reviews

1993, 26, 353-360

EXTINCTION OF SELF-INJURIOUS ESCAPE BEHAVIOR WITH AND WITHOUT INSTRUCTIONAL FADING

JENNIFER R. ZARCONE AND BRLAN A. IWATA THE UNIVERSITY OF FLORIDA

TIMOTHY R. VoLLmER LOUISIANA STATE UNIVERSITY

SUNEETA JAGTIANI THE KENNEDY INSTITUTE, BALTIMORE, MARYLAND

AND

RICHARD G. SMITH AND JODI L. MAzAuiSKI THE UNIVERSITY OF FLORIDA

Three individuals with developmental disabilities participated in a study of the treatment of self- injurious behavior (SIB) maintained by negative reinforcement (escape from educational tasks). Treatment was implemented in a multiple baseline design across subjects, in which two treatments were compared in a multielement format. Both treatment conditions induded an escape-extinction component in which SIB no longer produced escape. One of the conditions also induded a fading component in which the frequency of instructions was initially reduced to zero and then was gradually faded back in across sessions until the instructional rate matched that of the original baseline. Results indicated that extinction alone reduced SIB to the end-of-treatment criterion in fewer sessions than did extinction plus fading for all 3 subjects. For 2 of the 3 subjects, however, there was an initial increase in the frequency of SIB at the outset of treatment with extinction (an extinction burst) that was not observed when extinction was combined with the fading component. DESCRIPTORS: extinction, functional analysis, negative reinforcement, self-injurious behavior,

stimulus fading

Research has shown that self-injurious behavior (SIB) can function as an escape response when it terminates certain ongoing activities such as aca- demic or work tasks (e.g., Iwata, Dorsey, Slifer, Bauman, & Richman, 1982; Steege, Wacker, Berg, Cigrand, & Cooper, 1989). One approach to re- ducing the frequency of such behavior consists of presenting the activity while preventing escape; that is, withholding negative reinforcement. This pro- cedure has been described as "escape" extinction and has been shown to be an effective treatment

This research was supported by a grant from the Devel- opmental Disabilities Planning Council. We appreciate the valuable assistance of MartinaJonak, Peter Andree, and other staff members and volunteers.

Reprints may be obtained from Brian Iwata, Psychology Department, University of Florida, Gainesville, Florida 32611.

for escape or avoidance behavior (Heidom & Jen- sen, 1984; Repp, Felce, & Barton, 1988; Steege et al., 1989). For example, Iwata, Pace, Cowdery, Kalsher, and Cataldo (1990) implemented escape extinction with 6 individuals during training ses- sions. Instructions were delivered at regular inter- vals, and trials were not terminated if a subject exhibited SIB; instead, the subject was guided through the task and the next trial was presented. Thus, SIB did not produce escape from educational tasks and, as a result, was extinguished in all sub- jects.

One possible limitation of extinction is a tem- porary increase in responding during the initial stages of treatment. For example, some of the subjects in the studies by Repp et al. (1988) and Iwata et al. (1990) showed a "burst" of SIB when escape ex-

353

NUMBER3 (FALL 1993)JOURNAL OF APPLIED BEHAVIOR ANALYSIS

JENNIFER R. ZARCONE et al.

tinction was first implemented. This potential prob- lem can make extinction procedures for SIB or aggression difficult from a practical standpoint be- cause an increase in the behavior can be dangerous to either the individual or those in the individual's immediate environment. Thus, to the extent that escape extinction produces a typical extinction burst, its therapeutic use may be limited to individuals whose behavior problems do not pose severe risk.

Pace, Iwata, Cowdery, Andree, and McIntyre (1993) attempted to prevent the occurrence of ex- tinction bursts by using a stimulus-fading proce- dure in conjunction with escape extinction while treating the SIB of 3 individuals. During the first few treatment sessions, no instructional trials were presented. Subsequently, the rate of instructions was increased gradually (i.e., faded into the training sessions) until it eventually matched that of the original baseline. This procedure, combined with escape extinction, resulted in large and immediate reductions in SIB that were maintained throughout treatment. A component analysis conducted with 1 of the 3 subjects showed that SIB increased above its baseline level when extinction was implemented without the fading procedure.

Although the results reported by Pace et al. (1993) are quite promising, they must be consid- ered tentative because the effects of extinction plus fading were compared to those of extinction alone with only 1 subject and only during the initial treatment sessions. Thus, the extent to which an extinction burst would have occurred for 2 of the 3 subjects was not determined. Furthermore, even if extinction resulted in temporary increases in all subjects' SIB, it is possible that treatment may have proceeded more rapidly in the extinction-alone con- dition (which did not contain a protracted fading procedure). The purpose of this study was to extend the

findings ofPace et al. (1993) by using multielement designs to compare the therapeutic effects of ex- tinction with and without instructional fading on SIB maintained by escape. In addition, the pro- cedures were continued long enough to determine

the length of time (number of sessions) required to reach a prearranged end-of-treatment criterion.

METHOD

Subjects and Setting Three adult women living in a state residential

facility participated. All of them were diagnosed with profound mental retardation. Lisa was a 39- year-old woman with Down syndrome whose SIB consisted of hitting her face in the area of her cheeks and temples. Her SIB posed a moderate risk of tissue damage and significantly interfered with her daily training. Lisa was ambulatory and had no sensory or motor impairments. She followed simple (one-step) instructions but had no expressive lan- guage. Karen was a 26-year-old woman whose primary forms of SIB were hand biting and face slapping. She also hit other parts of her body or stationary objects and frequently engaged in dis- ruptive behavior (e.g., throwing task items) and aggressive behavior toward staff. Karen's SIB posed a moderate risk of tissue damage, particularly where she bit her hand. Karen was ambulatory and had no sensory or motor impairments. She followed most instructions, but her expressive skills were limited to a few utterances. Diane was a 32-year- old woman with Down syndrome. Her SIB con- sisted of head banging against stationary objects and hitting her head with her hand, fist, knee, or foot. Diane's SIB posed a severe risk of tissue dam- age, and she wore a protective helmet at all times except during meals and bathing. Diane had a cataract in one eye, but she was fully sighted in the other eye and had no motor impairments. She had good receptive language (she could follow two-step instructions) and an expressive vocabulary that con- sisted ofapproximately a dozen manual signs. None of the subjects received psychotropic medication during the course of the study.

The study was conducted on the grounds of the facility at a day program for the assessment and treatment of SIB. Two to five sessions were run individually with each subject 5 days per week in therapy rooms (approximately 4 m by 6 m or 7

354

EXTINCTION AND FADING

m by 13 in). Sessions lasted for 15 min, separated by breaks lasting 10 to 15 min.

Response Measurement and Reliability

Self-injurious responses were defined as follows: face/head hitting-audible contact of a hand, knee, or leg against any part of the face or head; slap- ping-audible contact of any part of the body against another part of the body (other than the face or head) or against a stationary object (e.g., furniture, floor, or wall of therapy room); head banging-audible contact of the head against a stationary object; and hand biting-closure of the teeth on any part of the skin from fingertips to wrist. Data were also collected on disruption and aggression (for Karen), compliance with instruc- tions, and the rate at which instructions were pre- sented by the therapist. An observer was present during each session and recorded subject and ex- perimenter behavior on a hand-held computer (As- sistant, Model A 102) during continuous 10-s in- tervals. Session data were converted to responses per minute (SIB, aggression, disruption, and in- structions) and percentage compliance (percentage of instructions with which a subject complied).

Interobserver agreement was assessed by having a second observer simultaneously but independently collect data during 29% of the sessions of the func- tional analysis (for Karen and Lisa), 29% of all baseline sessions, and 25% of all treatment sessions. Agreement percentages were calculated based on interval-by-interval comparisons of the observers' records, in which the smaller number of responses in each interval was divided by the larger number of responses. These fractions were then summed across all intervals and divided by the total number of intervals in the session to get the percentage agreement between the two observers. Mean overall agreement scores and ranges for SIB during as- sessment (for Karen and Lisa), baseline, and treat- ment, respectively, were as follows: Karen-97.3% (91.4% to 100%), 91% (89.6% to 93%), 98.1% (77.6% to 100%); Diane-89.8% (85.1% to 91.7%) and 99.4% (96.5% to 100%); Lisa-

99.4% (97% to 100%), 95.6% (91.0% to 98.8%), and 98.6% (93.8% to 100%).

Experimental Sequence and Designs

The first phase of the study consisted of a func- tional analysis assessment, in which a series of con- ditions was presented in a multielement format (Sidman, 1960; Ulman & Sulzer-Azaroff, 1975). The purpose of this analysis was to identify the variables maintaining Lisa's and Karen's SIB. Di- ane did not participate in this phase of the study. A previous functional analysis revealed that Diane's SIB was maintained primarily by attention. Sub- sequently, the behavior was treated successfully with a differential reinforcement procedure that was un- related to the present experiment and is not de- scribed here. Informal follow-up observations at Diane's residence indicated that she continued to exhibit SIB in several isolated situations. Before undertaking a complete repetition of her functional analysis, additional descriptive data were collected and indicated that SIB occurred only following instructions requiring physical activity (e.g., "walk over here"). Based on these data, we suspected that Diane's SIB was maintained by escape. This con- clusion was verified when a baseline condition was initiated; therefore, Diane was induded only in the treatment phase of this study.

Following completion of the assessment, baseline data were collected for all 3 subjects. Treatment was then introduced according to a multiple base- line across subjects design. For each subject, two treatment conditions were presented in a multiele- ment format. One treatment consisted of extinction of escape behavior; the other consisted of escape extinction plus demand-frequency fading. Each phase is described below.

Functional Analysis Lisa and Karen were exposed to four assessment

conditions. A brief description of each condition is provided here; complete details can be found in Iwata et al. (1982). During the attention condition, the subject was in a therapy room with a variety

355

JENNIFER R. ZARCONE et al.

of leisure materials available. At the beginning of the session, the experimenter entered the room, informed the subject, "I will be here if you need me," and then ignored the subject. Contingent on the occurrence of SIB, the experimenter approached the subject, provided attention in the form of con- cern and disapproval of the behavior (e.g., "stop that, you'll hurt yourself'), and briefly interrupted the SIB through response blocking. During the demand condition, the experimenter presented ac- ademic or physical tasks at the rate of one every 30 s. The experimenter delivered praise and phys- ical contact (pats on the back) when the subject complied with the task. If the subject exhibited SIB during any part of the instructional sequence, the experimenter terminated the trial and imple- mented a time-out until the next scheduled instruc- tion. During the alone condition, the subject was in the therapy room alone without any leisure ma- terials. The final condition, play, served as a control. In this condition, leisure materials were made avail- able, and the experimenter provided attention to the subject every 30 s. Any SIB that occurred during the session was ignored.

Treatment Baseline. Procedures in effect during baseline

were identical to those in the demand condition of the functional analysis. The experimenter presented learning trials once every 30 s and delivered praise and physical contact contingent on compliance. If compliance was not initiated within 5 s of the initial instruction, the experimenter modeled the response and, if necessary, provided physical guidance 5 s later. Trials were terminated contingent on the oc- currence of SIB at any time during the instructional sequence. The tasks required moderate physical ac- tivity (e.g., "stand up," "come over here," "put on your shoe"), were similar to those found in the subjects' individual educational plans, and consis- tently produced SIB during the functional analysis assessment or during informal sessions conducted at the beginning of the study. For each subject, baseline sessions were conducted by two experi- menters who later would be paired with the dif-

ferent treatment procedures. The order of sessions throughout baseline and treatment was semiran- dom, with no more than three consecutive sessions presented by one experimenter.

Escape extinction. Treatment was implemented in a manner identical to that described by Iwata et al. (1990) and Pace et al. (1993). All aspects of this condition were the same as in baseline, except that SIB did not produce escape (termination of a trial). If the subject emitted SIB at any point during the trial, physical guidance was provided imme- diately, and the session continued accordingly.

Escape extinction plus instructional fading. Consequences for compliance, noncompliance, and SIB were the same as those in the extinction con- dition. In addition, all instructions were eliminated from the initial treatment session and subsequently were faded in across sessions based on a subject's observed rates of SIB. The frequency of instruc- tional trials per session was increased by one if the subject's SIB was at or below 0.5 responses per minute during a previous session (this rule was modified after Session 39 for Diane; instructions were increased by two per session if her rate of SIB was at or below 0.5 per minute). The eventual goal was to increase the rate of instructions until their frequency matched that of the baseline and ex- tinction conditions (two per minute). The extinc- tion-plus-fading procedure was thus very similar to that used by Pace et al. (1993), with the additional feature of explicit fading criteria.

Treatment Effectiveness In order to evaluate the relative effects of the

two treatment procedures, a criterion was estab- lished prior to treatment to indicate when a treat- ment condition was considered to be effective. This criterion was set initially at or below 0.5 self-in- jurious responses per minute for five consecutive sessions, with an instructional rate of two per min- ute. Karen's criterion was later modified to one self- injurious response per minute for five sessions in a row. Given Karen's baseline rate of SIB and the rate of instructions, this criterion was determined to be a clinically significant reduction. When the

356

EXTINCTION AND FADING

16criterion was met in one treatment condition, that condition was discontinued.

RESULTS

Results of the functional analysis assessment for Karen and Lisa are shown in Figure 1. Both subjects exhibited the highest rates of SIB during the de- mand condition, indicating that their SIB was maintained primarily by escape from instructions. These data replicate the findings of previous re- search (e.g., Iwata et al., 1990; Steege et al., 1989) indicating that contingent task removal in the form of brieftime-out can serve as negative reinforcement for SIB.

Figure 2 shows the results obtained during the treatment comparison. Karen's data show a high rate of SIB during baseline. When extinction was introduced, her SIB increased above its baseline rate (i.e., there was an extinction burst). Although her rate of SIB subsequently decreased during ex- tinction, it remained somewhat variable, and a dose examination of her data revealed that a significant proportion of her remaining SIB occurred imme- diately following the completion of a trial. As a result, an additional contingency was added to the extinction condition after Session 91. If Karen ex- hibited SIB between trials, she was presented with a new trial. This contingency was associated with a further decrease in SIB, and she met her termi- nation criterion in this condition in 50 treatment sessions (after Session 109). During the initial ses- sions of the extinction-plus-fading condition, no extinction burst was evident; her rate of SIB de- creased immediately, and the frequency of instruc- tions was faded quickly. However, when the rate of instructions reached approximately 0.6 per min- ute (nine trials per session), Karen's SIB became more variable and the number of sessions at each instructional rate increased. For example, 27 ses- sions were conducted at the rate of 1.33 trials per minute (20 trials per session) before she met cri- terion to fade to the next instructional rate. Al- though Karen eventually met the criterion to in- crease the rate of instructions two more times, it

12

wI-

z

wCLL0) wU) On

a:

16

4-

Lisa

5V e1 1 2 25

5 1 0 1 5 20 25 30 35

SESSIONS Figure 1. Responses per minute of SIB across assessment

conditions.

appeared that the fading procedure did not have the expected effect in reducing her SIB. Thus, the extinction-plus-fading condition was discontinued after 133 sessions.

Although Diane's rate of SIB during baseline was very high during two initial sessions, the other 12 sessions of baseline stabilized at approximately three responses per minute. Following baseline, her rate of SIB rapidly decreased during the initial sessions of the extinction condition, and she sub- sequently met criterion for treatment completion after only nine extinction sessions. In the extinction- plus-fading condition, Diane's rate ofSIB decreased to zero during the first treatment session and re- mained low throughout the condition as the fre- quency of instructions was increased rapidly to two per minute. She completed treatment in this con- dition after 34 sessions. Thus, both treatments were effective in reducing Diane's SIB, and no burst was observed during the extinction condition.

Lisa's rate of SIB also was somewhat variable

357

JENNIFER R. ZARCONE et al.

w

0u

zcr 0CL

0c ct)w a:

BL TREATMENT (EXTINCTION vs EXTINCTION + FADING)

I| Karen

z

w 0L CO F

CO)

z

25 50 75 100 125 150 175 200

SESSIONS

Figure 2. Responses per minute of SIB (left axis) and instructional trials per minute during the extinction-plus-fading condition (right axis) during baseline and treatment. The arrow on Karen's graph indicates when contingent instructions were added in the extinction condition.

during initial baseline sessions, but it stabilized during the 15 sessions at approximately two re- sponses per minute. When extinction was intro- duced, there was a dramatic increase in SIB during the first two treatment sessions. Her rate of SIB eventually decreased in this condition, and treat- ment was completed when criterion was reached after a total of 22 sessions. During extinction plus

fading, Lisa's rate of SIB was immediately reduced at the outset oftreatment. As the rate of instructions increased, Lisa's rate of SIB continued to remain low, and treatment was completed after 54 sessions.

Following the completion of treatment, staff members at each subject's residence were trained to use escape extinction for SIB when conducting instructional programs. For Karen's and Lisa's staff

358

EXTINCTION AND FADING

members, training (which induded modeling, prompting, and feedback) continued until they could perform the procedures consistently. Staff members at Diane's residence were trained more informally. For all subjects, ongoing instructional programs were modified to indude an escape ex- tinction contingency for SIB. Follow-up observa- tions conducted over a 6-month period revealed that subjects' SIB remained low (i.e., at or below end-of-treatment levels).

DISCUSSION Results of this study extend the findings of Pace

et al. (1993) by providing a component analysis of the effects of escape extinction and instructional fading on the self-injurious behavior of 3 subjects. By examining extinction with and without fading in a multielement format, the relative contribution of fading was directly assessed. Results indicated that extinction alone was an effective treatment for SIB but that it was associated with an initial burst of responding for 2 of the 3 subjects (Karen and Lisa). No burst of SIB was observed, however, when instructional fading was used in conjunction with extinction; SIB decreased rapidly at the outset of treatment for all 3 subjects.

Although there was a higher initial rate of SIB in the extinction condition for Karen and Lisa, Karen's only successful treatment was extinction; she never completed the extinction-plus-fading con- dition. For Diane and Lisa, treatment during ex- tinction plus fading required over twice as many sessions as the extinction-alone condition, even when Diane's fading procedure was accelerated by in- creasing instructions by two per minute. It is pos- sible that fading could have occurred even more rapidly for Diane and Lisa, resulting in fewer treat- ment sessions in the extinction-plus-fading condi- tion. Karen's data, however, suggest that an ac- celerated fading procedure would have been ineffective for her, given that the instructional rate never increased beyond 1.5 per minute. Thus, the rapidity with which fading can occur, as well as the final rate of instructions that can be achieved, may differ across individuals.

For the 2 subjects who successfully completed

both treatment conditions (Diane and Lisa), the number of sessions to the completion of treatment may not be the best index by which the two treat- ments can be compared. For both individuals, the total (cumulative) number ofresponses emitted was higher in the extinction condition than in the ex- tinction-plus-fading condition. For example, Lisa's data show that she met criterion in the extinction condition after only 22 treatment sessions, but dur- ing that time 995 responses occurred. During ex- tinction plus fading, only 268 responses occurred over 54 treatment sessions. Results were less pro- nounced for Diane. She exhibited 176 responses over 28 treatment sessions during extinction, com- pared to 146 responses over 34 treatment sessions during extinction plus fading. The higher total number of responses during extinction is directly related to the burst of SIB at the outset of that treatment condition.

The effects of extinction on Karen's SIB were somewhat confounded. The introduction of con- tingent instructions after 41 sessions in the extinc- tion condition made it impossible to determine whether the punishment contingency or the addi- tional extinction trials finally reduced her SIB. It was not possible to add this contingency to the extinction-plus-fading condition in an attempt to equate the procedures because the fading aspect would have been eliminated. Thus, when extinction plus fading was not successful after over 130 treat- ment sessions, treatment was terminated. Never- theless, the two procedures had very different effects on SIB at the outset of treatment, making Karen's data valid for the purposes of determining whether fading could eliminate an extinction burst. The results of this study indicate that there are

potential advantages and disadvantages to both treatment procedures. Although SIB may be elim- inated more quickly when extinction alone is used, the rate of SIB may increase beyond acceptable limits before it is reduced. When instructional fad- ing is implemented in conjunction with extinction, there seems to be an immediate reduction in SIB at the outset of treatment. However, the combined treatment procedure was not effective for 1 subject and required many sessions for the other 2. Thus,

359

360 JENNIFER R. ZARCONE et al.

potential severity of the behavior and length of time available for treatment are two factors that may determine whether or not fading procedures should be used in conjunction with extinction.

Future research on the treatment ofescape-main- tained SIB might focus on further development of procedures based on stimulus fading. In the present study, fading occurred along the dimension of fre- quency (instructions per minute). Other research (e.g., Weeks & Gaylord-Ross, 1981) has shown that task difficulty is another variable that may increase escape behavior. Thus, it should be possible to develop assessment procedures to determine the aversive characteristics of instructional tasks (fre- quency, difficulty, duration, etc.) for each individ- ual and to use the results as the basis for developing fading procedures along multiple dimensions.

Another interesting extension of the present study would be an evaluation of instructional fading as the sole form of intervention. If the rate of SIB can be reduced to near zero at the outset of treatment merely by altering instructional variables, perhaps fading procedures can be implemented without ex- tinction. This approach to treatment would elim- inate any possible negative side effects associated with the use of extinction, although the rate of fading may be exceedingly slow, and any escape behavior that occurred during treatment would nec- essarily be reinforced. The first limitation may be merely a practical one. However, the second may be critical in light of the results of other recent studies (Wacker et al., 1990; Zarcone, Iwata, Hughes, & Vollmer, 1993) indicating that extinc- tion is an important and perhaps necessary com- ponent of treatment procedures based on the re- inforcement of alternative behaviors, such as functional communication training (Carr & Du- rand, 1985) and high-probability request sequences (Mace & Belfiore, 1990). Thus, although the re- sults of a number of studies, induding the present one, indicate that extinction alone can reduce the frequency of behavior disorders such as SIB, ad- ditional research is needed to identify the critical components of several combined treatment ap- proaches, induding those based on instructional manipulations such as fading.

REFERENCES

Carr, E. G., & Durand, V. M. (1985). Reducing behavior problems through functional communication training. Journal of Applied Behavior Analysis, 18, 111-126.

Heidorn, S. D., & Jensen, C. C. (1984). Generalization and maintenance of the reduction of self-injurious be- havior maintained by two types of reinforcement. Be- haviour Research and Therapy, 22, 581-586.

Iwata, B. A., Dorsey, M. F., Slifer, K. J., Bauman, K. E., & Richman, G. S. (1982). Toward a functional analysis of self-injury. Analysis and Intervention in Develop- mental Disabilities, 3, 1-20.

Iwata, B. A., Pace, G. M., Cowdery, G. E., Kalsher, M. J., & Cataldo, M. F. (1990). Experimental analysis and extinction of self-injurious escape behavior. Journal of Applied Behavior Analysis, 23, 11-27.

Mace, F. C., & Belfiore, P. (1990). Behavioral momentum in the treatment of escape-motivated stereotypy. Journal of Applied Behavior Analysis, 2, 507-514.

Pace, G. M., Iwata, B. A., Cowdery, G. E., Andree, P. J., & McIntyre, T. (1993). Stimulus (instructional) fading during extinction of self-injurious escape behavior. Jour- nal of Applied Behavior Analysis, 26, 205-212.

Repp, A. C., Felce, D., & Barton, L. E. (1988). Basing the treatment of stereotypic and self-injurious behaviors on hypotheses of their causes. Journal of Applied Be- havior Analysis, 21, 281-289.

Sidman, M. (1960). Tactics of scientific research. New York: Basic Books.

Steege, M. W., Wacker, D. P., Berg, W. K., Cigrand, K. C., & Cooper, L. J. (1989). The use of behavioral assessment to prescribe and evaluate treatments for se- verely handicapped children. Journal of Applied Be- havior Analysis, 22, 23-33.

Ulman, J. D., & Sulzer-Azaroff, B. (1975). Multielement baseline design in educational research. In E. Ramp & G. Semb (Eds.), Behavior analysis: Areas of research and application (pp. 359-376). Englewood Cliffs, NJ: Prentice-Hall.

Wacker, D. P., Steege, M. W., Northup, J., Sasso, G., Berg, W., Reimers, T., Cooper, L., Cigrand, K., & Donn, L. (1990). A component analysis of functional commu- nication training across three topographies of severe be- havior problems.Journal ofApplied Behavior Analysis, 23, 417-429.

Weeks, M., & Gaylord-Ross, R. (1981). Task difficulty and aberrant behavior in severely handicapped students. Journal of Applied Behavior Analysis, 14, 449-463.

Zarcone, J. R., Iwata, B. A., Hughes, C. E., & Vollmer, T. R. (1993). Momentum versus extinction effects in the treatment of self-injurious escape behavior. Journal of Applied Behavior Analysis, 26, 135-136.

Received January 19, 1993 Initial editorial decision March 8, 1993 Revision received April 19, 1993 Final acceptance April 19, 1993 Action Editor, Robert Horner