant impulses become repressed, then fused with anxiety/rage, press for dis- charge, meet with more parental repression/ tension and become "insatiable longings of a secondary narcissism of a dissociated human character"

Sensory deprivation at critical developmental periods

Mother has a lack of "sending power"

Double-bind attitude (superficial warmth combined with severe coldness and rejection) of the mother makes the infant "want to escape"

Active avoidance of eye-to-eye contact and defi- cient touching of the infant by the mother.

and not childhood. While some theorists have not been explicit as to how the parent's personal defi- ciencies were translated into poor child care practices, others have blamed poor feeding or poor speech practices, for example.

Any parental deficiency, what- ever its specific qualities, theoreti- cally led to the infant's social with- drawal which in turn led to failure to acquire normal speech and

other intellectual and social skills. Hence, the "nurture" theorists be- lieved that normal biological intel- ligence was locked within the au- tistic child. If the right treatment key could be found, the child would accelerate in progress and eventually become normal or even supernormal in intelligence.

There were also two varieties of the "nature-nurture" interaction theory (table 7). In one version all

Table 7. Examples of nature-nurture theories of infantile autism

Despert (1971)

O'Moore(1972)

Miller (1974)

Tinbergen and Tin- bergen (1976)

Disturbance processing "l-other" concept leading to resistance in developing social contacts

Autism tends to develop in language-impaired children, most likely caused by fetal anoxia plus psychological trauma

Evidence for brain abnormality inconclusive. Pos- sibly interference in early, subtle organization of newborn behavior leads to a perceptual disorder due to inadequate sensorimotor, tactile, and kinesthetic handling

Causes organic and experiential. Some cases caused more by organic factor while others have more psychogenic causes

428 SCHIZOPHRENIA BULLETIN

autistic infants were seen as bio- logically deficient and the parents viewed as failing to give proper emotional support to a "vulnera- ble" infant. In the second version, some infants were viewed as "or- ganically" damaged and others as biologically normal. In this second version, the parents of "nonorganic" infants were de- scribed as failing to a greater de- gree than parents of "organic" in- fants.

In the pure "nature" or organic theory (table 8), the parents were viewed as contributing little more to the child's illness than symptom variations. The illness was consid- ered basically the expression of a biological abnormality. As in the "nature-nurture" interaction theory, investigators disagreed about the kind of biological dys- function that might be present in the infant.

Types of Consistencies in Theo- ries. Through all the variations of parental deficiency theories ran a consistent thread of emphasis on the parental failure to adequately and warmly stimulate their autistic infants. In some theories, this ade- quate stimulation failure was im- plied, for some theorists dwelt on the greater degrees of psychopath- ology in the parents and did not state explicitly the mechanism of the developmental failure of the autistic child. Logically, such de- velopmental failure could come from inadequate handling prac- tices making life uncomfortable and incomprehensible to the in- fant. Alternatively, the infant could mimic the sick personality of the parents.

In the many theories concerning the kind of biological defect, there

was rather consistent emphasis on a deficient nervous system. It is in- teresting to note that only a few authors considered the deficiency to be the rather common neurolog- ical dysfunction such as might af- fect the infant's general intelli- gence, as well as specifically the language centers. Most often, some more esoteric and difficult to test biological deficiency was pro- pounded concerning the reticular activating system, the limbic lobe, or the whole timing mechanism of somatic development, for exam- ple.

Clearly, to test these three opposing theories and even a few of their variations, researchers needed to consider both the psy- chological nature of the parents and their child-rearing practices and the neurological status of the children. The reason it was logical to begin with such a homely bio- logical procedure as a neurological evaluation was the proclivity for a maldevelopment or insult to any portion of the nervous system to be accompanied by other physical maldevelopments. These signs, small as they might be, such as de- formed earlobes or cafe-au-lait skin spots, would be the outward tracks that accompanied the bio- logical dysfunction in the brain. These kinds of outward tracks are common in other groups of neurologically dysfunctioning chil- dren such as the generally re- tarded (Kennard 1960). Such a neurological study, in addition to giving us indirect evidence of CNS dysfunction, or of normal func- tion, might also give clues as to the nature of the insult in any given child or subgroups of chil- dren. It also might help define neurologically normal and abnor- mal subgroups of autistic children.

Types of Control Groups Needed. Two sets of control groups were necessary to test the three theo- ries: (1) parents whose children were all normal and one of their normal children, and (2) parents who had at least one nonpsychotic child, subnormal in general intelli- gence or having a specific learning disability such as dysphasia. The control parents would need to be matched to autistic parents on im- portant demographic variables. The children should be age- and sex-matched. All the subnormal children should have use of all four extremities and special senses so that the same developmental tests could be used on all child groups.

Findings That Would Prove or Disprove Theories. It is possible to outline a series of logical con- structs of the findings from the 1970s that would prove or dis- prove each one of the opposing three theories (see table 9, parts A and B). If the pure nurture theory were correct, investigators should find that parents of autistic chil- dren in comparison to control par- ents mishandled their autistic in- fants or provided more inappropriate identification models. Unhealthy personality types, mental illness, or extreme character traits should be more common in parents of autistics. Because most autistic children have severe symptoms before their third birthday, this parental defi- ciency would have had to occur before this time. Investigators should find no more signs of neu- rological dysfunction in autistic than in normal children.

If the pure nature theory were

VOL. 7, NO. 3, 1981 429

Table 8. Examples of nature theories of infantile autism

Huttand Hutt (1970) Frith (1971), Hermelin

(1971) Money, Bobrow, and

Clarke (1971) Myklebust, Killen, and

Bannochie (1972) Bender (1973)

Zarcone et al. (1973)

Baltaxe and Simmons (1975)

DeMyer (19755)

Hauser, DeLong, and Rosman (1975)

Hertzig and Walker (1975)

Simon (1975)

Boucher(1976a) Cohen, Caparulo, and

Shaywitz (1976) Darby (1976) Omitzand Ritvo (1976)

Porges (1976) Chess (1977)

Fish (1977)

Churchill (1978)

Damascio and Maurer (1978)

Hieretal. (1978) Rutter(1978)

Coleman (1979) Lovaas, Koegel, and

Schreibman (1979)

Hyperactive reticular activating system leading to reactive effort to reduce sensory input Imbalance between extracting external rules and internal rules interfering with

organism-environment interchange Autoantibodies to CNS

Dysfunction of right hemisphere with improper nonverbal processing

Early decompensation in genetically vulnerable individual due to organic stress in utero, perinatally, or early childhood

Low 5-HT in rapid eye movement sleep center (brainstem) results in a mixed dream- wake state

May be left hemisphere. Underlying deficit not language specific but cognitive

Profound language difficulty may be sole cause in some children but not in others who may also have a visual motor dyspraxia leading to the characteristic verbal and non- verbal communication problems

Asymmetrically bilateral temporal lobe disease but main location left hemisphere

Brain organization deficiency leading to disorder in information processing

Autistic is dysprosodic and cannot get meaning of intonation and stress features of speech. Location may be inferior colliculus of brainstem, a major center for selective attention to sounds. Variety of causes (birth anoxia, circulatory toxins, or phenyl- ketonuria)

Language impairment is symptom of general cognitive difficulties Midbrain and brainstem dysfunction involving catechojaminergic pathways but need

more evidence Cause is due to a number of diseases; symptoms due to a "final common pathway" Neuropathophysiological process affecting developmental rate, sensorimotor integra-

tion, language, cognition, intelligence, and ability to relate. Cause either idiopathic or known brain disease of various kinds

Imbalance in autonomic activities; possibly abnormal serotonin levels in CNS Final behavioral consequence from many different causes and can follow chronic viral

encephalitis. Onset not always before age 30 months Pandevelopmental retardation is an "infancy marker" for an inherited neurointegrative

defect Language deficits are necessary and sufficient cause of those phenomena common to

all autistics. Other types of disabilities may account for some of the differences among them

A variety of agents cause lesions in mesolimbic cortex in mesial frontal and temporal lobes, neostriatum, and anterior and medial nuclear groups of thalamus

No single abnormal brain location. Lesion(s) consequent to a variety of disease Nonspecific syndrome of biological impairment is most likely explanation. Bilateral le-

sion Not unitary disease. If localized lesion, it is in thalamus Stimulus overselectivity: Low performance level of autistic children related to their tend-

ency to respond to only limited number of cues in their environment

430 SCHIZOPHRENIA BULLETIN

Table 8. Examples of nature theories of infantile autism—Continued

Prior (1979)

Wing (1979)

Left hemisphere dysfunction with some language and cognitive functions possibly taken over by "relatively" strong right hemisphere mediation

Deficit in cognitive skills

correct, we should find that autis- tic children differed in some im- portant biological way from nor- mal children and possibly also from the subnormal children. There should be no important de- ficiency in the way autistic parents handled their autistic infants in

comparison to parents of subnor- mal infants.

If the nature-nurture theory were correct, the autistic and sub- normal children should be similar in neurological status, but parents of the latter group should provide more adequate infant care prac-

tices than parents of autistics. The alternate nature-nurture interac- tion theory was that there were two classes of autistic children—the neurologically nor- mal and the abnormal. In this case, we should find that those autistic children who were normal

Table 9. Results needed to support each of three basic theories of causation1

For nurture theory to be correct

For nature theory to be correct For nature-nuture interaction theory to be correct

Children 1. Autistic children should

have no more signs of neurological dysfunc- tion than normal chil- dren but fewer signs than subnormal chil- dren

Parents 1. Greater amounts of

psychopathology or

2. More deficient childcare practices (during infancy espe- cially) in parents of au- tistics than parents of normals or subnormals

1. Autistic children should have more signs of neuropathology than normal children

2. They could or could not resemble subnormal children in overt neuropathology but should have some dif- ferent modalities af- fected than subnormal children

1. Equal amounts of psychopathology

or 2. Child-care practices

should be insignifi- cantly different in au- tistic vs. normal and autistic vs. subnormal parents

Version A . Autistic children should have equal amounts and types of neurologi- cal dysfunction as sub- normal children

. Normal children should have lesser amounts of neurological dysfunc- tion than either autistic or subnormal children

Greater amounts of psychopathology

or More deficient childcare practices (especially during in- fancy) than subnormal parents who should be at least as good or better than normal par- ents

Version B Parent type a should

have a neurologically normal autistic child. Parent type b should have a neurologically abnormal child

1. There should be 2 classes of parents of autistics with respect to amounts of psychopathology or adequacy of parenting:

a. The emotionally ill or deficient parent

b. The emotionally healthy or suffi- cient parent

1 Adapted from DeMyer (1975/?).

VOL. 7, NO. 3, 1981 431

neurologically would have had less adequate parenting than those who had signs of neurological dys- function.

Major Import of Findings From the 1970s. Little support has been found for either the nurture or the nature-nurture interaction theo- ries. Parents of autistic children were discovered to be similar in parenting practices and emotional health to parents of both normal and biologically deficient control children. Only a higher incidence of reactive depression from the stress of rearing a deviant child was found in both mothers of au- tistics and retardates in compari- son to normal control groups. These findings have become so firmly established that nurture and nature-nurture interactional theory articles were infrequently published during the 1970s.

Evidence for the nature theory is strong. The results of most care- fully designed studies using the two necessary control groups have shown that more autistic children than normal controls have come from troubled pregnancies and dif- ficult labors, and have had more signs of neurological dysfunction at all periods of life. In contrast, nearly every study showed how much autistic children were like retardates and other neurobiologically handicapped groups with regard to incidence of neurological dysfunction and measured intelligence. Even in those very few autistic children who possess overtly a clean neuro- logical bill of health and "normal" intelligence, some theorists (Cohen, Caparulo, and Shaywitz 1976) are speculating that new di- agnostic techniques are likely to

locate heretofore hidden brain def- icits.

Yet to be explained is what makes one brain dysfunctional child autistic and another one re- tarded and yet another one special learning disabled. Investigators have yet to locate the exact func- tional and anatomical or biochemi- cal differences between autistics and other neurologically handicapped groups. The brain is generally assumed to be the site of the lesion. The cause of the lesion is assumed to be any of the many agents that are known to damage the brain. A sign of the times is that defective biogenic amine me- tabolism is involved as a possible causative agent by some theorists. Currently some investigators (Ornitz and Ritvo 1976) speak of autism as classifiable on the basis of etiology as (1) idiopathic (due to unknown cause) and (2) associated with diagnosable disease. The ana- tomic site(s) offered as possibilities cover many areas of the central nervous system from the limbic lobe to the higher frontal and pari- etal cortical association centers (see summary of the Neurobiology section).

The question of the pathogenetic route of the language symptoms has intrigued several theorists. Some state or imply that pathology in the brain centers controlling language might be the core lesion while others believe that language symptoms primarily result from cognitive deficiencies. Any theo- rist who grapples with this contro- versy must also contend with the fact that neither "language" nor "cognition" has a good opera- tional definition (Churchill 1978). Furthermore, the brain structures subserving all aspects of

language, let alone cognition, have not been defined. Until such defi- nitions are forthcoming and brain localization studies make enor- mous strides, these two con- tending theories cannot be put to a good test. The perceptual incon- stancy theorists have not dis- cussed in any detail the mecha- nism by which language symp- toms develop, but they generally locate the lesion in the midbrain or brains tern.

The connections between the so- cial symptoms and cognitive and language symptoms have been discussed relatively less often by organic theorists, although social symptoms have usually been re- garded as secondary to language/body communication deficits or to cognitive or percep- tual organization deficits. In Kanner's original theory, he saw the social deficits as inborn, a pro- vocative idea that may apply to some but not all autistics.

As Hingtgen and Bryson noted (1972), there are still basic defi- ciencies in the theories, the most notable being the proclivity for theorists to reason from an incom- plete data base. Either they seize upon partial aspects of the syn- drome or they base their ideas on limited observations of a few children—the most noteworthy example being that of Tinbergen (1974) in his Nobel Prize accept- ance speech. He largely took the side of the "nurture" theory based on observation of three autistic children and an unknown number of normal children. The errors in his theory and Tinbergen's igno- rance of important facts concern- ing autism were knowledgeably rebutted by Schopler (1974) and Wing and Ricks (1976).

As we begin the 1980s, all but a

432 SCHIZOPHRENIA BULLETIN

few researchers hypothesize that autism is an organic disorder involving the brain. While many organic theorists believe that any event that can damage the brain can cause the symptomatic picture of autism, few people agree on the site or sites of the lesion(s). A key controversy is whether language or cognitive defects are the central functional deficit. One major goal for the 1980s is to begin more ex- tensive investigations comparing nonautistic brain dysfunctional children to autistic children. Data from these studies should help to shed light on important etiological factors that lead to major differ- ences in social and communication differences between the two groups.

Current Status

In our earlier review (Hingtgen and Bryson 1972) two dominant hypotheses were said to have in- fluenced much of the research up to that time: (1) that the childhood psychoses, including infantile au- tism, were thought to represent the earliest manifestations of adult schizophrenia, and (2) that psy- chotic children were believed to be potentially capable of normal func- tioning in virtually all areas of de- velopment. Neither hypothesis has been supported by the re- search data, and during the decade of the 1970s it was the rare investi- gator who even gave lip-service to such previously held notions. The great body of evidence generated during the last 10 years of research has indicated that infantile autism is a type of developmental disor- der accompanied by severe and, to a large extent, permanent intellectual/behavioral deficits. It is

viewed as a distinct entity with clearly no relationship to the adult psychoses. This profound change in attitude has received the strongest endorsement in two prestigious national publications. The leading journal in this area changed its name in 1978 from Journal of Autism and Childhood Schizophrenia to Journal of Autism and Developmental Disorders; a n d , of perhaps even greater significance, the new DSM-IH (American Psy- chiatric Association 1980) no long- er includes infantile autism under the heading of childhood schizo- phrenia, but rather under the cate- gory of pervasive developmental disorders.

Still unresolved, however, is the relationship between infantile au- tism, other forms of developmen- tal disorders, and mental retarda- tion. Although the current research singles out degree of so- cial distance and degree of com- municative speech as the two most consistent differentiating items, the optimal rating instrument, with appropriate controls for both mental and chronological age, re- mains to be constructed. In addi- tion, of the many classification systems introduced in the last 10 years, no one system clearly stands out as having superior heu- ristic value regarding etiological, prognostic, or treatment consider- ations. Perhaps the multi-axial systems still provide the greatest hope for future work.

The prevalence rate of infantile autism is probably between 4 and 5 per 10,000 population under 15 years of age, while in subnormal populations such as those with congenital rubella and mental re- tardation, it is probably much higher: 500 to 800 per 10,000. There is nearly universal agree-

ment that more males have autism than females, the ratio being about 4 to 1, and that autism is not relat- ed to birth order. There is less agreement about maternal age and national origin of parents. Twin and other sibling studies lend weak support to genetic inherit- ance of the complete autistic syn- drome, although specific deficits in language-cognitive components may be genetically linked. The previous common finding of higher socioeconomic status being related to autism was probably due to referral bias. Higher socio- economic fathers are more likely than lower class fathers to apply to clinics specializing in autism where most studies on autism are done.

In sharp contrast to early portrayals of parents of autistic children as "refrigerator" person- alities, the last decade of investi- gation has characterized these par- ents as similar to those with children exhibiting other severe childhood disturbances. Thus par- ents of autistic children have been found to display no more signs of mental or emotional illness than parents of children with organic disorders (with or without psycho- sis). In addition, they do not mani- fest extreme personality traits such as coldness, obsessiveness, social anxiety, or rage, nor do they pos- sess specific deficits in infant and child care.

Although rearing an autistic child is stressful, and mild de- pressive symptoms are not uncom- mon, especially in mothers, there is no evidence that parents of au- tistic children develop severe per- sonality changes or major psychi- atric symptomatology. While there appear to be conflicting data about the presence or absence of thought

VOL. 7, NO. 3, 1981 433

disorders resulting from anxious confusion about their autistic child, parents of autistic children are now generally perceived to be trustworthy in their evaluation of their child's problems and are typ- ically involved as co-therapists in various treatment programs.

A major thrust of the 1970s was the growth of parent groups under the aegis of the National Society for Autistic Children. Parents meet together for emotional support, sharing of home care techniques and diagnostic/treatment/school resources, and to support re- search. The Society has done much to relieve parents of un- warranted guilt feelings and focus national attention on the realities of autism.

The studies of measured intelli- gence of the 1970s confirmed the findings of the 1960s that most au- tistic children score in the mentally retarded ranges, that these scores are stable over short and long peri- ods of time, and that splinter skills do not reduce the validity of the IQ, which is a good predictor of ultimate outcome. In addition, the conditions of social improvement or several years of treatment in au- tistic children are not, as a rule, accompanied by an increase in measured intelligence. An IQ of 40 or below, even when determined during preschool years, is predic- tive of a poor outcome. While not all children with IQs initially above 60 or 70 have a good out- come, those who make accelerated intellectual progress and a better social adjustment come from this higher IQ group. Autistic children in both preschool and later years are testable if appropriate mental age test items are used. As a group, autistics score nearer age norms on certain types of per-

ceptual-motor tests items, namely, fitting and assembly tasks (e.g., Seguin form board and WISC ob- ject assembly) than they do on ver- bal items, particularly verbal com- prehension and abstraction. This inability to use language abstractly on IQ tests parallels the disability in other language studies.

Current thinking assumes that gross disturbances in language de- velopment are a major characteris- tic, if not the primary defect, of in- fantile autism. If autistic children are not completely mute, as is fre- quently the case, their receptive and expressive language is primi- tive, echolalia occurs with high frequency, there is poor compre- hension of nonverbal gestures, and they have low levels of spon- taneous speech and speech for so- cial purposes. Cognitive/language deficits appear fundamental to the autistic syndrome. Indeed, vocal articulation appears as the sole area in which the autistic children with some language skills do not score lower than control groups from retarded or dysphasic popu- lations. Some initial studies using sign language provided hope that nonvocal communicative skills could be developed in autistic chil- dren. Unfortunately, sign lan- guage as used in more carefully controlled later studies failed to yield dramatic gains in language skills for most autistic children. However, these techniques could be effectively incorporated into an overall multisensory treatment program to increase the total behavioral (both verbal and nonverbal) repertoire significantly.

Important studies during the decade of the 1970s have success- fully demolished two widely held clinical impressions—namely, that autistic children have atypical lan-

guage because of parents' re- sponse to the child's attempts to communicate, and that autistic children refuse to use the pronoun "I." Neither hypothesis is taken seriously any longer by the great majority of researchers. Extensive use of behavioral procedures has produced significant improve- ments in speech in many autistic children. Nevertheless, in spite of many promising new treatment methods, the overall poor progno- sis for development of true spon- taneous, communicative language in most autistic children remains.

Since perceptual disturbances have been singled out as a possible underlying deficit responsible for the development of the autistic syndrome, research in perceptual processes appears crucial for ulti- mately understanding this disor- der. The phenomenon of stimulus overselectivity (i.e., in a specific learning task autistic children tend to respond only to a few cues from a larger range of available cues) has received a great deal of atten- tion since 1970. Although it is now apparent that stimulus over- selectivity, also having been ob- served in nonautistic retarded chil- dren, is closely related to lowered mental age, such a perceptual strategy could account for much of the paucity of behavior associated with infantile autism. A number of training methods have been de- vised to reduce the autistic child's overselective attention to cues for specific learning tasks.

Another interesting feature of learning in the autistic child is that some types of prompt fading tech- niques, which are usually quite ef- fective with retarded children, often impede, rather than facili- tate, learning simple and complex discrimination tasks. Other prob-

434 SCHIZOPHRENIA BULLETIN

lems in perceptual processing manifested by autistic children in- volve the imposition of idiosyn- cratic perceptual patterns on in- coming sets of stimuli, severe disturbances in handling multiple and complex cues, specific abnor- malities in audition, cross-modal stimulus association deficits, and failure to tolerate intertrial inter- vals of more than a few seconds. Although previously assumed to be an important aspect of the au- tistic syndrome, evidence of lack of eye-to-eye contact and lack of self-recognition is now seriously questioned. The major impact of research on perceptual processes in autistic children points to seri- ous disturbances, but the underly- ing mechanism(s) have yet to be delineated.

In behavioral studies, autistic children demonstrated levels of play (typically stereotyped or nonsymbolic), pantomine, hand use, body imitation, drawing geo- metric figures, and ball play that were typical of much younger chil- dren, even infants. Those autistic children with higher achievement levels in other test scores and in school also had higher levels of performance in all behaviors stud- ied. Parental reports of develop- mental milestones from autistic in- fancy through later childhood confirmed what the researchers measured: that autistic children were delayed in passing many milestones not only in language and motor skills but in use of eat- ing utensils, drawing figures, dancing, and rhythmic and music performance. Autistic children's inability to symbolize and imitate body action may be more severe than mentally retarded nonautistic children's and may relate to their inability to communicate nonver-

bally and thus in turn to their so- cial distance.

With the growing conviction during the 1960s that many, if not all, autistic children had neuro- biological abnormalities, investiga- tors of the 1970s used every feasi- ble technique to study this subject—for example, standard neurological evaluations, EEG, pneumoencephalograms, and computerized tomography. Autis- tic children in comparison to nor- mal children had more physical and laboratory signs of possible brain dysfunctions. In contrast, autistic and mentally retarded chil- dren had nearly equal numbers of signs of neuropathological import. Attempts to localize the lesion(s) are just beginning, but initial stud- ies indicated that both sides of the brain might be affected, the left side probably more severely in- volved than the right side. In many cases, however, a variety of brain abnormalities were reported in various sites, suggesting that the autistic syndrome could be re- lated to many different kinds of neurobiological insults.

The possibility that abnormal neurochemistry is responsible for the autistic picture has been heavi- ly investigated by chemical analy- ses of the various body fluids and their cellular constitutents. The serotonergic system received the most research attention. Several investigators working independ- ently found higher than normal levels of peripheral serotonin in groups of autistic children; howev- er, this abnormality may have little to do with producing primary symptoms of autism but more to do with mental retardation aspects of autism. Autistics with low IQs resembled nonautistic persons

with similarly low IQs in having higher than normal peripheral serotonin values. Also autistics with borderline or normal IQs had serotonin values within the normal range. Studies of other biogenic amines, amino acids, hormones, and trace elements have revealed no consistent biochemical abnor- malities specific to autism. Never- theless some provocative differ- ences between autistics and other diagnostic groups have been re- ported in urinary catecholamines and bufotenin; plasma norepi- nephrine, dopamine- /3-hydroxylase activity; CSF levels of homovanillic acid, 5-hydroxyindolacetic acid, and ethanolamine; serotonin efflux from red cells; and the endocrine system. These differences need to be confirmed and require further exploration in experiments careful- ly controlled for other biological factors.

While drug treatment is only partially successful in controlling some autistic symptoms in some individuals, the nonsedating antipsychotic agents (haloperidol: Haldol; fluphenazine: Prolixin; thioridazine: Mellaril; thiothixine: Navane; and molindone) appear more successful than the sedating agents (chlorpromazine: Thora- zine; and promazine: Sparine). Other less successful agents stud- ied include biogenic amine precur- sors, amphetamines, and thyroid hormones. Antidepressants and lithium may have limited useful- ness for certain symptoms. Effica- cy of high-dose vitamins is contro- versial, and well-controlled studies need to be done. Since side effects are common and can be se- rious and long lasting, the potent antipsychotic agents should be used mainly for treatment of se- vere symptoms uncontrolled by

VOL 7, NO. 3, 1981 435

behavioral techniques and for rela- tively limited periods of time.

Since traditional forms of psy- chotherapy have not proved effective with autistic children, systematic, intrusive, behavior- al/educational treatment programs were introduced in the early 1960s. This trend has increased in inten- sity in the 1970s and virtually dis- placed most other treatment ap- proaches used for the disorder. Application of behavioral methods to verbal behavior has been exten- sively studied both in mute autistic children and in those who had some language skills before treat- ment began. The latter group ap- pears to benefit most from these methods, often developing func- tionally useful language, if not true spontaneity of speech. In the former group of mute children, considerably less improvement is realized. Even here, however, ob- ject identification and elemental vocabularies are significant ad- vances over the complete absence of verbal communication. Inten- sive and continued work during treatment and followup are cru- cial, especially for children pos- sessing the lowest levels of verbal skills, if treatment gains are to be maintained.

Since one of the most important goals for most treatment programs is to prepare the autistic child to enter, at the very least, a special educational setting, a highly valu- able recent finding is the fact that a 1:1 therapist-child ratio is no long- er regarded as essential for the most effective treatment approach. Equivalent gains have been ob- tained in structured small classes. In addition, many studies have now indicated that greater prog- ress can be achieved in the educa- tional, cognitive, and linguistic

status of autistic children, when treated in an educational setting with intrusive behavioral methods than when treated with nonin- trusive approaches. Of equal im- portance is the fact that the former methods have in no way penalized the child with regard to general behavioral-social development. Thus, the overwhelming evidence suggests strongly that the treat- ment of choice for maximal expan- sion of the autistic child's behavioral repertoire is a system- atic behavioral/educational pro- gram, involving as many child contact hours as possible, and using therapists (including par- ents) who have been trained in the use of behavioral techniques.

Numerous improvements in specific behavioral techniques have resulted from intensive stud- ies of the critical variables influencing learning in autistic children. Many studies of factors such as time-out stimuli, re- sponding to multiple cues, mainte- nance and generalization of train- ing, sensory stimulation as reinforcers, and reduction of self- stimulatory behavior have resulted in the fine-tuning of behavioral methodology and the achievement of additional therapeutic gains. Al- though the rapid expansion in the use of behavioral/educational methods in the last 10 years has led to significant improvements in behavioral output in almost all children given the advantage of such programs, the overall prog- nostic outlook for a self-supportive adult life remains poor, except for those children already demon- strating key behavioral skills be- fore treatment. In spite of this rather pessimistic ultimate expec- tation, the advantage of behavioral treatment programs remains im-

pressive. They provide maximal opportunities for those autistic children with the greatest poten- tial, and, at the very least, furnish those children with the least po- tential the opportunity to acquire some significant behavioral skills.

A review of the major outcome studies of the last three decades shows that the majority (60 to 75 percent) of autistic people live in complete or semidependence with relatives or in long-term institu- tions. Only about 1 or 2 percent achieve a normal, independent status and 5 to 19 percent a bor- derline normal status. The factor most heavily associated with a good outcome is a relatively good general ability as measured during initial evaluation (e.g., IQ, verbal skills). Those who make acceler- ated progress, both in intelligence and decrease in symptom severity, come from the group of autistic children with IQs of 60 or above. Low-functioning autistics, while often making symptomatic gains, do not reach normal or nearly nor- mal levels in adulthood no matter what type of treatment or educa- tion they receive. The ultimate outcome for higher functioning autistics may be enhanced by good behavioral/educational programs. Family adequacy is not related to outcome. Parents frequently be- come worn out psychologically and physically by the time of adolesence of the autistic child and seek placement. Of course, in every case in which the autistic offspring remains dependent, eventual placement must be sought because of parental illness or advancing age.

Theoretical articles supporting a nurture or nature-nurture position were rarely published in the 1970s. The consensus of most investiga-

436 SCHIZOPHRENIA BULLETIN

tors leans toward the nature theo- ries and attributes infantile autism to some type of neurobiological defect that affects the normal func- tioning of the brain. Localization studies have revealed different types of lesions in various parts of the brain, suggesting that any ill- ness or agent that can damage the brain can also probably cause the autistic syndrome. The specific site(s) of the lesion(s) has (have) not been agreed upon. In addition, a major controversy is whether language or cognitive defects are the central functional deficit. A major goal for the 1980s is more extensive studies of nonautistic brain dysfunctional children for comparison to autistic children. These studies should enable re- searchers to explain why the two groups differ in social and commu- nicative skills and how this might be correlated to neurobiological defects.

References

Abassi, V.; Linscheid, T.; and Coleman, M. Triiodothyronine (T3) concentration and therapy in au- tistic children. Journal of Autism and Childhood Schizophrenia, 8:383-387, 1978.

Albert, R.S., and Davis, A.J. A re- liability study of interparental agreement on the Rimland Diag- nostic Check List. Journal of Clini- cal Psychology, 27:499-502, 1971. Allen, J.; DeMyer, M.K.; Norton, J. A.; Pontius, W.; and Yang, E. In- tellectuality in parents of psychot- ic, subnormal, and normal chil- dren. Journal of Autism and Childhood Schizophrenia, 3:311-326, 1971.

Alter, M., and Schulenberg, R. Dermatoglyphics in rubella syn-

drome. Journal of the American Medical Association, 197: 685-688, 1966.

American Psychiatric Association, DSM-II: Diagnostic and Statistical Manual of Mental Disorders. 2nd ed. Washington, DC: APA, 1968.

American Psychiatric Association. DSM-III: Diagnostic and Statistical Manual of Mental Disorders. 3rd ed. Washington, DC: APA, 1980.

Ando, H., and Tsuda, K. Intrafamilial incidence of autism, cerebral palsy, and mongolism. Journal of Autism and Childhood Schizophrenia, 5:267-274, 1975.

Annett, M. The binomial distribu- tion of right, mixed and left handedness. Quarterly Journal of Experimental Psychology, 19:327-333, 1967.

Anthony, J. An experimental ap- proach to the psychopathology of childhood. British Journal of Medi- cal Psychology, 31:211-223, 1958.

Arick, J.R., and Krug, D.A. Autis- tic children: A study of learning characteristics and programming needs. American Journal of Mental Deficiency, 83:200-202, 1978.

Baltaxe, C.A., and Simmons, J.Q. Language in childhood psychosis: A review. Journal of Speech and Hearing Disorders, 40:439-458, 1975.

Barrera, R.D.; Lobato-Barrera, D.; and Sulzer-Azaroff, B. A simulta- neous treatment comparison of three expressive language training programs with a mute autistic child. Journal of Autism and Devel- opmental Disorders, 10:21-37, 1980. Barry, R.J., and James, A.L. Handedness in autistics, retardates, and normals of a wide age range. Journal of Autism and Childhood Schizophrenia, 8:315-323, 1978.

Bartak, L., and Rutter, M. Educa- tional treatment of autistic chil- dren. In: Rutter, M., ed. Infantile Autism: Concepts, Characteristics and Treatment. London: Churchill- Livingstone, 1971. pp. 258-280.

Bartak, L., and Rutter, M. Special educational treatment of autistic children: A comparative study—1. Design of study and characteristics of units. Journal of Child Psychology and Psychiatry, 14:161-179, 1973.

Bartak, L., and Rutter, M. The use of personal pronouns by autistic children. Journal of Autism and Childhood Schizophrenia, 4:217-222, 1974.

Bartak, L., and Rutter, M. Differ- ences between mentally retarded and normally intelligent autistic children. Journal of Autism and Childhood Schizophrenia, 6:109-120, 1976.

Bartak, L.; Rutter, M.; and Cox, A. A comparative study of infantile autism and specific developmental receptive language disorders: I. The children. British Journal of Psy- chiatry, 126:127-145, 1975.

Bartolucci, G., and Albers, R.J. Deictic categories in the language of autistic children. Journal of Au- tism and Childhood Schizophrenia, 4:131-141, 1974.

Bartolucci, G., and Pierce, S.J. A preliminary comparison of phono- logical development in autistic, normal, and mentally retarded subjects. British Journal of Disorders of Communication, 12:137-147, 1977.

Bartolucci, G.; Pierce, S.J.; and Streiner, D. Cross-sectional stud- ies of grammatical morphemes in autistic and mentally retarded chil- dren. Journal of Autism and Devel- opmental Disorders, 10:39-50, 1980. Bartolucci, G.; Pierce, S.J.; Streiner, D.; and Eppel, P.T. Pho-

VOL 7, NO. 3, 1981 437

nological investigation of verbal autistic and mentally retarded sub- jects. Journal of Autism and Child- hood Schizophrenia, 6:303-316, 1976. Bellak, L. Schizophrenia: A Review of the Syndrome. New York: Logos Press, 1958.

Belmaker, R.H.; Hattab, J.; and Ebstein, R.P. Plasma dopamine-/3- hydroxylase in childhood psycho- sis. Journal of Autism and Childhood Schizophrenia, 8:293-298, 1978.

Belmont, L., and Birch, H.G. Lat- eral dominance and right-left awareness in normal children. Child'Development, 34:257-270, 1963.

Benaroya, S.; Wesley, S.; Ogilvie, H.; Klein, L.S.; and Clarke, E. Sign language and multisensory input training of children with communication and related devel- opmental disorders: Phase II. Jour- nal of Autism and Developmental Dis- orders, 9:219-220, 1979.

Benaroya, S.; Wesley, S.; Ogilvie, H.; Klein, L.S.; and Meany, M. Sign language and multisensory input training of children with communication and related devel- opmental disorders. Journal of Au- tism and Childhood Schizophrenia, 7:23-31, 1977.

Bender, L. The life course of chil- dren with schizophrenia. American Journal of Psychiatry, 130:783-786, 1973.

Bender, L., and Faretra, G. The re- lationship between childhood schizophrenia and adult schizo- phrenia. In: Kaplan, A.R., ed. Ge- netic Factors in Schizophrenia. Springfield, IL: Charles C Thomas, 1973. pp. 28-64.

Black, M.; Freeman, B.J.; and Montgomery, J. Systematic obser- vation of play behavior in autistic children. Journal of Autism and

Childhood Schizophrenia, 5:363-371, 1975. Blackstock, E.G. Cerebral asym- metry and the development of early infantile autism. Journal of Autism and Childhood Schizophrenia, 8:339-353, 1978.

Bomberg, D.; Szurek, S.; and Etemad, J. A statistical study of a group of psychotic children. In: Szurek, S., and Berlin, I., eds. Clinical Studies in Childhood Psy- choses. New York: Brunner/Mazel Publishers, Inc., 1973. pp. 303-347.

Boucher, J. Is autism-primarily a language disorder? British Journal of Disorders of Communication, 11:135-143, 1976a. Boucher, J. Articulation in early childhood autism. Journal of Au- tism and Childhood Schizophrenia, 6:297-302, 1976b. Boucher, J. Hand preference in au- tistic children and their parents. Journal of Autism and Childhood Schizophrenia, 7:177-187, 1977. Boucher, J. Echoic memory capaci- ty in autistic children. Journal of Child Psychology and Psychiatry, 19:161-166, 1978.

Boullin, D.J.; Bhagavan, H.N.; Coleman, M.; O'Brien, R.A.; and Youdim, M.B.H. Platelet monoamine oxidase in children with infantile autism. Medical Biol- ogy, 53:210-213, 1975.

Boullin, D.J.; Coleman, M.; and O'Brien, R.A. Abnormalities in platelet 5-hydroxytryptamine ef- flux in patients with infantile au- tism. Nature, 266:371-372, 1970.

Boullin, D.J.; Coleman, M.; O'Brien, R.A.; and Rimland, B. Laboratory predictions of infantile autism based on 5-hydroxytrypta- mine efflux from blood platelets and their correlation with the

Rimland E-2 score. Journal of Au- tism and Childhood Schizophrenia, 1:63-71, 1971.

Boullin, D.J., and O'Brien, R.A. Uptake and loss of 14C-dopamine by platelets from children with in- fantile autism. Journal of Autism and Childhood Schizophrenia, 2:67-74, 1972.

Brady, D.O., and Smouse, A.D. A simultaneous comparison of three methods for language training with an autistic child: An experi- mental single case analysis. Journal of Autism and Childhood Schizo- phrenia, 8:271-279, 1978.

Brambilla, F.; Viani, F.; and Rossotti, V. Endocrine aspects of child psychoses. Diseases of the Nervous System, 30:627-632, 1969.

Browning, E.R. The effectiveness of long and short verbal com- mands in inducing correct re- sponses in three schizophrenic children. Journal of Autism and Childhood Schizophrenia, 4:293-300, 1974.

Browning, R.M. Treatment effects of a total behavior modification program with five autistic chil- dren. Behaviour Research and Ther- apy, 9:319-327, 1971.

Bryson, C.Q. Short-term memory and cross-modal information proc- essing in autistic children. Journal of Learning Disabilities, 5:81-91, 1972.

Byassee, J.E., and Murrell, S.A. Interaction patterns in families of autistic, disturbed, and normal children. American Journal of Or- thopsychiatry, 45:473-478, 1975.

Campbell, M.; Anderson, L.; Meier, M.; Cohen, I.; Small, A.; Samit, C ; and Sachar, E. A com- parison of haloperidol and behav- ior therapy and their interaction in autistic children. Journal of the

438 SCHIZOPHRENIA BULLETIN

American Academy of Child Psychia- try, 17:640-655, 1978a.

Campbell, M.; Fish, B.; David, R.; Shapiro, T.; Collins, P.; and Koh, C. Response to triiodothyronine and dextroamphetamine: A study of preschool schizophrenic chil- d r e n . Journal of Autism and Child- hood Schizophrenia, 2:343-358, 1972a.

Campbell, M.; Fish, B.; David, R.; Shapiro, T.; Collins, P.; and Koh, C. Liothyronine treatment in psy- chotic and non-psychotic children under 6 years. Archives of General Psychiatry, 29:602-608, 1973.

Campbell, M.; Fish, B.; Korein, ].; Shapiro, T.; Collins, P.; and Koh, C. Lithium and chlorpromazine: A controlled crossover study of hyperactive severely disturbed young children. Journal of Autism and Childhood Schizophrenia, 2:234-263, 1972b.

Campbell, M.; Fish, B.; Shapiro, T.; and Floyd, A. Thiothixene in young disturbed children: A pilot study. Archives of General Psychia- try, 23:70-72, 1970.

Campbell, M.; Fish, B.; Shapiro, T.; and Floyd, A. Study of molindone in disturbed preschool children. Current Therapeutic Re- search, 13:28-33, 1971a.

Campbell, M.; Fish, B.; Shapiro, T.; and Floyd, A. Imipramine in preschool autistic and schizo- p h r e n i c children. Journal of Autism and Childhood Schizophrenia, 1:267-282, 1971b.

Campbell, M.; Fish, B.; Shapiro, T.; and Floyd, A. Acute responses of schizophrenic children to a sed- ative and a "stimulating" neuroleptic: A pharmacologic yardstick. Current Therapeutic Re- search, 14:759-766, 1972c.

Campbell, M.; Friedman, E.; DeVito, E.; Greenspan, L.; and Collins, P.J. Blood serotonin in psychotic and brain damaged chil- d r e n . Journal of Autism and Child- hood Schizophrenia, 4:33-41, 1974.

Campbell, M.; Friedman, E.; Green, W.H.; Collins, P.J.; Small, A.M.; and Breuer, H. Blood serotonin in schizophrenic chil- dren. A preliminary study. Inter- national Pharmacopsychiatry, 10:213-221, 1975.

Campbell, M.; Friedman, E.; Green, W.H.; Small, A.M.; and Burdock, E.I. Blood platelet monoamine oxidase activity in schizophrenic children and their families. A preliminary study. Neuropsychobiology, 2:239-246, 19760.

Campbell, M.; Geller, B.; Small, A.M.; Petti, T.A.; and Ferris, S.H. Minor physical anomalies in young psychotic children. American Jour- nal of Psychiatry, 135:573-575, 1978b.

Campbell, M.; Hardesty, A.S.; Breuer, H.; and Polevoy, N. Child- hood psychosis in perspective. A follow-up of 10 children. Journal of the American Academy of Child Psy- chiatry, 17:14-28, 1978c.

Campbell, M.; Hardesty, A.S.; and Burdock, E.I. Demographic and perinatal profile of 105 autistic children: A preliminary report. Psychopharmacology Bulletin, 14:36-39, 1978.

Campbell, M.; Petti, T.A.; Green, W.H.; Cohen, I.L.; Genieser, N.B.; and David, R. Some physical pa- rameters of young autistic chil- d r e n . Journal of Child Psychiatry, 19:193-212, 1980.

Campbell, M.; Small, A.; Collins, P.; Friedman, E.; David, R.; and Genieser, N. Levodopa and

levoamphetamine: A crossover study in young schizophrenic chil- d r e n . Current Therapeutic Research, 19:70-86, 1976b.

Campbell, M.; Small, A.; Hol- lander, C ; Korein, J.; Cohen, I.; Kalmijn, M.; and Ferris, S. A controlled crossover study of triiodothyronine in autistic chil- d r e n . Journal of Autism and Child- hood Schizophrenia, 8:371-381, 197Sd.

Campion, E., and Tucker, G. A note on twin studies, schizophre- nia and neurological impairment. Archives of General Psychiatry, 29:460-464, 1973.

Cantwell, D.P., and Baker, L. Imi- tations and echoes in autistic and dysphasic children. Journal of the American Academy of Child Psychia- try, 17:614-624, 1978a.

Cantwell, D.P., and Baker, L. The language environment of autistic and dysphasic children. Journal of the American Academy of Child Psy- chiatry, 17:604-613, 1978b.

Cantwell, D.P.; Baker, L.; and Rutter, M. Families of autistic and dysphasic children: II. Mother's speech to the children. Journal of Autism and Childhood Schizophrenia, 7:313-327, 1977.

Cantwell, D.; Baker, L.; and Rut- ter, M. A comparative study of in- fantile autism and specific devel- opmental receptive language disorder: IV. Analysis of syntax and language function. Journal of Child Psychology and Psychiatry, 19:351-362, 1978a.

Cantwell, D.P.; Baker, L.; and Rutter, M. Family factors. In: Rut- ter, M., and Schopler, E., eds. Au- tism: A Reappraisal of Concepts and Treatment. New York: Plenum Press, 1978b. pp. 269-296.

VOL 7, NO. 3, 1981 439

Capute, A.J.; Derivan, A.T.; Chauvel, P.J.; and Rodriguez, A. Infantile autism: I. A prospective study of the diagnosis. Develop- mental Medicine and Child Neurolo- gy, 17:58-62, 1975.

Carr, E.G. The motivation of self- injurious behavior: A review of some hypotheses. Psychological Bulletin, 84:800-816, 1977.

Carr, E.G. Teaching autistic chil- dren to use sign language: Some research issues. Journal of Autism and Developmental Disorders, 9:345-359, 1979.

Carr, E.G.; Binkoff, J.A.; Kologinsky, E.; and Eddy, M. Ac- quisition of sign language by au- tistic children. I: Expressive labelling. Journal of Applied Behav- ior Analysis, 11:489-501, 1978. Carr, E.G.; Newsom, C D . ; and Binkoff, J.A. Stimulus control of self-destructive behavior in a psy- chotic child. Journal of Abnormal Child Psychology, 4:139-153, 1976. Carr, E.G.; Schreibman, L.; and Lovaas, O.I. Control of echolalic speech in psychotic children. Jour- nal of Abnormal Child Psychology, 3:331-351, 1975.

Casey, L.O. Development of com- municative behavior in autistic children: A parent program using manual signs. Journal of Autism and Childhood Schizophrenia, 8:45-59, 1978.

Chess, S. Follow-up report on au- tism in congenital rubella. Journal of Autism and Childhood Schizo- phrenia, 7:68-81, 1977.

Chess, S. Discussion: Language, cognition, and autism by Rutter; Studies of the autistic syndromes by Coleman. In: Katzman, R., ed. Congenital and Acquired Cognitive Disorders. Research Publication, The Association for Research in

Nervous and Mental Disease, Vol. 57, 1979. pp. 277-280.

Churchill, D.W. The relation of in- fantile autism and early childhood schizophrenia to developmental language disorders of childhood. Journal of Autism and Childhood Schizophrenia, 2:182-197, 1972.

Churchill, D.W. Language of Autis- tic Children. New York: John Wiley and Sons, 1978.

Churchill, D.W., and Bryson, C.Q. Looking and approach behavior of psychotic and normal children as a function of adult attention or pre- occupation. Comprehensive Psychi- atry, 13:171-177, 1972.

Claghorn, J. A double-blind com- parison of haloperidol (Haldol) and thioridazine (Mellaril) in out- patient children. Current Therapeu- tic Research, 14:785-789, 1972.

Clancy, H.; Dugdale, A.; and Rendle-Short, J. The diagnosis of infantile autism. Developmental Medicine and Child Neurology, 11:432-442, 1969.

Clark, P., and Rutter, M. Compli- ance and resistance in autistic chil- dren. Journal of Autism and Child- hood Schizophrenia, 7:33-48, 1977.

Cohen, D.J.; Caparulo, B.K.; Gold, J.R.; Waldo, M.C.; Shaywitz, B.A.; Ruttenberg, B.A.; and Rimland, B. Agreement in diagnosis: Clinical assessment and behavior rating scales for pervasively disturbed children. Journal of the American Academy of Child Psychiatry, 17:589-603, 1978.

Cohen, D.J.; Caparulo, B.; and Shaywitz, B. Primary childhood aphasia and childhood autism. Journal of the American Academy of Child Psychiatry, 15:604-645, 1976.

Cohen, D.J.; Caparulo, B.K.; Shaywitz, B.A.; and Bowers, M.B., Jr. Dopamine and serotonin me-

tabolism in neuropsychiatrically disturbed children: CSF homo- vanillic acid and 5-hydrox- yindoleacetic acid. Archives of Gen- eral Psychiatry, 34:545-550, 1977.

Cohen, D.J.; Shaywitz, B.A.; Johnson, W.T.; and Bowers, M.B., Jr. Biogenic amines in autistic and atypical children: Cerebrospinal fluid measures of homovanillic acid and 5-hydroxyindoleacetic acid. Archives of General Psychiatry, 31:845-853, 1974.

Cohen, D.J., and Young, J.G. Re- view article: Neurochemistry and child psychiatry. Journal of the American Academy of Child Psychia- try, 16:353-411, 1977. Cohen, D.J.; Young, J.G.; and Roth, J.H. Platelet monoamine oxi- dase in early childhood autism. Archives of General Psychiatry, 34:534-537, 1977.

Cohen, I.L.; Campbell, M.; Posner, D.; Small, A.M.; Triebel, D.; and Anderson, L.T. Behavioral effects of haloperidol in young au- tistic children. Journal of Child Psy- chiatry, 19:665-677, 1980. Colby, K.M., and Kraemer, H.C. An objective measurement of nonspeaking children's perform- ance with a computer-controlled program for the stimulation of lan- guage behavior. Journal of Autism and Childhood Schizophrenia, 5:139-146, 1975.

Colby, K.M., and Parkison, C. Handedness in autistic children. Journal of Autism and Childhood Schizophrenia, 7:3-9, 1977.

Coleman, M. Studies of the autis- tic syndromes. In: Katzman, R., ed. Congenital and Acquired Cogni- tive Disorders. Research Publica- tion, The Association for Research in Nervous and Mental Disease, Vol. 57, 1979. pp. 265-275.

440 SCHIZOPHRENIA BULLETIN

Coleman, M.; Campbell, M.; Freedman, L.S.; Roffman, M.; Ebstein, R.P.; and Goldstein, M. Serum dopamine-beta- hydroxylase levels in Down's syn- drome. Clinical Genetics, 5:312-315, 1974. Colman, R.; Frankel, F.; Ritvo, E.; and Freeman, B.J. The effects of fluorescent and incandescent illu- mination upon repetitive behav- iors in autistic children. Journal of Autism and Childhood Schizophrenia, 6:157-162, 1976.

Condon, W.S. Multiple response to sound in dysfunctional chil- dren. Journal of Autism and Child- hood Schizophrenia, 5:37-56, 1975. Corbett, J.; Harris, R.; Taylor, E.; and Trimble, M. Progressive disintegrative psychoses of child- hood. Journal of Child Psychology and Psychiatry, 18:211-219, 1977. Cowan, P.A.; Hodinott, B.A.; and Wright, B.A. Compliance and re- sistance in the conditioning of au- tistic children: An exploratory study. Child Development, 36:913-923, 1965. Cox, A.; Rutter, M.; Newman, S.; and Bartak, L. A comparative study of infantile autism and spe- cific developmental receptive lan- guage disorder: II. Parental char- acteristics. British Journal of Psychiatry, 126:146-159, 1975. Creak, M. Schizophrenic syn- drome in childhood: Further prog- ress report of a working party. De- velopmental Medicine and Child Neurology, 6:530-535, 1964.

Creedon, M.P., ed. Appropriate Be- havior Through Communication. Chicago: Michael Reese Medical Center, Dysfunctioning Child Center Publication, 1975. Curcio, F. Sensorimotor func- tioning and communication in

mute autistic children. Journal of Autism and Childhood Schizophrenia, 8:281-292, 1978. Curcio, F., and Piserchia, E.A. Pantomimic representation in psy- chotic children. Journal of Autism and Childhood Schizophrenia, 8:181-189, 1978.

Damasio, A.R., and Maurer, R.G. A neurological model for child- hood autism. Archives of Neurology, 35:777-786, 1978.

Damasio, H.; Maurer, R.G.; Damasio, A.R.; and Chui, H.C. Computerized tomographic scan findings in patients with autistic behavior. Archives of Neurology, 37:504-510, 1980. Darby, J.K. Neuropathologic as- pects of psychosis in children. Journal of Autism and Childhood Schizophrenia, 6:339-352, 1976. Davids, A. Childhood psychosis: The problem of differential diag- nosis. Journal of Autism and Child- hood Schizophrenia, 5:129-138,1975. DeMyer, M.K. The nature of the neuropsychological disability in autistic children. Journal of Autism and Childhood Schizophrenia, 5:109-128, 1975a.

DeMyer, M.K. Research in infan- tile autism: A strategy and its re- sults. Biological Psychiatry, 10:433^*52, 1975b.

DeMyer, M.K. Parents and Children in Autism. Washington, DC: Victor H. Winston & Sons, 1979.

DeMyer, M.K.; Alpern, G.D.; Barton, S.; DeMyer, W.; Churchill, D.W.; Hingtgen, J.N.; Bryson, C.Q.; Pontius, W.; and Kimberlin, C. Imitation in autistic, early schizophrenic, and non-psychotic subnormal children. Journal of Au- tism and Childhood Schizophrenia, 2:264-287, 1972a.

DeMyer, M.K.; Barton, S.; Alpern, G.D.; Kimberlin, C ; Allen, J.; Yang, E.; and Steele, R. The meas- ured intelligence of autistic chil- dren. Journal of Autism and Child- hood Schizophrenia, 4:42-60, 1974. DeMyer, M.K.; Barton, S.; DeMyer, W.E.; Norton, J.A.; Allen, J.; and Steele, R. Prognosis in autism: A follow-up study. Journal of Autism and Childhood Schizophrenia, 3:199-246, 1973.

DeMyer, M.K.; Barton, S.; and Norton, J.A. A comparison of adaptive, verbal, and motor pro- files of psychotic and non- psychotic subnormal children. Journal of Autism and Childhood Schizophrenia, 2:359-377, 1972.

DeMyer, M.K.; Bryson, C.Q.; and Churchill, D.W. The earliest indi- cators of pathological develop- ment: Comparison of symptoms during infancy and early child- hood in normal, subnormal, schiz- ophrenic and autistic children. In: Biological and Environmental Deter- minants of Early Development. Re- search Publication, The Associa- tion for Research in Nervous and Mental Disease, Vol. 51, 1973. pp. 298-332.

DeMyer, M.K.; Churchill, D.W.; Pontius, W.; and Gilkey, K.M. A comparison of five diagnostic sys- tems for childhood schizophrenia and infantile autism. Journal of Au- tism and Childhood Schizophrenia, 1:175-189, 1971a.

DeMyer, M.K.; Pontius, W.; Norton, J.A.; Barton, S.; Allen, J.; and Steele, R. Parental practices and innate activity in normal, au- tistic, and brain-damaged infants. Journal of Autism and Childhood Schizophrenia, 2:49-66, 1972b. DeMyer, M.K.; Schwier, H.; Bryson, C.Q.; Solow, E.B.; and

VOL 7, NO. 3, 1981 441

Roeske, N. Free fatty acid re- sponse to insulin and glucose stimulation in schizophrenic, au- tistic, and emotionally disturbed children. Journal of Autism and Childhood Schizophrenia, 1A36-452, 1971b.

Despert, J.L. Reflections on early infantile autism. Journal of Autism and Childhood Schizophrenia, 1:363-367, 1971.

DeVilliers, J.G., and Naughton, J.M. Teaching a symbol language to autistic children. Journal of Con- sulting and Clinical Psychology, 42:lir-117, 1974.

Douglas, V.I., and Sanders, F.A. A pilot study of Rimland's diag- nostic checklist with autistic and mentally retarded children. Journal of Child Psychology and Psychiatry, 9:105-109, 1968.

Egel, A.L., Koegel, R.L., and Schreibman, L. Review of educational-treatment procedures for autistic children. In: Mann, L., and Sabatino, D.A., eds. The Fourth Review of Special Education. New York: Grune & Stratton, 1980. pp. 109-149.

Eggers, C. Course and prognosis of childhood schizophrenia. Jour- nal of Autism and Childhood Schizo- phrenia, 8:21-36, 1978.

Eisenberg, L. The autistic child in adolescence. American Journal of Psychiatry, 112:607-613, 1956.

Eisenberg, L. The classification of the psychotic disorders in child- hood. In: Eron, L.D., ed. Classifi- cation of Behavior Disorders. Chicago: Aldine Publishing Com- pany, 1966. pp. 87-122.

Ekstein, R., and Friedman, S.W. Infantile autism: From entity to process. Reiss-Davis Clinic Bulletin, 11:70-85, 1974.

Engelhardt, D.; Polizos, P.; Waizer, J.; and Hoffman, S. A double-blind comparison of fluphenazine and haloperidol in outpatient schizophrenic children. Journal of Autism and Childhood Schizophrenia, 3:128-137, 1973.

Etemad, J.G., and Szurek, S.A. A modified follow-up study of a group of psychotic children. In: Szurek, S.A., and Berlin, I.N., eds. Clinical Studies in Childhood Psychoses. New York: Brunner/Mazel, 1973. pp. 348-371.

Etzel, B.C., and LeBlanc, J.M. The simplest treatment alternative: The law of parsimony applied to choosing appropriate instructional control and errorless-learning pro- cedures for the difficult-to-teach child. Journal of Autism and Devel- opmental Disorders, 9:361-382, 1979.

Faretra, G.; Dooher, L.; and Dowling, J. Comparison of haloperidol and fluphenazine in disturbed children. American Jour- nal of Psychiatry, 126:1670-1673, 1970.

Fassler, J., and Bryant, N.D. Dis- turbed children under reduced au- ditory input: A pilot study. Excep- tional Children, 38:197-204, 1971.

Fein, D.; Tinder, P.; and Waterhouse, L. Stimulus generali- zation in autistic and normal chil- dren. Journal of Child Psychology and Psychiatry, 20:325-335, 1979.

Ferster, C.B., and DeMyer, M.K. The development of performances in autistic children in an automat- ically controlled environment. Journal of Chronic Diseases, 13:312-345, 1961.

Fischer, I., and Glanville, B.W.K. Programmed teaching of autistic children: Scholastic progress over one year. Archives of General Psy- chiatry, 23:90-94, 1970.

Fish, B. Neurobiologic antecedents of schizophrenia in children: Evi- dence for an inherited congenital neurointegrative defect. Archives of General Psychiatry, 34:1297-1313, 1977.

Fish, B., and Hagin, R. Visual- motor disorders in infants at risk for schizophrenia. Archives of Gen- eral Psychiatry, 28:900-904, 1973.

Florsheim, J., and Peterfreund, O. The intelligence of parents of psy- chotic children. Journal of Autism and Childhood Schizophrenia, 4:61-70, 1974.

Folstein, S., and Rutter, M. Genet- ic influences and infantile autism. Nature, 265:726-728, 1977.

Foxx, R.M., and Azrin, N.H. The elimination of autistic self- stimulatory behavior by overcorrection. Journal of Applied Behavior Analysis, 6:1-14, 1973.

Fraknoi, J., and Ruttenberg, B.A. Formulation of the dynamic eco- nomic factors underlying infantile autism. Journal of the American Academy of Child Psychiatry, 10:713-738, 1971.

Frank, S.M.; Allen, D.A.; Stein, L.; and Myers, B. Linguistic per- formance in vulnerable and autis- tic children and their mothers. American Journal of Psychiatry, 133:909-915, 1976.

Frankel, F. Experimental studies of autistic children in the classroom. In: Ritvo, E.; Freeman, B.J.; Ornitz, E.; and Tanguay, P., eds. Autism: Diagnosis, Current Research and Management. New York: Spec- trum Publications, 1976. pp. 185-194.

Frankel, F.; Freeman, B.J.; Ritvo, E.; Chikami, B.; and Carr, E. Ef- fects of frequency of photic stimu- lation upon autistic and retarded

442 SCHIZOPHRENIA BULLETIN

children. American Journal of Men- tal Deficiency, 81:32-40, 1976. Frankel, F.; Freeman, B.J.; Ritvo, E.; and Pardo, R. The effect of en- vironmental stimulation upon the stereotyped behavior of autistic children. Journal of Autism and Childhood Schizophrenia, 8:389-394, 1978.

Frankel, F., and Graham, V. Sys- tematic observation of classroom behavior of retarded and autistic preschool children. American Jour- nal of Mental Deficiency, 81:73-84, 1976.

Frankel, F., and Simmons, J.Q. Self-injurious behavior in schizo- phrenics and retarded children. American Journal of Mental Defi- ciency, 80:512-522, 1976.

Freeman, B.J.; Frankel, F.; and Ritvo, E. The effects of response contingent vestibular stimulation on the behavior of autistic and re- t a r d e d children. Journal of Autism and Childhood Schizophrenia, 6:353-358, 1976.

Freeman, B.J.; Ritvo, E.; Guthrie, D.; Schroth, P.; and Ball, J. The Behavior Observation Scale for Autism: Initial methodology, data analysis, and preliminary findings on 89 children. Journal of the Amer- ican Academy of Child Psychiatry, 17:576-588, 1978.

Frith, U. Studies in pattern detec- tion in normal and autistic chil- dren: I. Immediate recall of audito- ry sequences. Journal of Abnormal Psychology, 76:413-420, 1970a.

Frith, U. Studies in pattern detec- tion in normal and autistic chil- dren: II. Reproduction and pro- duction of color sequences. Journal of Experimental Child Psychology, 10:120-135, 1970b.

Frith, U. Spontaneous patterns produced by autistic, normal and

subnormal children. In: Rutter, M., ed. Infantile Autism: Concepts, Characteristics and Treatment. Edinburgh: Churchill and Livingstone, 1971. pp. 113-131.

Frith, U. Cognitive mechanisms in autism: Experiments with color and tone sequence production. Journal of Autism and Childhood Schizophrenia, 2:160-173, 1972.

Fulkerson, S.C., and Freeman, W.M. Perceptual-motor deficiency in autistic children. Perceptual and Motor Skills, 50:331-336, 1980.

Fulwiler, R.L., and Fouts, R.S. Ac- quisition of American sign lan- guage by a non-communicating autistic child. Journal of Autism and Childhood Schizophrenia, 6:43-51, 1976.

Gallagher, J.J., and Wiegerink, R. Educational strategies for the au- tistic child. Journal of Autism and Childhood Schizophrenia, 6:15-26, 1976.

Gittelman, M., and Cleeman, J. Serum magnesium level in psy- chotic and normal children. Behavioral Neuropsychiatry, 1:51-52, 1969.

Gittelman, M., and Cleeman, J. Serum magnesium differences: Further evidence for discontinuity between adult and childhood schizophrenia. Mental Hygiene, 55:492-494, 1971.

Goetz, L.; Schuler, A.; and Sailor, W. Teaching functional speech to the severely handicapped: Current issues. Journal of Autism and Devel- opmental Disorders, 9:325-343, 1979.

Goldfarb, W. Childhood Schizo- phrenia. Cambridge, MA: Harvard University Press, 1961.

Goldfarb, W. Childhood psycho- sis. In: Mussen, P.H., ed. Carmichael's Manual of Child Psy-

chology. Vol. 2. New York: John Wiley & Sons, 1970. pp. 765-830.

Goldfarb, W.; Spitzer, R.L.; and Endicott, J.A. A study of psycho- pathology of parents of psychotic children by structured interview. Journal of Autism and Childhood Schizophrenia, 6:327-338, 1976.

Goldfarb, W.; Yudkovitz, E.; and Goldfarb, N. Verbal symbols to designate objects: An experimental study of communication in mothers of schizophrenic children. Journal of Autism and Childhood Schizophrenia, 3:281-298, 1973.

Goldstein, M. Dopamine-beta- hydroxylase and endogenous total 5-hydroxyindole levels in autistic patients and controls. In: Coleman, M., ed. The Autistic Syn- dromes. New York: North- Holland, 1976. pp. 57-58.

Graziano, A.M. A group treatment approach to multiple problem be- haviors of autistic children. Excep- tional Children, 36:765-770, 1970.

Greenbaum, G. An evaluation of niacinamide in the treatment of childhood schizophrenia. American Journal of Psychiatry, 127:89-92, 1970.

Gubbay, S.S.; Lobascher, M.; and Kingerlee, P. A neurological ap- praisal of autistic children: Results of a Western Australian survey. Developmental Medicine and Child Neurology, 12:422-429, 1970.

Guthrie, R.D., and Wyatt, R.J. Bi- ochemistry and schizophrenia. III. A review of childhood psychosis. Schizophrenia Bulletin, 12:18-32, 1975.

Halpern, W.I. The schooling of au- tistic children: Preliminary find- i n g s . American Journal of Orthopsy- chiatry, 40:665-671, 1970.

Handford, H.A. Brain hypoxia, minimal brain dysfunction, and

VOL. 7, NO. 3, 1981 443

schizophrenia. American Journal of Psychiatry, 132:192-194, 1975.

Hartley, H.G.; Stahl, S.M.; and Freedman, D.X. Hyperserotone- mia and amine metabolites in au- tistic and retarded children. Ar- chives of General Psychiatry, 34:521-531, 1977.

Hanson, D.R., and Gottesman, I.I. The genetics, if any, of infantile autism and childhood schizophre- nia. Journal of Autism and Childhood Schizophrenia, 6:209-234, 1976.

Haracopos, D., and Kelstrup, A. Psychotic behavior in children under the institutions for the men- tally retarded in Denmark. Journal of Autism and Childhood Schizo- phrenia, 8:1-12, 1978.

Hargrave, E., and Swisher, L. Modifying the verbal expression of a child with autistic behaviors. Journal of Autism and Childhood Schizophrenia, 5:147-154, 1975.

Harper, ]., and Williams, S. Early environmental stress and infantile autism. The Medical Journal of Australia, 1:341-346, 1974.

Harper, ]., and Williams, S. Infan- tile autism: The incidence of na- tional groups in a New South Wales survey. The Medical Journal of Australia, 1:299-301, 1976.

Harris, S.L., and Wolchik, S.A. Suppression of self-stimulation: Three alternative strategies. Jour- nal of Applied Behavior Analysis, 12:185-198, 1979.

Hauser, S.L.; DeLong, G.R.; and Rosman, N.P. Pneumoencephalo- graphic findings in the infantile autism syndrome: A correlation with temporal lobe disease. Brain, 98:667-688, 1975.

Hauser, S.L., and Williams, R.S. "A Neuropathologic Study of Three Cases of Idiopathic Autism" (Abstract). Presented at the 55th

Annual Meeting of the American Association of Neuropathologists, Kansas City, KS, June 7-10, 1979.

Hayes, R.W., and Gordon, A.G. Auditory abnormalities in autistic children, lancet, 11:767, 1977.

Hermelin, B. Rules and language. In: Rutter, M., ed. Infantile Autism: Concepts, Characteristics and Treat- ment. Edinburgh: Churchill and Livingstone, 1971. pp. 98-111.

Hermelin, B. Locating events in space and time: Experiments with autistic, blind, and deaf children. Journal of Autism and Childhood Schizophrenia, 2:288-298, 1972.

Hermelin, B., and O'Connor, N. Psychological Experiments With Au- tistic Children. Oxford: Pergamon Press, 1970.

Hermelin, B., and O'Connor, N. Spatial coding in normal, autistic and blind children. Perceptual and Motor Skills, 33:127-132, 1971.

Hertzig, M.E., and Walker, H.A. Symptom formation as an expres- sion of disordered information processing in schizophrenic chil- dren. Journal of Autism and Child- hood Schizophrenia, 5:13-24, 1975.

Hier, D.E.; LeMay, M.; and Rosenberger, P.B. Autism: Associ- ation with reversed cerebral asym- metry. Neurology, 28:348-349, 1978.

Hier, D.E.; LeMay, M.; and Rosenberger, P.B. Autism and un- favorable left-right asymmetries of the brain. Journal of Autism and De- velopmental Disorders, 9:153-159, 1979.

Hilbun, W.B. Dermatoglyphic findings on a group of psychotic children. The Journal of Nervous and Mental Disease, 151:352-358, 1970.

Hill, S.D.; Wagner, E.A.; Shedlarski, J.G.; and Sears, S.P.

Diurnal cortisol and temperature variation of normal and autistic children. Developmental Psychobi- ology, 10:579-583, 1977.

Himwich, H.E.; Jenkins, R.L.; Fujimori, M.; Narasimhachari, N.; and Ebersole, M. A biochemical study of early infantile autism. Journal of Autism and Childhood Schizophrenia, 2:114-126, 1972.

Hingtgen, J.N., and Bryson, C.Q. Recent developments in the study of early childhood psychoses: In- fantile autism, childhood schizo- phrenia, and related disorders. Schizophrenia Bulletin, 5:8-54, 1972.

Hung, D.W. Using self- stimulation as reinforcement for autistic children. Journal of Autism and Childhood Schizophrenia, 8:355-366, 1978.

Husted, J.R.; Hall, P.; and Agin, B. The effectiveness of time-out in reducing maladaptive behavior of autistic and retarded children. The Journal of Psychology, 79:189-196, 1971.

Hutt, S., and Hutt, C. Behavior Studies in Psychiatry. Oxford: Pergamon Press, 1970.

Jackson, M.J., and Garrod, P.J. Plasma zinc, copper, and amino acid levels in the blood of autistic children. Journal of Autism and Childhood Schizophrenia, 8:203-208, 1978.

James, A.L., and Barry, R.J. A re- view of psychophysiology in early onset psychosis. Schizophrenia Bul- letin, 6:506-525, 1980. Johnson, R.J.; Wiersema, V.; and Kraft, I.A. Hair amino acids in childhood autism. Journal of Au- tism and Childhood Schizophrenia, 4:187-188, 1974.

Kanner, L. Autistic disturbances of affective contact. Nervous Child, 2:217-250, 1943.

444 SCHIZOPHRENIA BULLETIN

Kanner, L. Problems of nosology and psychodynamics of early in- fantile autism. American Journal of Orthopsychiatry, 19:416-426, 1949.

Kanner, L. Follow-up study of 11 autistic children originally re- ported in 1943. Journal of Autism and Childhood Schizophrenia, 1:119-145, 1971.

Kanner, L. Childhood Psychosis: Ini- tial Studies and New Insights. Wash- ington, DC: V.H. Winston & Sons, 1973.

Kennard, M.A. Value of equivocal signs in neurological diagnosis. Neurology, 10:753-764, 1960.

Kikuchi, M.; Morimoto, Y.; Mach- ida, S.; Yamazaki, A.; Yogashi, Y.; Shtara, M.; Saito, Y.; Ito, N.; and Hamada, M. Family dynamics of autistic children: Report on paren- tal intelligence (abstract). Journal of Autism and Childhood Schizophrenia, 1:236, 1971.

King, P.D. Early infantile autism: Relation to schizophrenia. Journal of the American Academy of Child Psychiatry, 14:666-682, 1975.

Knobloch, H., and Pasamanick, B. Some etiologic and prognostic fac- tors in early infantile autism and psychosis. Pediatrics, 55:182-191, 1975.

Koegel, R.L.; Dunlap, G.; and Dyer, K. Intertrial interval dura- tion and learning in autistic chil- dren. Journal of Applied Behavior Analysis, 13:91-99, 1980.

Koegel, R.L., and Egel, A.L. Moti- vating autistic children. Journal of Abnormal Psychology, 88:418^*26, 1979.

Koegel, R.L.; Egel, A.L.; and Dun- lap, G. Learning characteristics of autistic children. In: Sailor, W.; Wilcox, B.; and Brown, L., eds. Methods of Instruction With Severely

Handicapped Students. Baltimore: Brookes Publishers, 1980. pp. 259-301.

Koegel, R.L.; Firestone, P.B.; Kramme, K.W.; and Dunlap, G. Increasing spontaneous play by suppressing self-stimulation in au- tistic children. Journal of Applied Behavior Analysis, 7:521-528, 1974.

Koegel, R.L., and Lovaas, O.I. Comments on autism and stimulus over selectivity. Journal of Abnormal Psychology, 87:563-565, 1978.

Koegel, R.L., and Rincover, A. Treatment of psychotic children in a classroom environment: I. Learn- ing in a large group. Journal of Ap- plied Behavior Analysis, 7:45-59, 1974.

Koegel, R.L., and Rincover, A. Some detrimental effects of using extra stimuli to guide learning in normal and autistic children. Jour- nal of Abnormal Child Psychology, 4:59-71, 1976.

Koegel, R.L., and Rincover, A. Re- search on the difference between generalization and maintenance in extra-therapy responding. Journal of Applied Behavior Analysis, 10:1- 12, 1977.

Koegel, R.L.; Russo, D.C.; and Rincover, A. Assessing and train- ing teachers in the generalized use of behavior modification with au- tistic children. Journal of Applied Behavior Analysis, 10:197-205, 1977.

Koegel, R.L., and c aireibman, L. Teaching autistic children to re- spond to simultaneous multiple cues. Journal of Experimental Child Psychology, 24:299-311, 1977.

Koegel, R.L.; Schreibman, L.; Britten, K.; and Laitinen, R. The effects of schedule of reinforce- ment on stimulus overselectivity in autistic children. Journal of Autism

and Developmental Disorders, 9:383- 397, 1979.

Kolvin, I. Studies in the childhood psychoses: I. Diagnostic criteria and classification. British Journal of Psychiatry, 118:381-384, 1971.

Kolvin, I.; Humphrey, M.; and McNay, A. Studies in the child- hood psychoses: VI. Cognitive fac- tors in childhood psychoses. Brit- ish Journal of Psychiatry, 118:415- 419, 1971.

Kolvin, I.; Ounsted, C ; Richard- son, L.M.; and Garside, R.F. Stud- ies in the childhood psychoses: III. The family and social background in childhood psychoses. British Journal of Psychiatry, 118:396-402, 1971.

Kolvin, I.; Ounsted, C ; and Roth, M. Studies in the childhood psy- choses: V. Cerebral dysfunction and childhood psychoses. British Journal of Psychiatry, 118:407-414. 1971.

Konstantareas, M.M.; Webster, C D . ; and Oxman, J. Manual lan- guage acquisition and its influence on other areas of functioning in four autistic-like children. Journal of Child Psychology and Psychiatry, 20:337-350, 1979.

Kovattana, P.M., and Kraemer, H.C. Response to multiple visual cues of color, size and form by au- tistic children. Journal of Autism and Childhood Schizophrenia, 4:251-261, 1974.

Kozloff, M.A. Reaching the Autistic Child: A Parent Training Program. Champaign, IL: Research Press, 1973.

Lake, C.R.; Ziegler, M.G.; and Murphy, D.L. Increased norepi- nephrine levels and decreased dopamine-/3-hydroxylase activity in primary autism. Archives of General Psychiatry, 34:553-566, 1977.

VOL 7, NO. 3, 1981 445

LaVigna, G.W. Communication training in mute autistic adoles- cents using the written word. Jour- nal of Autism and Childhood Schizo- phrenia, 7:135-149, 1977.

Lennox, C ; Callias, M.; and Rut- ter, M. Cognitive characteristics of parents. Journal of Autism and Childhood Schizophrenia, 7:243-261, 1977.

Lichstein, K.L., and Schreibman, L. Employing electric shock with autistic children: A review of the side effects. Journal of Autism and Childhood Schizophrenia, 6:163-173, 1976. "

Litrownik, A.J.; Mclnnis, E.T.; Wetzel-Pritchard, A.M.; and Filipelli, D.L. Restricted stimulus control and inferred attentional deficits in autistic and retarded children Journal of Abnormal Psy- chology, 87:554-562, 1978.

Lobascher, M.E.; Kingerlee, P.E.; and Gubbay, S.S. Childhood au- tism: An investigation of aetiolog- ical factors in twenty-five cases. British Journal of Psychiatry, 117:525-529, 1970.

Lockyer, L., and Rutter, M. A five to fifteen year follow-up study of infantile psychosis: IV. Patterns of cognitive ability. British Journal of Social and Clinical Psychology, 9:152-163, 1970.

Lotter, V. Epidemiology of autistic conditions in young children: I. Prevalence. Social Psychiatry, 1:124-137, 1966.

Lotter, V. Social adjustment and placement of autistic children in Middlesex: A follow-up. Journal of Autism and Childhood Schizophrenia, 4:11-32, 1974.

Lotter, V. Childhood autism in Africa. Journal of Child Psychology and Psychiatry, 19:231-244, 1978.

Lovaas, O.I. The Autistic Child: Language Development Through Be- havior Modification. New York: Irvington Publishers, Inc., 1977.

Lovaas, O.I.; Koegel, R.L.; and Schreibman, L. Stimulus overse- lectivity in autism: A review of re- search. Psychological Bulletin, 86:1236-1254, 1979.

Lovaas, O.I.; Koegel, R.; Sim- mons, J.Q.; and Long, J.S. Some generalization and follow-up measures on autistic children in behavior therapy. Journal of Ap- plied Behavior Analysis, 6:131-166, 1973.

Lovaas, O.I.; Litrownik, A.; and Mann, R. Response latencies to auditory stimuli in autistic chil- dren engaged in self-stimulatory behavior. Behaviour Research and Therapy, 9:39-49, 1971.

Lovaas, O.I., and Schreibman, L. Stimulus overselectivity of autistic children in a two stimulus situa- tion. Behaviour Research and Thera- py, 9:305-310, 1971.

Lovaas, O.I.; Schreibman, L.; and Koegel, R.L. A behavior modifica- tion approach to the treatment of autistic children. Journal of Autism and Childhood Schizophrenia, 4:111- 129, 1974.

Lovaas, O.I.; Schreibman, L.; Koegel, R.; and Rehm, R. Selective responding by autistic children to multiple sensory input. Journal of Abnormal Psychology, 77:211-222, 1971.

Lovaas, O.I.; Varni, J.W.; Koegel, R.L.; and Lorsch, N. Some obser- vations on the nonextinguish- ability of children's speech. Child Development, 48:1121-1127, 1977.

Mahler, K.R.; Harper, J.F.; Ma- cleay, A.; and King, M.G. Peculi- arities in the endocrine response to insulin stress in early infantile au-

tism. Journal of Nervous and Mental Disease, 161:180-184, 1975.

Marcus, L.M. Patterns of coping in families of psychotic children. American Journal of Orthopsychiatry, 47:388-399, 1977.

Marcus, L.M.; Lansing, M.; An- drews, C.E.; and Schopler, E. Im- provement of teaching effective- ness in parents of autistic children. Journal of the American Academy of Child Psychiatry, 17:625-639, 1978.

Margolies, P.J. Behavioral ap- proaches to the treatment of early infantile autism: A review. Psycho- logical Bulletin, 84:249-264, 1977.

Marshall, N.R., and Hegrenes, J.R. Programmed communication therapy for autistic mentally re- tarded children. Journal of Speech and Hearing Disorders, 35:70-83, 1970.

Martin, G. Brief time-outs as con- sequences for errors during train- ing programs with autistic and re- tarded children: A questionable procedure. The Psychological Re- cord, 25:71-89, 1975.

Masagatani, G.N. Hand-gesturing behavior in psychotic children. The American Journal of Occupational Therapy, 27:24-29, 1973.

Massie, H.N. Blind ratings of mother-infant interaction in home movies of pre-psychotic and nor- mal infants. American Journal of Psychiatry, 135:1371-1374, 1978.

Masters, J.C., and Miller, D.E. Early infantile autism: A methodo- logical critique. Journal of Abnormal Psychology, 75:342-343, 1970.

McAdoo, W.G., and DeMyer, M.K. Personality characteristics of parents. In: Rutter, M., and Schopler, E., eds. Autism: A Reap- praisal of Concepts and Treatment. New York: Plenum Press, 1978a. pp. 251-267.

446 SCHIZOPHRENIA BULLETIN

McAdoo, W.G., and DeMyer, M.K. Research related to family factors in autism. Journal of Pediat- ric Psychology, 2:162-166, 1978b. McAndrew, J.; Case, Q.; and Tref- fert, D. Effects of prolonged phenothiazine intake on psychotic and other hospitalized children. Journal of Autism and Childhood Schizophrenia, 2:75-91, 1972.

McLean, L.P., and McLean, J.E. A language training program for nonverbal autistic children. Journal of Speech and Hearing Disorders, 39:186-193, 1974.

Miller, R.T. Childhood schizo- phrenia: A review of selected liter- ature. International Journal of Mental Health, 3:3-46, 1974.

Miller, A., and Miller, E.E. Cogni- tive developmental training with elevated boards and sign lan- guage. Journal of Autism and Child- hood Schizophrenia, 3:65-85, 1973. Mogar, R., and Aldrich, R. The use of psychedelic agents with au- tistic schizophrenic children. Be- havioral Neuropsychiatry, 1:44-50, 1969.

Money, J.; Bobrow, N.A.; and Clarke, F.C. Autism and autoim- mune disease: A family study. Journal of Autism and Childhood Schizophrenia, 1:146-160, 1971.

Morton-Evans, A., and Hensley, R. Paired associate learning in early infantile autism and recep- tive developmental aphasia. Jour- nal of Autism and Childhood Schizo- phrenia, 8:61-69, 1978.

Myklebust, H.; Killen, J.; and Bannochie, M. Emotional charac- teristics of learning disability. Jour- nal of Autism and Childhood Schizo- phrenia, 2:151-159, 1972.

Narasimhachari, N., and Him- wich, H.E. Biochemical studies of

early infantile autism. Biological Psychiatry, 10:425-432, 1975.

Netley, C , and Lockyer, L. Meth- odological problems in the study of parental cognitive characteris- tics. Journal of Autism and Childhood Schizophrenia, 8:115-117, 1978.

Netley, C ; Lockyer, L.; and Greenbaum, G.H.C. Parental characteristics in relation to diag- nosis and neurological status in childhood psychosis. British Jour- nal of Psychiatry, 127:440-444, 1975.

Neuman, C.J., and Hill, S.D. Self- recognition and stimulus prefer- ence in autistic children. Develop- mental Psychobiology, 11:571-578, 1978.

Noach, M. Concept formation in the speaking autistic child. Interna- tional Journal of Mental Health, 3:100-109, 1974.

Nordquist, V.M., and Wahler, R.G. Naturalistic treatment of an autistic child. Journal of Applied Be- havior Analysis, 6:79-87, 1973.

O'Moore, M. A study of the aetiol- ogy of autism from a study of birth and family characteristics. Journal of the Irish Medical Association, 65:114-120, 1972.

Ornitz, E.M.; Guthrie, D.; and Farley, A.H. The early develop- ment of autistic children. Journal of Autism and Childhood Schizophrenia, 7:207-229, 1977.

Ornitz, E.M., and Ritvo, E. The syndrome of autism: A critical re- view. American Journal of Psychia- try, 133:609-621, 1976.

Oxman, J.; Webster, D.C.; and Konstantareas, M.M. Condon's multiple-response phenomenon in severely dysfunctional children: An attempt at replication. Journal of Autism and Childhood Schizophre- nia, 8:395-402, 1978.

Parker, G. Season of birth in New South Wales. The Medical Journal of Australia, 2:563-566, 1978.

Peiper, A. Cerebral Function in In- fancy and Childhood. New York: Consultants Bureau, 1963. p . 243.

Perry, T.L.; Hansen, S.; and Christie, R.G. Amino compounds and organic acids in CSF, plasma, and urine of autistic children. Bio- logical Psychiatry, 13:575-586, 1978.

Philips, G.M., and Dyer, C. Late onset echolalia in autism and allied disorders. British Journal of Disor- ders of Communication, 12:47-59, 1977.

Pierce, S., and Bartolucci, G. A syntactic investigation of verbal autistic, mentally retarded and normal children. Journal of Autism and Childhood Schizophrenia, 7:121-134, 1977.

Piggott, L.R., and Gottlieb, J.S. Childhood schizophrenia—What is it? Journal of Autism and Child- hood Schizophrenia, 3:96-105, 1973.

Piggott, L.R., and Simson, C.B. Changing diagnosis of childhood psychosis. Journal of Autism and Childhood Schizophrenia, 5:239-245, 1975.

Plummer, S.; Baer, D.M.; and LeBlanc, J.M. Functional consider- ations in the use of procedural timeout and an effective alterna- tive. Journal of Applied Behavior Analysis, 10:689-705, 1977.

Polan, C.C., and Spencer, B.L. Checklist of symptoms of autism in early life. West Virginia Medical Journal, 55:198-204, 1959.

Polizos, P.; Engelhardt, D.; Hoff- man, S.; and Waizer, J. Neurolog- ical consequences of psychotropic drug withdrawal in schizophrenic children. Journal of Autism and Childhood Schizophrenia, 3:247-253, 1973.

VOL. 7, NO. 3, 1981 447

Pollack, M.; Gittelman, M.; Miller, R.; Berman, P.; and Bakwin, R. A developmental pediatric, neuro- logical, psychological and psychi- atric comparison of psychotic chil- dren and their sibs. American Journal of Orthopsychiatry, 40:329- 330, 1970.

Porges, S.W. Peripheral and neurochemical parallels of psycho- pathology: A psychophysiological model relating autonomic imbal- ance to hyperactivity, psychop- athy, and autism. Advances in Child Development and Behavior, 11:35-65, 1976.

Prior, M.R. Cognitive abilities and disabilities in infantile autism: A review. Journal of Abnormal Child Psychology, 7:357-380, 1979.

Prior, M.R. Conditional matching learning set performance in autis- tic children. Journal of Child Psy- chology and Psychiatry, 18:183-189, 1977a.

Prior, M.R. Psycholinguistic disa- bilities of autistic and retarded children. Journal of Mental Deficien- cy Research, 21:37-45, 1977b.

Prior, M., and Bence, R. A note on the validity of the Rimland diag- nostic checklist. Journal of Clinical Psychology, 31:510-513, 1975.

Prior, M.; Boulton, D.; Gajzago, C ; and Perry, D. The classification of childhood psychoses by numer- ical taxonomy. Journal of Child Psy- chology and Psychiatry, 16:321-330, 1975.

Prior, M.R., and Chen, C.S. Learning set acquisition in autistic children. Journal of Abnormal Psy- chology, 84:701-708, 1975.

Prior, M.R., and Chen, C.S. Short- term and serial memory in autistic, retarded, and normal children. Journal of Autism and Childhood Schizophrenia, 6:121-131, 1976.

Prior, M., and Macmillan, M.B. Maintenance of sameness in chil- dren with Kanner's syndrome. Journal of Autism and Childhood Schizophrenia, 3:154-167, 1973.

Rees, S.C., and Taylor, A. Prog- nostic antecedents and outcome in a follow-up study of children with a diagnosis of childhood psycho- sis. Journal of Autism and Childhood Schizophrenia, 5:309-322, 1975.

Ricks, D.M., and Wing, L. Lan- guage, communication, and the use of symbols in normal and au- tistic children. Journal of Autism and Childhood Schizophrenia, 5:191- 221, 1975.

Rimland, B. The differentiation of childhood psychoses: An analysis of checklists for 2,218 psychotic children. Journal of Autism and Childhood Schizophrenia, 1:161-174, 1971.

Rimland, B.; Callaway, E.; and Dreyfus, P. The effects of high doses of vitamin B6 on autistic children: A double-blind crossover study. American Journal of Psychia- try, 135:472-475, 1978. Rincover, A. Sensory extinction: A procedure for eliminating self- stimulatory behavior in develop- mentally disabled children. Journal of Abnormal Child Psychology, 6:299-310, 1978a.

Rincover, A. Variables affecting stimulus fading and discriminative responding in psychotic children. Journal of Abnormal Psychology, 87:541-553, 1978b.

Rincover, A., and Koegel, R.L. Setting generality and stimulus control in autistic children. Journal of Applied Behavior Analysis, 8: 235-246, 1975.

Rincover, A., and Koegel, R.L. Classroom treatment of autistic children: II. Individualized in-

struction in a group. Journal of Ab- normal Child Psychology, 5:113-126, 1977.

Rincover, A.; Newsom, C D . ; Lovaas, O.I.; and Koegel, R.L. Some motivational properties of sensory stimulation in psychotic children. Journal of Experimental Child Psychology, 24:312-323, 1977.

Ritvo, E. Annotation. Biochemical studies of children with the syn- dromes of autism, childhood schizophrenia and related devel- opmental disabilities: A review. Journal of Child Psychology and Psy- chiatry, 18:373-379, 1977.

Ritvo, E.; Cantwell, D.; Johnson, E.; Clements, M.; Benbrook, F.; Slagle, S.; Kelley, P.; and Ritz, M. Social class factor in autism. Jour- nal of Autism and Childhood Schizo- phrenia, 1:297-310, 1971a.

Ritvo, E., and Freeman, B.J. Na- tional Society for Autistic Children definition of the syndrome of au- tism. Journal of Autism and Child- hood Schizophrenia, 8:162-169, 1978.

Ritvo, E.; Yuwiler, A.; Geller, E.; Kales, A.; Rashkis, S.; Schicor, A.; Plotkin, S.; Axelrod, R.; and How- ard, C. Effects of L-dopa in autism. Journal of Autism and Childhood Schizophrenia, 1:190-205, 1971b.

Ritvo, E.; Yuwiler, A.; Geller, E.; Ornitz, E.M.; Saeger, K.; and Plotkin, S. Increased blood seroto- nin and platelets in early infantile autism. Archives of General Psychia- try, 23:566-572, 1970.

Rivinus, T.M.; Jamison, D.L.; and Graham, P.J. Childhood organic neurological disease presenting as psychiatric disorder. Archives of Disease in Childhood, 50:115-119, 1975.

Rock, N. Childhood psychosis and long-term chemo- and psychother-

448 SCHIZOPHRENIA BULLETIN

apy. Diseases of the Nervous System, 35:303-308, 1974.

Roff, J.D.; Knight, R.; and Wert- heim, E. A factor-analytic study of childhood symptoms antecedent to schizophrenia. Journal of Abnormal Psychology, 85:543-549, 1976.

Romanczyk, R.G.; Diament, C ; Goren, E.R.; Trunell, G.; and Harris, S.L. Increasing isolate and social play in severely disturbed children: Intervention and post- intervention effectiveness. Journal of Autism and Childhood Schizophre- nia, 5:57-70, 1975.

Rosenbaum, M.S., and Breiling, J. The development and functional control of reading-comprehension behavior. Journal of Applied Behav- ior Analysis, 9:323-33*3, 1976.

Rosenberger, P.B., and Nigam, M.P. "Experimental Phenylalani- nemia and Learning Rate." Pres- ented at meeting of Child Neurol- ogy Society, Dartmouth College, New Hampshire, September 13-15, 1979.

Roth, J.A.; Young, J.G.; and Cohen, D.J. Platelet monoamine oxidase activity in children and ad- olescents. Life Sciences, 18:919-924, 1976.

Russell, A.T.; Cantwell, D.P.; Mattison, R.; and Will, L. A com- parison of DSM-II and DSM-III in the diagnosis of childhood psychi- atric disorders. III. Multiaxial fea- t u r e s . Archives of General Psychia- try, 36:1223-1226, 1979.

Russo, D.C.; Koegel, R.L.; and Lovaas, O.I. A comparison of human and automated instruction of autistic children. Journal of Ab- normal Child Psychology, 6:189-201, 1978.

Ruttenberg, B. A psychoanalytic understanding of infantile autism and its treatment. In: Churchill,

D.; Alpern, G.; and DeMyer, M.K., eds. Infantile Autism: Pro- ceedings of the Indiana University Colloquium. Springfield, IL: Charles C Thomas, 1970. pp. 145-184.

Rutter, M. Concepts of autism: A review of research. Journal of Child Psychology and Psychiatry, 9:1-25, 1968.

Rutter, M. Childhood schizophre- nia reconsidered. Journal of Autism and Childhood Schizophrenia, 2:315- 337, 1972.

Rutter, M. Brain damage syndromes in childhood: Concepts and find- ings. Journal of Child Psychology and Psychiatry, 18:1-21, 1977.

Rutter, M. Diagnosis and defini- tion of childhood autism. Journal of Autism and Childhood Schizophrenia, 8:139-161, 1978.

Rutter, M. Language, cognition and autism. In: Katzman, R., ed. Congenital and Acquired Cognitive Disorders. Research Publication, The Association for Research in Nervous and Mental Disease, Vol. 57, 1979. pp. 247-264.

Rutter, M., and Bartak, L. Special educational treatment of autistic children: A comparative study— II. Follow-up findings and impli- cations for services. Journal of Child Psychology and Psychiatry, 14:241-270, 1973.

Rutter, M.; Bartak, L.; and New- man, S. Autism—A central disor- der of cognition and language? In: Rutter, M., ed. Infantile Autism: Concepts, Characteristics and Treat- ment. London: Churchill- Livingstone, 1971. pp. 148-171.

Rutter, M., and Lockyer, L. A five to fifteen year follow-up study of infantile psychosis. I. Description of the sample. British Journal of Psychiatry, 113:1169-1182, 1967.

Rutter, M.; Shaffer, D.; and Shepherd, M. A Multiaxial Classifi- cation of Child Psychiatric Disorders. Geneva: World Health Organiza- tion, 1975. Rutter, M., and Sussenwein, F. A developmental and behavioral ap- proach to the treatment of pre- school autistic children. Journal of Autism and Childhood Schizophrenia, 1:376-397, 1971.

Sailor, W., and Taman, T. Stimu- lus factors in the training of prepo- sitional usage in three autistic chil- d r e n . Journal of Applied Behavior Analysis, 5:183-192, 1972.

Salvin, A.; Routh, D.K.; Foster, R.E.; and Lovejoy, K.M. Acquisi- tion of modified American sign language by a mute autistic child. Journal of Autism and Childhood Schizophrenia, 7:359-371, 1977.

Sankar, D.V.S. Uptake of 5-hy- droxytryptamine by isolated plate- lets in childhood schizophrenia and autism. Neuropsychobiology, 3:234-239, 1977.

Schopler, E. The stress of autism as ethology. Journal of Autism and Childhood Schizophrenia, 4:193-196, 1974. Schopler, E. Toward reducing be- havior problems in autistic chil- d r e n . Journal of Autism and Child- hood Schizophrenia, 6:1-13, 1976.

Schopler, E.; Brehm, S.S.; Kins- bourne, M.; and Reichler, R.J. Ef- fect of treatment structure on de- velopment in autistic children. Archives of General Psychiatry, 24:415-421, 1971.

Schopler, E., and Loftin, J. Thought disorders in parents of psychotic children. A function of test anxiety. Archives of General Psychiatry, 20:174-181, 1969.

Schopler, E., and Reichler, R.J. Parents as co-therapists in the

VOL. 7, NO. 3, 1981 449

treatment of psychotic children. Journal of Autism and Childhood Schizophrenia, 1:87-102, 1971.

Schopler, E., and Reichler, R.J. How well do parents understand their own psychotic child? Journal of Autism and Childhood Schizophre- nia, 2:387-400, 1972.

Schopler, E.; Reichler, R.J.; De- Villis, R.F.; and Daly, K. Toward objective classification of child- hood autism: Childhood Autism Rating Scale (CARS). Journal of Au- tism and Developmental Disorders, 10:91-103, 1980.

Schreibman, L. Effects of within- stimulus and extra-stimulus prompting on discrimination learning in autistic children. Jour- nal of Applied Behavior Analysis, 8:91-112, 1975.

Schreibman, L., and Carr, E.G. Elimination of echolalic respond- ing to questions through the train- ing of a generalized verbal re- sponse. Journal of Applied Behavior Analysis, 11:453-463, 1978.

Schreibman, L.; Koegel, R.L.; and Craig, M.S. Reducing stimulus overselectivity in autistic children. Journal of Abnormal Child Psycholo- gy, 5:425-436, 1977.

Shapiro, T. The quest for a linguis- tic model to study the speech of autistic children: Studies on echo- ing. Journal of the American Academy of Child Psychiatry, 16:608-619, 1977.

Shapiro, T.; Chiarandini, I.; and Fish, B. Thirty severely disturbed children: Evaluation of their lan- guage development for classifica- tion and prognosis. Archives of General Psychiatry, 30:819-825, 1974.

Shapiro, T., and Huebner, H.F. Speech patterns of five psychotic children now in adolescence. Jour-

nal of the American Academy of Child Psychiatry, 15:278-293, 1976.

Shapiro, T., and Kapit, R. Linguis- tic negation in autistic and normal children. Journal of Psycholinguistic Research, 7:337-351, 1978.

Shapiro, T., nnd Lucy, P. Echoing in autistic children: A chronomet- ric study of semantic processing. Journal of Child Psychology and Psy- chiatry, 19:373-378, 1978.

Shapiro, T.; Roberts, A.; and Fish, B. Imitation and echoing in young schizophrenic children. Journal of the American Academy of Psychiatry, 9:548-567, 1970.

Sherman, T.W., and Webster, C D . The effects of stimulus- fading on acquisition of a visual position discrimination in autistic, retarded, and normal children. Journal of Autism and Childhood Schizophrenia, 4:301-312, 1974.

Silberg, J.L. The development of pronoun usage in the psychotic child. Journal of Autism and Child- hood Schizophrenia, 8:413-425, 1978.

Simmons, J.Q., and Baltaxe, C. Language patterns of adolescent autistics. Journal of Autism and Childhood Schizophrenia, 5:333-351, 1975.

Simmons, J.Q.; Benor, D.; and Daniel, D. The variable effects of LSD-25 on the behavior of a heter- ogeneous group of childhood schizophrenics. Behavioral Neuro- psychiatry, 4:10-16, 1972.

Simon, N. Echolalic speech in childhood autism: Consideration of possible underlying loci of brain damage. Archives of General Psychi- atry, 32:1439-1446, 1975.

Singh, N.N. Reprogramming the social environment of an autistic child. New Zealand Medical Journal, 87:135-138, 1978.

Small, J.G. Psychiatric disorders and EEG. In: Textbook of Electroen- cephalography. Baltimore: Urban and Schwarzenberg, Inc., in press.

Solnick, J.V.; Rincover, A.; and Peterson, C.R. Some determinants of the reinforcing and punishing effects of timeout. Journal of Ap- plied Behavior Analysis, 10:415-424, 1977.

Spence, M.A. Genetic studies. In: Ritvo, E., ed. Autism: Diagnosis, Current Research and Management. New York: Halstead/Wiley, 1976. pp. 169-174.

Spence, M.A.; Simmons, J.Q.; Brown, N.A.; and Wikler, L. Sex ratios in families of autistic chil- dren. American Journal of Mental Deficiency, 77:405-407, 1973.

Steg, J.P., and Rapoport, J.L. Minor physical anomalies in nor- mal, neurotic, learning disabled, and severely disturbed children. Journal of Autism and Childhood Schizophrenia, 5:299-307, 1975.

Stevens-Long, J., and Rasmussen, M. The acquisition of simple and compound sentence structure in an autistic child. Journal of Applied Behavior Analysis, 7:473-479, 1974.

Strain, P.S.; Kerr, M.M.; and Rag- land, E.U. Effects of peer-medi- ated social initiations and prompting/reinforcement proce- dures on the social behavior of au- tistic children. Journal of Autism and Developmental Disorders, 9:41- 54, 1979.

Stubbs, E.G., and Magenis, R.E. HLA and autism. Journal of Autism and Developmental Disorders, 10:15- 19, 1980.

Sverd, J.; Kupietz, S.; Winsberg, B.; Hurwic, M.; and Becker, L. Ef- fects of L-5-hydroxytryptophan in autistic children. Journal of Autism

450 SCHIZOPHRENIA BULLETIN

and Childhood Schizophrenia, 8:171- 180, 1978.

Szurek, S.A. Playfulness, creativ- ity and schisis. In: Szurek, S.A., and Berlin, I.N., eds. Clinical Stud- ies in Childhood Psychoses. New York: Brunner/Mazel, 1973. pp. 10-28.

Takahashi, S.; Kanai, H.; and Miyamoto, Y. Reassessment of ele- vated serotonin levels in blood platelets in early infantile autism. journal of Autism and Childhood Schizophrenia, 6:317-326, 1976.

Takahashi, S.; Kanai, H.; and Miyamoto, Y. Monoamine oxidase activity in blood platelets from au- tistic children. Folia Psychiatrica et Neurologica Japonica, 31:597-603, 1977.

Tinbergen, E.A., and Tinbergen, N. The aetiology of childhood au- tism: A criticism of the Tinbergens' theory: A rejoinder. Psychological Medicine, 6:545-549, 1976.

Tinbergen, N. Ethology and stress diseases. Science, 185:20-23, 1974.

Towbin, A. Central nervous sys- tem damage in the human fetus and newborn infant. American Journal of Diseases of Children, 119:529-542, 1970.

Treffert, D.A. Epidemiology of in- fantile autism. Archives of General Psychiatry, 22:431-438, 1970.

Treffert, D.A.; Me Andrew, J.B.; and Dreifuerst, P. An inpatient treatment program and outcome for 57 autistic and schizophrenic children. Journal of Autism and Childhood Schizophrenia, 3:138-153, 1973.

Trunnell, T.L. Thought disturb- ance in schizophrenia: Replication study utilizing Piaget's theories. Archives of General Psychiatry, 13:9-18, 1965.

Tymchuk, A.J.; Simmons, J.Q.; and Neafsey, S. Intellectual char- acteristics of adolescent childhood psychotics with high verbal ability. Journal of Mental Deficiency Re- search, 21:133-138, 1977.

Van Krevelen, D.A. Early infantile autism and autistic psychopathy. Journal of Autism and Childhood Schizophrenia, 1:82-86, 1971.

Waizer, J.; Polizos, P.; Hoffman, S.; Engelhardt, D.; and Margolis, R. A single-blind evaluation of thiothixene with outpatient schiz- ophrenic children. Journal of Au- tism and Childhood Schizophrenia, 2:378-386, 1972.

Walker, H.A. A dermatoglyphic study of autistic patients. Journal of Autism and Childhood Schizophrenia, 7:11-21, 1977a.

Walker, H.A. Incidence of minor physical anomaly in autism. Jour- nal of Autism and Childhood Schizo- phrenia, 7:165-176, 1977b.

Walker, H.A., and Birch, H.G. Lateral preference and right-left awareness in schizophrenic chil- dren. The Journal of Nervous and Mental Disease, 151:341-351, 1970.

Walker, H.A., and Birch, H.G. In- tellectual patterning in schizo- phrenic children. Journal of Autism and Childhood Schizophrenia, 4:143- 161, 1974.

Ward, A.J. The use of structural therapy in the treatment of autistic children. Psychotherapy: Theory, Research and Practice, 9:46-50, 1972. Ward, A.J. Early childhood autism and structural therapy: Outcome after 3 years. Journal of Consulting and Clinical Psychology, 46:586-587, 1978.

Watt, N.F. Patterns of childhood social development in adult schiz- ophrenics. Archives of General Psy- chiatry, 35:160-165, 1978.

Webster, C D . ; McPherson, H.; Sloman, L.; Evans, M.A.; and Kuchar, E. Communicating with an autistic boy by gestures. Journal of Autism and Childhood Schizophre- nia, 3:337-346, 1973.

Wells, K.C.; Forehand, R.; and Hickey, K. Effects of a verbal warning and overcorrection on stereotyped and appropriate be- haviors. Journal of Abnormal Child Psychology, 5:387-403, 1977.

Wells, K.C.; Forehand, R.; Hickey, K.; and Green, K.D. Effects of a procedure derived from the over- correction principle on manipu- lated and nonmanipulated behav- iors. Journal of Applied Behavior Analysis, 10:679-687, 1977.

Wenar, C , and Ruttenberg, B.A. The use of BRIAC for evaluating therapeutic effectiveness. Journal of Autism and Childhood Schizophre- nia, 6:175-191, 1976.

Widelitz, M.M., and Feldman, W. Pink spot in childhood schizo- phrenia. Behavioral Neuropsychia- try, 1:29-30, 1969.

Wilhelm, H., and Lovaas, O.I. Stimulus overselectivity: A com- mon feature in autism and mental retardation. American Journal of Mental Deficiency, 81:26-31, 1976.

Williams, S., and Harper, J. A study of aetiological factors at crit- ical periods of development in au- tistic children. Australian and New Zealand Journal of Psychiatry, 7:163- 168, 1973.

Wing, L. The handicaps of autistic children—A comparative study. Journal of Child Psychology and Psy- chiatry, 10:1-40, 1969.

Wing, L. Perceptual and language development in autistic children: A comparative study. In: Rutter, M., ed. Infantile Autism: Concepts,

VOL. 7, NO. 3, 1981 451

Characteristics and Treatment. Lon- don: Churchill, 1971. pp. 173-197.

Wing, L. The current status of childhood autism. Psychological Medicine, 9:9-12, 1979.

Wing, L. Childhood autism and * social class: A question of selec- tion? British Journal of Psychiatry, 137:410-417, 1980.

Wing, L., and Gould, J. Systematic recording of behavior and skills of retarded and psychotic children. Journal of Autism and Childhood Schizophrenia, 8:79-97, 1978.

Wing, L.; Gould, J.; Yeates, S.R.; and Brierley, L.M. Symbolic play in severely mentally retarded and in autistic children. Journal of Child Psychology and Psychiatry, 18:167- 178, 1977.

Wing, L., and Ricks, D.M. The ae- tiology of childhood autism: A crit- icism of the Tinbergens' ethologi- cal theory. Psychological Medicine, 6:533-543, 1976.

Wing, L.; Yeates, S.R.; Brierley, L.M.; and Gould, J. The preva- lence of early childhood autism: Comparison of administrative and epidemiological studies. Psycho- logical Medicine, 6:89-100, 1976.

Wolf, E.G.; Wenar, C ; and Rut- tenberg, B.A. A comparison of personality variables in autistic and mentally retarded children. Journal of\Autism and Childhood Schizophrenia, 2:92-108, 1972.

Wolff, S., and Barlow, A. Schizoid personality in childhood: A com- parative study of schizoid, autistic and normal children. Journal of

Child Psychology and Psychiatry, 20:29-46, 1979.

Wolff, W.M., and Morris, L.A. In- tellectual and personality charac- teristics of parents of autistic chil- d r e n . Journal of Abnormal Psychology, 77:155-161, 1971.

Yamazaki, K.; Saito, Y.; Okada, F.; Fujieda, T.; and Yamashita, I. An application of neuroendocrin- ological studies in autistic children and Heller's syndrome. Journal of Autism and Childhood Schizophrenia, 5:323-332, 1975.

Young, J.G.; Caparulo, B.K.; Shaywitz, B.A.; Johnson, W.T.; and Cohen, D.J. Childhood au- tism. Cerebro-spinal fluid exami- nation and immunoglobulin levels. Journal of the American Academy of Child Psychiatry, 16:174-179, 1977. Young, J.G.; Cohen, D.J.; Brown, S.L.; and Caparulo, B.K. De- creased urinary free catechol- amines in childhood autism. Jour- nal of the American Academy of Child Psychiatry, 17:671-678, 1978.

Young, J.G.; Kyprie, R.M.; Ross, N.T.; and Cohen, D.J. Serum dopamine-beta-hydroxylase activi- ty: Clinical applications in child psychiatry. Journal of Autism and Developmental Disorders, 10:1-14, 1980. Yuwiler, A.; Ritvo, E.; Bald, D.; Kipper, D.; and Koper, A. Exami- nation of circadian rhythmicity of blood serotonin and platelets in autistic and nonautistic children. Journal of Autism and Childhood Schizophrenia, 1:421-435, 1971. Yuwiler, A.; Ritvo, E.R.; Geller, E.; Glousman, R.; Schneiderman,

G.; and Matsuno, D. Uptake and efflux of serotonin from platelets of autistic and nonautistic chil- d r e n . Journal of Autism and Child- hood Schizophrenia, 5:83-98, 1975.

Zarcone, V.; Kales, A.; Scharf, M.; Tan, T.; Simmons, J.; and Dement, W. Repeated oral ingestion of 5-hydroxytryptophan: The effect on behavior and sleep process in two schizophrenic children. Ar- chives of General Psychiatry, 28:843- 846, 1973.

Zifferblatt, S.M.; Burton, S.D; Horner, R.; and White, T. Estab- lishing generalization effects among autistic children. Journal of Autism and Childhood Schizophrenia, 7:337-347, 1977.

Acknowledgments

The authors thank Sandra Barton, Phyllis Bramer, Janice Hartman, and Deena Spencer for their tech- nical assistance during all stages of the preparation of this review.

The Authors

Marian K. DeMyer, M.D., is Pro- fessor of Psychiatry and Joseph N. Hingtgen, Ph.D., is Professor of Psychology and Neurobiology in the Department of Psychiatry and the Institute of Psychiatric Re- search, Indiana University School of Medicine, Indianapolis, IN. Roger K. Jackson, M.D., is a senior resident in psychiatry at the Indiana University School of Med- icine.