NURS 589 STATS

AJSLP

Research Note

Analysis of Story Recall in Military Veterans With and Without Mild Traumatic Brain Injury: Preliminary Results

Karen Gallaghera and Tamiko Azumaa

Purpose: The purpose of this study was to determine whether detailed analysis of story recall performance reveals significant differences between veterans with and without history of mild traumatic brain injury (mTBI). Method: Twenty-one military veterans participated, with 7 reporting history of mTBI. All participants were administered the Logical Memory I and II subtests from the Wechsler Memory Scale–Fourth Edition (Wechsler, 2009). Responses were scored for total correct ideas (TCI) and total words produced (TWP). Results: Although the groups did not differ in scaled scores, other measures did reveal significant differences. After a delay, the mTBI group showed a greater drop in TCI relative

to the control group. Additionally, the control group showed an increase in TWP when the recall was delayed versus immediate; a pattern not observed for the mTBI group. Conclusions: The mTBI and control groups did not significantly differ in scaled scores. However, group differences were observed in TCI and TWP. The findings suggest that, relative to the control group, the mTBI group were less successful in retrieving episodic information and eliciting self-cueing. Small sample size limited data interpretation, and larger sample sizes are needed to confirm the findings. The results indicate that veterans with mTBI may present with symptoms persisting beyond the acute state of the injury.

M ilitary veterans comprise a subgroup of individ- uals with a high incidence of traumatic brain injury (TBI) complicated by factors either not

seen, or seen at significantly lower rates, in the civilian population. In veterans, the largest proportion of brain injuries are classified as mild (mild traumatic brain injury [mTBI]; MacGregor, Morrison, Quinn, & Garlarneau, 2011; Terrio et al., 2009). Recently, there has been increased attention focused on mTBI and possible long-term conse- quences. Researchers have studied subpopulations of individ- uals at higher risk for brain injury, including professional athletes, child/adolescent athletes (e.g., Covassin, Moran, & Wilhelm, 2013; Strain et al., 2015; Tapper, Gonzalez, Roy, & Niechwiej-Szwedo, 2017), and military service members (e.g., Nelson et al., 2010; Rigg & Mooney, 2011; Terrio et al., 2009; Verfaellie, Lafleche, Spiro, & Bousquet, 2014).

Given the diverse nature of these higher risk groups, it can be assumed that there will be differences in the situational aspects surrounding the injury, the associated symptoms, and concomitant conditions. Thus, individuals may receive the same diagnosis, but appropriate assessment and treat- ment can differ across groups. Military veterans, in particu- lar, require additional consideration in the assessment of mTBI because of military culture issues that can influence the time frame of injury diagnosis (Rigg & Mooney, 2011). Additionally, service members and veterans with a history of mTBI are more likely than civilians to report unremitting cognitive symptoms. Approximately 5% to 20% of athletes with mTBI reportedly present with long-term symptoms of mTBI (Jotwani & Harmon, 2010) compared with 7.5% to 40% of veterans reporting chronic postconcussive symp- toms (Cooper et al., 2015). These differences contribute to the complexity of diagnosis and subsequent treatment planning for mTBI-related symptoms in veterans.

Service-Related Factors in Assessment and Treatment

Service-related factors can further complicate assess- ment of longer term cognitive deficits related to mTBI. For example, although delayed reporting of mTBI occurs

aDepartment of Speech and Hearing Science, Arizona State University, Tempe

Correspondence to Karen Gallagher: [email protected]

Editor: Margaret Blake Associate Editor: Carl Coelho

Received October 31, 2016 Revision received March 22, 2017 Accepted November 21, 2017 https://doi.org/10.1044/2017_AJSLP-16-0208 Publisher Note: This article is part of the Special Issue: Select Papers From the 46th Clinical Aphasiology Conference.

Disclosure: The authors have declared that no competing interests existed at the time of publication.

American Journal of Speech-Language Pathology • Vol. 27 • 485–494 • March 2018 • Copyright © 2018 American Speech-Language-Hearing Association 485

in civilian settings (Hoffer, 2015), delayed or lack of injury report is much more likely to occur for veterans/service members, especially in combat settings (Defense and Vet- erans Brain Injury Center, 2016). Veterans may be more inclined to delay or avoid diagnosis due to military culture issues, including the context of active combat situations, lack of observable injury, and the belief that seeking health care is a sign of weakness (Hyatt, Davis, & Barroso, 2014). This perception of weakness is particularly prevalent when the symptoms are not physically visible and may be per- ceived as mental health related (Cooper et al., 2015; Hoge et al., 2008). Additionally, military administrative factors can cause hesitation in the reporting of mTBI symptoms. With a reported injury, military protocols related to assess- ment and treatment could delay the stateside return date of the deployed service member, disrupt or cause failure in training, or potentially delay exiting service at the end of an enlistment (Rigg & Mooney, 2011).

Despite the incidence of unremitting symptoms in veterans, some researchers conclude that there are no long- term neuropsychological deficits associated with service- related mTBI (e.g., Hartikainen et al., 2010; Jak et al., 2015). Long-term symptom complaints have been attributed to posttraumatic stress disorder (PTSD) or other variables, such as reduced effort or mental health issues (Lange et al., 2012; Nelson et al., 2010; Verfaellie et al., 2014). Kontos et al. (2015) concluded that history of mTBI secondary to blast injury (common in combat environments) can impact cognitive functions, but the study participants were only followed for 20 days postinjury. Despite ongoing research, the issue of long-term neuropsychological consequences of mTBI in military veterans remains equivocal.

In veterans with mTBI, there are also complicating factors related to the service environments, specific assess- ments, and the transition from military to civilian life. Vet- erans may only become aware of subtle cognitive deficits arising from mTBI after they exit the highly structured environment of the military and transition to college or the civilian workforce (DiRamio et al., 2008). This transition can be intense, affecting all facets of life, such as social interactions, personal relationships, and vocational demands. It requires new learning and may exert increased cognitive demands resulting in a more pronounced presentation of subtle symptoms. Given the subtle nature of these deficits and their exacerbation in complex, high-demand environ- ments, existing standardized assessment protocols may lack the ecological validity to quantify memory deficits associ- ated with mTBI in any population.

Limitations in Traditional Neuropsychological Assessment of mTBI

Standard neuropsychological tests often assess mem- ory with simple tasks, such as digit span or word recall. Simple verbal memory tasks that involve rote memory lack ecological value and are often not sensitive to subtle mem- ory impairments arising from mTBI (e.g., Geary et al., 2011). Digit span, including digits forward and reversed,

has been broadly used in research as a measure of work- ing memory (Alloway et al., 2006). There is debate regard- ing the nature of these measures. Some researchers propose that digits forward and digits reversed rely on separate cog- nitive abilities and that combining task scores obscures criti- cal performance differences (Lezak et al., 2012; Reynolds, 1997). Other researchers argue that the digit span tasks tap into the same cognitive processes (Rosen & Engle, 1997; St. Claire-Thompson & Allen, 2013). Beyond digit span tasks, memory is often assessed using word list learning tasks, such as the Verbal Paired Associates I and II subtests in the Wechsler Memory Scale–Fourth Edition (WMS-IV; Wechsler, 2009) and California Verbal Learning Test (Delis, Kramer, Kaplan, & Ober, 1987). Although these tasks pro- vide more information regarding episodic and semantic mem- ory skills (Tayim, Flashman, Wright, Roth, & McAllister, 2016), they do not necessarily reflect real-world memory demands, such as active recall of personal events.

Story Retelling in Individuals With mTBI Relative to digit or word list recall, story retelling is

a complex task integrating multiple cognitive–linguistic skills. The person must remember not only specific details but also sequence and thematic information. The Logical Memory subtests of the WMS-IV (Wechsler, 2009) include both an immediate recall and a surprise delayed recall of the stories. The delayed recall allows the assessment of lon- ger term encoding and retrieval of complex narrative infor- mation. Story recall tasks seem analogous to the memory demands of academic and work environments, in which in- dividuals must frequently remember information from lec- tures or conversations. The Wechsler Memory Scale has frequently been used to assess veterans with TBI (Burriss et al., 2008; Ryan, Lee-Wilk, Kok, & Wilk, 2011). Using the Logical Memory subtest, Burriss et al. (2008) found that combat veterans with PTSD performed poorer on the delayed recall portion than combat veterans and noncombat veterans with no PTSD diagnosis. The researchers included depression as a covariate in their analyses, but they did not account for history of mTBI in their sample.

Although the Wechsler Memory Scale provides infor- mation regarding general memory abilities, standard scaled scores alone may not reveal important aspects of the mem- ory deficit. Standard raw scores are based on counts of informational unit chunks, which can include more than one specific detail. For example, informational chunks fre- quently combine numbers and units of measures (e.g., “four inches”), yet each chunk is counted as a single item for scor- ing. The count of informational chunks does not differenti- ate the recall of specific information (i.e., the person may recall the measurement of “inches” but fail to correctly re- call the number “four”). By merging meaningful elements in a story, relevant errors or inclusion of partial details may be missed in the calculation. Standard scoring also does not provide information regarding self-cueing strategies. Ana- lyzing the content of other utterances may provide insight into the individual’s episodic memory retrieval process.

486 American Journal of Speech-Language Pathology • Vol. 27 • 485–494 • March 2018

Including these additional measures in the scoring of recall responses may provide a more complete memory profile for veterans with mTBI.

Story retelling and narrative analysis have long been utilized to assess cognitive–linguistic deficits associated with moderate and severe TBI in the absence of aphasia (Chapman et al., 1992; Coelho, Grela, Corso, Gamble, & Feinn, 2005; Marini et al., 2011). Researchers report that microstructural aspects of narrative production are gener- ally commensurate with those produced by non-brain-injured peers, but global measures, including content information unit (CIU), efficiency measures, and story grammar analy- ses, are more sensitive to mTBI-related deficits (Carlomagno et al., 2011; Marini, Zettin, & Galetto, 2014; Matsuoka et al., 2012). Recently, Galetto, Andreetta, Zettin, and Marini (2013) compared 10 adults with mTBI to non-brain- injured peers across a variety of cognitive, linguistic, and narrative assessments. They found no grammatical, lexical, or phonological differences between the groups, but the mTBI group scored significantly poorer than controls in story grammar measures. Overall, the results suggest that narrative analysis is appropriate and sensitive to deficits related to mTBI severity. Because these studies focused on story grammar and CIU analysis of spontaneous discourse, picture-based stimuli were used to elicit the narratives. For studies that use single pictures to elicit spontaneous discourse, the resulting responses may not provide mean- ingful information regarding the individual’s episodic mem- ory. Using picture series or nonverbal story presentations can be effective in eliciting responses that reflect episodic memory processes, but such stimuli seem less naturalistic than verbal story presentation and recall.

Episodic Memory in Individuals With mTBI Story retelling tasks are used to assess verbal episodic

memory in a variety of populations. Davis, Alea, and Bluck (2015) examined accuracy in socially relevant story retell- ing, with productions coded for both gist and detail accu- racy. Their study focused on aging and gender effects, but other researchers similarly examined recall of narrative details in individuals with TBI. Kennedy and Nawrocki (2003) compared recall of details and main ideas in TBI survivors and healthy controls. In responding to gist and fact questions, the TBI group were less accurate in their recall than the control group. Successful strategic retrieval of complex information from episodic memory relies on higher order abilities that are associated with prefrontal areas, the same frontal areas and networks that are fre- quently damaged in TBI (Stuss, 2011). Episodic memory retrieval deficits are often observed in individuals with TBI, including those with mTBI (Miotto et al., 2010; Tayim et al., 2016). However, the deficits associated with mTBI may be difficult to detect using broad measures or stan- dardized test scores. A more refined examination may be required to assess subtle differences in overall productivity, efficiency, and recall strategies. Self-cueing strategies have been promoted to facilitate the recall of verbal information

and improve discourse production. For example, Antonucci (2009) discussed the effective use of semantic features analy- sis (SFA) as a self-cueing strategy in discourse. Wambaugh, Mauszycki, Cameron, Wright, and Nessler (2013) included a discourse task and CIU measurement in their treatment study on SFA in individuals with chronic aphasia. They specifically utilized a “mediating strategy” treatment fol- lowing the more traditional SFA treatment. Because the discourse sample was semantically unrelated to the trained SFA stimuli, the researchers observed that the SFA skills did not directly improve lexical retrieval but, rather, were utilized as a self-cueing strategy. Several participants in- creased efficiency in terms of words per minute without a corresponding increase in relevant content—a pattern indic- ative of self-cueing. Semantic self-cueing warrants further examination in the context of episodic memory given the connection between semantic processing and memory encod- ing and retrieval (Craik and Tulving 1975; Greenberg & Verfaellie, 2010; Nelson, Kitto, Galea, McEvoy, & Bruza, 2013). Similar to Wambaugh et al. (2013), the current study analyzed the number of correct ideas produced by partici- pants in the story retelling task. Additionally, an examina- tion of the total number of words produced during story recall may provide a measure of self-initiated, retrieval strat- egies. A participant who uses self-cueing strategies through- out recall should produce more words than a participant who stops initiating, or never initiates, self-cueing strategies.

Aims The primary purpose of this study was to determine

whether detailed analysis of story recall performance re- veals significantly poorer performance in military veterans with a history of mTBI compared with veterans without a history of mTBI. The specific aims and hypotheses of the study were as follows:

1. To examine whether standard scaled scores reveal sig- nificant differences between veterans with mTBI and control veterans. It was hypothesized that the stan- dard scaled scores of the Logical Memory 1 and 2 would not be sensitive to mTBI deficits, and no group differences would be observed in either the immediate or delayed recall conditions.

2. To determine whether more detailed analysis of cor- rect information produced would detect differences between the mTBI and control groups across recall conditions. It was hypothesized that group differ- ences would not be observed in the immediate recall condition, but veterans with mTBI would produce fewer correct idea units than control veterans in the delayed recall condition. This pattern would suggest that mTBI impacts longer term encoding of verbal information more than working memory storage.

3. To assess differences between the mTBI and control groups in word production across the immediate and delayed recall conditions. It was hypothesized that the groups would be similar in word production in

Gallagher & Azuma: Story Recall in Veterans With TBI 487

the immediate condition, but the control group will produce more words than the mTBI group in the de- layed condition. This difference in word production would be related to the increased use of self-cueing strategies in the control group relative to the mTBI group.

Method Participants

Twenty-one military veterans enrolled as undergradu- ate students at Arizona State University participated in this study. Participants were recruited via e-flyers sent through the ASU Pat Tillman Veterans Center. Veterans were all Post-9/11 GI Bill benefit eligible students, who were work- ing on their first college degrees. All participants completed an online survey that included questions on their educa- tional, military service, and medical history. As part of the questionnaire, participants were asked to provide a self- rating of their memory. The question stated: “Select the statement that best describes your OVERALL memory (your memory in your everyday life)”: 5 = My memory is much better than other people my age; 4 = My memory is somewhat better than other people my age; 3 = My mem- ory is about the same as other people my age; 2 = My memory is somewhat worse than other people my age; 1 = My mem- ory is much worse than other people my age. Self-ratings of 1 or 2 were considered indicative of memory difficulties.

Seven of the 21 participants reported history of mTBI, either diagnosed or undiagnosed. Undiagnosed mTBI was verified based on responses to questions regarding symptoms included on the Mayo Classification System for Traumatic Brain Injury Severity (Malec et al., 2007) and the American Congress of Rehabilitative Medicine. Specific symptom self- report related to a described blast injury or head trauma included loss of consciousness less than 30 min, posttrauma anterograde amnesia of less than 24 hr, or a combination of blurred vision, confusion, dazed, dizziness, nausea, or headache. No participant reported a history of stroke or other neurological impairment. All participants reported normal or corrected-to-normal vision and hearing. Demo- graphic information for veterans with and without his- tory of mTBI is shown in Table 1. The groups did not

significantly differ in age, t(19) = −1.31, p = .205, grade point average (t < 1), number of college credits earned (t < 1), or number of months between high school graduation/General Equivalency Diploma (GED) and enrollment in first college class (t < 1). Prior to testing, a certified speech-language pathologist judged articulatory accuracy, rate of speech, and vocal quality during conversational speech. All partici- pants were judged to have “no perceptible speech or vocal parameters indicative of a possible disorder.”

Procedure Logical Memory I and II

All participants were administered Stories B and C from the Logical Memory I (LM I) and Logical Memory II (LM II) subtests from the WMS-IV as part of a larger battery of cognitive testing. The entire cognitive testing ses- sion lasted approximately 90–120 min (duration was depen- dent on the number of requested breaks). For the LM I subtest (Immediate Recall), the participant was told two stories. Both stories were orally presented by the researcher with no visual cues. Immediately after each story, the par- ticipant was asked to recall the story in as much detail as possible, using as many of the same words as they could re- member. The participant was not told that they would be required to retell the story following a delay. For the LM II subtest (Delayed Recall), the participant was asked to recall each story again following a 20–30-min delay. Between the immediate and delayed recall, participants completed an unrelated computerized task involving math or spatial judg- ments or a pencil-and-paper survey. All recall responses were digitally recorded for transcription and later scoring. Two researchers independently transcribed the stories ver- batim, including all interjections, mazes, and comments. Discrepancies between the two transcriptions were resolved by a third transcriber.

Scoring Standard scaled scores reflected scoring that followed

the WMS-IV manual for conversion to scaled scores and were calculated by two independent scorers who were trained in WMS-IV procedures. Discrepancies were resolved by a third scorer trained in the procedures. The transcriptions were then scored for number of total correct ideas and total

Table 1. Demographic information for veteran control and mTBI groups (range in brackets).

Characteristic Control veterans (n = 14) Veterans with mTBI (n = 7)

Age in years (SD) 29.29 (3.2) 32.57 (8.4) Mean GPA (SD) 3.08 (0.57) [2.21–4.00] 3.08 (0.57) [2.50–4.00] Mean college credits earned (SD) 86.5 (56.8) [6–167] 92.6 (75.6) [0–196] No. of months before college enrollment (SD) 50.7 (32.6) [3–108] 42.4 (47.1) [0–120] No. of males (%) 7 (50.0%) 7 (100%) No. reporting PTSD (%) 3 (21.4%) 3 (42.9%) No. reporting depression (%) 4 (28.6%) 3 (42.9%) No. reporting memory problems (%) 5 (35.7%) 5 (71.4%)

Note. Number of months before college enrollment: number of months between high school graduation/General Equivalency Diploma (GED) and enrollment in first college class. mTBI = mild traumatic brain injury; GPA = grade point average; PTSD = posttraumatic stress disorder.

488 American Journal of Speech-Language Pathology • Vol. 27 • 485–494 • March 2018

words produced, measures developed to more discretely analyze productivity and efficiency as related to the story content. Total correct ideas reflected the number of correct individual facts recalled from each story. Ideas were clas- sified into specific informational units for scoring. For ex- ample, the story excerpt “she had four small children, the rent was due, and they had not eaten for two days” con- tained the idea units: 1) four, 2) children, 3) rent was due, 4) had not eaten, 5) two, 6) days. This scoring of correct ideas differs from the standard score calculation specified in the WMS-IV manual in that specific ideas were each counted separately rather than part of a group of ideas. For exam- ple, in the standard logical memory (LM) scoring, “two days” is counted as one concept, and scoring allows for any phrase indicating “a couple of days.” Two independent raters scored all transcriptions for total correct ideas. Across the 80 transcriptions, there was 93.75% agreement between the two scorers. No discrepancy was greater than 1 point, and score discrepancies were either resolved through a re- peated scoring of the transcript or discussion between the scorers. Total words produced reflected the total number of words that the participant used to retell the story. Filler words (e.g., “um”), phrases unrelated to the story content, and phrases indicating the participant had stopped trying to retrieve story information (e.g., “I don’t remember any- thing more.”) were not included in the count. However, any phrases that reflected continued attempts to retrieve story information (e.g., “I don’t know the word exactly…” and “I’m trying to remember the city…”) or semantically related to the story content were included in the count. For exam- ple, the statement, “…was she robbed? I think she was robbed on 5th Street.” All transcriptions were scored by two independent scorers for total words produced with 98.75% agreement between the scorers.

Results To examine group differences in production in the

two recall conditions, a 2 (Delay Condition: Immediate vs. Delayed) × 2 (Group: Control vs. mTBI) mixed-factor analysis of variance (ANOVA) was conducted on each measure. Pairwise comparisons between mTBI and control groups were evaluated with independent-groups t tests. Pairwise comparisons between immediate and delayed in Table 2 shows the mean LM scaled scores, total cor- rect ideas, and total words for both groups across delay conditions.

LM Scaled Scores The 2 (Delay Condition) × 2 (Group) ANOVA con-

ducted on LM scaled scores revealed no significant effect of delay, F(1, 19) = 3.13, p = .093, η2 = .142. Overall, there was no difference in immediate versus delayed perfor- mance. There was no significant effect of group: The two groups did not differ in overall recall performance (F < 1). The interaction between the variables was marginal, F(1, 19) = 4.54, p = .07, η2 = .163. Post hoc analysis (analysis of

covariance [ANCOVA]), including Gender, PTSD, and Depression as covariates revealed no significant associated differences in LM scaled scores and no significant inter- actions with delay (all Fs < 1).

Total Correct Ideas The 2 (Delay Condition) × 2 (Group) ANOVA con-

ducted on the total correct ideas revealed a significant main effect of delay, F(1, 19) = 22.21, p < .001, η2 = .539. Significantly more correct ideas were recalled in the imme- diate condition relative to the delayed condition. The effect of group was not significant (F < 1). Overall, the two groups performed equivalently. However, there was a significant Delay × Group interaction, F(1, 19) = 4.54, p = .047, η2 = .193. To examine the nature of the interaction, pairwise comparisons were conducted using repeated- measures t tests (for immediate vs. delay comparisons) and independent-groups t tests (for group comparisons). The control group produced more total correct ideas in the im- mediate versus the delay conditions, t(13) = 2.50, p = .025. The mTBI group similarly produced more correct ideas in the immediate condition, t(6) = 3.60, p = .013, but showed a larger drop between the immediate and delayed conditions than the control group (mean delay differ- ence: mTBI = 7.00 vs. control = 2.65). Post hoc analysis (ANCOVA), including Gender, PTSD, and Depression as covariates revealed no significant associated differences in correct ideas and no significant interactions with delay (all Fs < 1). See Figure 1 for mean LM scaled scores and total correct ideas for both groups.

Total Words Produced The 2 (Delay Condition) × 2 (Group) ANOVA con-

ducted on total words produced revealed a significant effect of delay, F(1, 19) = 1.14, p = .03, η2 = .057. More total words were produced in the delayed condition than in the immediate condition. There was also a significant effect of group, F(1, 19) = 4.46, p = .05, η2 = .19, with the control group producing significantly more total words than the mTBI group. A significant Delay × Group interaction was also observed, F(1, 19) = 5.08, p = .036, η2 = .211. Addi- tional pairwise comparisons were conducted to examine the nature of the interaction. In the immediate condition, the control and mTBI groups did not significantly differ in total words (t < 1, p = .364). However, in the delayed condition, the mTBI group produced significantly fewer total words relative to the control group, t(19) = 2.60, p = .018. This group difference was driven by the control group producing more total words in the delayed versus immediate conditions, t(19) = −2.59, p = .022; a difference not observed for the mTBI group (t < 1, p = .508). Post hoc analysis (ANCOVA), including Gender, PTSD, and Depression as covariates, revealed no significant associated differences in total words produced and no significant interactions with delay (all Fs < 1). See Figure 2 for mean

Gallagher & Azuma: Story Recall in Veterans With TBI 489

total words produced across delay conditions for both groups.

Discussion In our sample of military veterans, analysis of LM

scaled scores did not reveal significant performance dif- ferences between the mTBI and control groups. The LM scaled scores generally fell within 1.5 SDs of the mean. Be- cause most studies only analyze scaled scores, the lack of mTBI effect on LM scaled scores is consistent with pre- vious studies reporting small (or no) enduring effects of mTBI on neuropsychological test performance, even when individuals self-report cognitive symptoms (Hartikainen et al., 2010; Jak et al., 2015). Indeed, in our study sample, 71.4% of the mTBI group self-reported memory problems versus the 35.7% of the control group. The LM standard scoring procedures do not allow for the detection of re- trieval failures or cueing strategies. Though subtle, these retrieval impairments may nonetheless have important im- plications for functional memory performance. Given that accurate discourse retelling/recall is a critical skill in both professional and social settings, any memory retrieval im- pairment could have wide-ranging consequences.

In this study, story recall performance was further analyzed using other measures that provided more detailed response information, including total correct ideas and total words produced. These additional measures revealed diverging response patterns for veterans with and without history of mTBI. For total correct ideas, the control and mTBI groups both produced more correct ideas in the im- mediate condition than in the delay condition, but the mTBI group showed a larger drop after the delay. The lower

recall performance after a delay observed in the mTBI group is consistent with previous findings related to mTBI (Fisher, Ledbetter, Cohen, Marmor, & Tulsky, 2000) and more severe TBI (Carlozzi, Grech, & Tulsky, 2013). Thus, the detrimental effect of delay on story recall seemed to impact the veterans with mTBI more than the control veterans. Immediate recall was less likely to be affected in mTBI as it relies more strongly on rehearsal strategies and short-term storage of information. In contrast, accurate recall of information following a delay requires appropriate encoding and retrieval of the individual story elements, the semantic relationships among those ele- ments, and the sequence of the events. These processes are associated with the prefrontal cortex, an area commonly affected by mTBI (Stuss, 2011). The participants also com- pleted unrelated tasks between the immediate and delayed recall of the stories. Thus, the poorer recall of the mTBI group in the delayed condition might reflect an increased sensitivity to interference effects. Additional research with a larger sample of veterans is needed to confirm the sensitivity of this measure to the effects of mTBI on discourse memory.

Total words produced also revealed different response patterns in veterans with and without a history of mTBI. In the immediate condition, the two groups produced a sim- ilar number of total words. However, after a delay, the vet- erans in the mTBI group produced significantly fewer total words than veterans in the control group. This group differ- ence was primarily driven by the control veterans produc- ing significantly more words in the delayed condition than in the immediate condition. The total words produced in the attempted recall of story information provide insight into retrieval strategies employed by the participants.

An informal examination was performed to assess the nature of the group differences in total words and self-cueing

Table 2. Mean logical memory scaled scores, mean total correct ideas, and mean total words for veteran control and mTBI groups across delay conditions (standard deviations in parentheses; range in brackets).

Delay condition Control veterans

Veterans with mTBI

Group difference (mTBI–control) t-test results

LM scaled score: immediate 8.21 (3.5) 8.71 (2.1) +0.50 [1.0–12.0] [5.0–11.0]

LM scaled score: delayed 8.29 (2.9) 7.00 (3.1) −1.29 [3.0–13.0] [4.0–11.0]

Delay difference +0.08 −1.71 (Delayed–immediate)

Total correct ideas: immediate

21.29 (9.0) 21.71 (4.6) +0.42 t(19) = .117, p = .908 [7.0–33.0] [13.0–26.0]

Total correct ideas: delayed

18.64 (8.3) 14.71 (6.9) −3.93 t(19) = 1.08, p = .294 [4.0–31.0] [7.0–25.0]

Delay difference −2.65* −7.00* (Delayed–immediate) t(13) = 2.50, p = .025 t(6) = 3.60, p = .013

Total words: immediate

88.36 (28.0) 74.86 (14.7) −13.50 t(19) = 0.929, p = .364 [31–125] [56–95]

Total words: delayed

106.36 (37.5) 68.43 (16.9) −37.93* t(19) = 2.60, p = .018 [30–167] [43–85]

Delay difference +18.00* −6.43 (Delayed–immediate) t(19) = −2.59, p = .022

Note. mTBI = mild traumatic brain injury; LM = logical memory.

*p < .05.

490 American Journal of Speech-Language Pathology • Vol. 27 • 485–494 • March 2018

in the delayed recall condition. Of the 14 control veterans, nine (64%) showed evidence of self-cueing attempts in the form of extraneous, but story-relevant, phrases. For exam- ple, one control veteran stated, “…she had been robbed

or mugged, or held up.” The distinction between being “robbed” and “held up” is significant in the standard WMS-IV scoring. Other control veterans demonstrated attempts at phonemic self-cueing (e.g., “Was it an M state?

Figure 1. Mean standard scores (top panel) and total correct ideas (bottom panel) for veteran control and mTBI groups across immediate and delayed conditions. mTBI = mild traumatic brain injury.

Gallagher & Azuma: Story Recall in Veterans With TBI 491

Missouri?” “Her last name starts with a B, I think…”) or semantic self-cueing (e.g., “…they felt, I don’t know the word exactly, but compelled to take up a collection for her…”). In contrast, only two (28.6%) of the seven mTBI participants showed any evidence of self-cueing attempts in their delayed recall. Some veterans with mTBI provided very sparse responses with no attempt to retrieve beyond a few basic ideas, such as “I believe his name was Joseph or Joe and he caught a weather statement on the news of in- clement weather” and “I just remember Garcia.” Others stated their difficulty in remembering the story details but did not actively try to retrieve the information. For exam- ple, one veteran with mTBI responded, “A girl was robbed, I can’t remember the name of the street nor can I remember the time. [I can’t remember. That’s all I have].” The brack- eted extraneous comments reflect no attempt at self-cueing or additional retrieval strategies. Because such story-irrelevant comments were not included in the total word measure, the group difference in the delayed condition suggests that the control veterans were more likely to engage in self-initiated retrieval strategies than veterans with mTBI. This finding is consistent with previous studies that focused on the use of self-cueing strategies in individuals with TBI (Miotto et al., 2012; Ptak, van der Linden, & Schnider, 2010). An alternative explanation for the decreased self-cueing observed in the mTBI group is that the participants had simply provided all of the information they had stored and no amount of self- cueing would be effective. Additionally, it is possible that participants engaged in internal retrieval strategies that were not reflected in their verbal responses.

The patterns observed for both additional measures are complementary: The veterans with mTBI produced fewer correct ideas and fewer total words than control vet- erans, but only when recall of the information was delayed. Including the analysis of total words and correct ideas along with scaled scores in story recall responses could re- veal subtle memory deficits in veterans reporting unremit- ting mTBI symptoms. Indeed, the nature of impairments associated with mTBI may warrant more detailed analysis of narrative recall in veterans with mTBI. LM scaled scores are based on counts of larger units of information, which may overestimate or underestimate memory retrieval success. Subtle mTBI deficits may only be discernable when smaller units of meaning are analyzed. Retrieval fail- ure of small information units can have large consequences. For example, it is a nontrivial matter if one recalls “two days” as “two weeks.”

Another possible influential factor in memory perfor- mance in our veteran groups is the presence of concomitant service-related conditions, such as PTSD (Burriss, Ayers, Ginsberg, & Powell, 2008; Cooper et al., 2015). In our sam- ple, the proportion of veterans reporting PTSD was higher in the mTBI group (n = 3) than in the control group (n = 3). Given that PTSD has been shown to be a significant vari- able affecting cognitive performance in individuals with mTBI (Lange et al., 2012; Nelson et al., 2010; Verfaellie et al., 2014), post hoc analyses examining PTSD were con- ducted on our study sample. When the veterans were sepa- rated into PTSD (n = 6) and no PTSD (n = 15) groups, no significant differences were observed for any measure

Figure 2. Mean total words produced for control and mTBI groups across immediate and delayed conditions. mTBI = mild traumatic brain injury.

492 American Journal of Speech-Language Pathology • Vol. 27 • 485–494 • March 2018

(LM I and LM II scaled scores and correct ideas, all t < 1; total words for LM II, t(19) = 1.31, p = 2.06). Additionally, when the primary analyses were repeated with PTSD in- cluded as a covariate, PTSD did not account for a signifi- cant amount of the variance (all F < 1).

The findings of this preliminary study have implica- tions for assessment and treatment but require confirmation with a larger sample of veterans. In our study, the delay- associated memory decline seems to reflect both difficulties in the retrieval of ideas and failure to actively engage in self-cueing. If this pattern is observed during assessment, clinical intervention could focus on strengthening the encod- ing of information and increasing the use of retrieval strate- gies. Future research could also investigate the role of executive function in episodic memory retrieval of narra- tive discourse. Previous researchers have attributed episodic memory impairments related to mTBI to executive func- tion deficits associated with TBI (Miotto et al., 2010; Tayim et al., 2016). Executive function can also affect other as- pects of memory recall performance. For example, execu- tive function may be related to the probability of self-cueing strategy use. Future studies should investigate the role of executive function in the successful retrieval of discourse information from episodic memory. Narrative discourse analysis might also provide additional insight into mTBI- related deficits. Although narrative discourse analysis of spontaneous speech has been shown to be a sensitive mea- sure for TBI (Chapman et al., 1992; Coelho et al., 2005; Jorgensen & Togher, 2009; Peach & Coelho, 2016), it is less often used to specifically examine episodic memory perfor- mance for orally presented discourse. Such analyses may further elucidate the nature of cognitive–linguistic deficits related to mTBI.

As previously noted, military veterans with mTBI can report chronic cognitive symptoms that may not be confirmed by standard neuropsychological assessments. This study demonstrates that additional measures of mem- ory recall performance can reveal mTBI effects that are not observed when only standard scale scores are consid- ered. These more detailed measures can provide more spe- cific information regarding the nature of the memory difficulties experienced by veterans with mTBI.

Acknowledgments The authors would like to thank the staff and leadership at

the Pat Tillman Veterans Center at Arizona State University.

References Alloway, T., Gathercole, S., & Pickering, S. (2006). Verbal and

visuospatial short‐term and working memory in children: Are they separable? Child Development, 77(6), 1698–1716.

Antonucci, S. M. (2009). Use of semantic feature analysis in group aphasia treatment. Aphasiology, 23(7/8), 854–866.

Burriss, L., Ayers, E., Ginsberg, J., & Powell, D. A. (2008). Learn- ing and memory impairment in PTSD: Relationship to depres- sion. Depression and Anxiety, 25(2), 149–157.

Carlomagno, S., Giannotti, S., Vorano, L., & Marini, A. (2011). Discourse information content in non-aphasic adults with brain injury: A pilot study. Brain Injury, 25(10), 1010–1018.

Carlozzi, N., Grech, J., & Tulsky, D. (2013). Memory functioning in individuals with traumatic brain injury: An examination of the Wechsler Memory Scale–Fourth Edition (WMS–IV). Journal of Clinical and Experimental Neuropsychology, 35, 906–914.

Chapman, S. B., Culhane, K. A., Levin, H. S., Harward, H., Mendelsohn, D., Ewing-Cobbs, L., . . . Bruce, D. (1992). Narrative discourse after closed head injury in children and adolescents. Brain and Language, 43(1), 42–65.

Coelho, C. A., Grela, B., Corso, M., Gamble, A., & Feinn, R. (2005). Microlinguistic deficits in the narrative discourse of adults with traumatic brain injury. Brain Injury, 19(13), 1139–1145.

Cooper, D. B., Bunner, A. E., Kennedy, J. E., Balldin, V., Tate, D. F., Eapen, B. C., & Jaramillo, C. A. (2015). Treatment of persistent post-concussive symptoms after mild traumatic brain injury: A systematic review of cognitive rehabilita- tion and behavioral health interventions in military service members and veterans. Brain Imaging and Behavior, 9(3), 403–420.

Covassin, T., Moran, R., & Wilhelm, K. (2013). Concussion symp- toms and neurocognitive performance of high school and col- lege athletes who incur multiple concussions. The American Journal of Sports Medicine, 41(12), 2885–2889.

Craik, F., & Tulving, E. (1975). Depth of processing and the re- tention of words in episodic memory. Journal of Experimental Psychology, 104(3), 268.

Davis, D., Alea, N., & Bluck, S. (2015). The difference between right and wrong: Accuracy of older and younger adults’ story recall. International Journal of Environmental Research and Public Health, 12(9), 10861–10088.

Defense and Veterans Brain Injury Center. (2016). Defense and Veterans Brain Injury Center. Retrieved from http://dvbic. dcoe.mil/

Delis, D. C., Kramer, J. H., Kaplan, E., & Ober, B. A. (1987). CVLT, California Verbal Learning Test: Adult Version: Manual. San Antonio, TX: The Psychological Corporation.

DiRamio, D., Ackerman, R., & Mitchell, R. (2008). From combat to campus: Voices of student-veterans. Journal of Student Affairs Research and Practice, 48, 279–296.

Fisher, D., Ledbetter, M., Cohen, N., Marmor, D., & Tulsky, D. (2000). WAIS-III and WMS-III profiles of mildly to se- verely brain-injured patients. Applied Neuropsychology, 7(3), 126–132.

Galetto, V., Andreetta, S., Zettin, M., & Marini, A. (2013). Patterns of impairment of narrative language in mild traumatic brain injury. Journal of Neurolinguistics, 26(6), 649.

Geary, E. K., Kraus, M. F., Rubin, L. H., Pliskin, N. H., & Little, D. M. (2011). Verbal learning strategy following mild trau- matic brain injury. Journal of the International Neuropsycho- logical Society, 17(4), 709–719.

Greenberg, D. L., & Verfaellie, M. (2010). Interdependence of epi- sodic and semantic memory: Evidence from neuropsychology. Journal of International Neuropsychological Society, 16(5), 748–753.

Hartikainen, K. M., Wäljas, M., Isoviita, T., Dastidar, P., Liimatainen, S., Solbakk, A., . . . Öhman, J. (2010). Persis- tent symptoms in mild to moderate traumatic brain injury associated with executive dysfunction. Journal of Clinical and Experimental Neuropsychology, 32(7), 767–774.

Gallagher & Azuma: Story Recall in Veterans With TBI 493

Hoffer, M. E. (2015). Mild traumatic brain injury: Neurosensory effects. Current Opinion in Neurology, 28(1), 74–77.

Hoge, C. W., McGurk, D., Thomas, J. L., Cox, A. L., Engle, C. C., & Castro, C. A. (2008). Mild traumatic brain injury in U.S. soldiers returning from Iraq. The New England Journal of Medicine, 358(5), 453–463.

Hyatt, K., Davis, L. L., & Barroso, J. (2014). Chasing the care: Soldiers experience following combat-related mild traumatic brain injury. Military Medicine, 179(8), 849–855.

Jak, A. J., Gregory, A., Orff, H. J., Colón, C., Steele, N., Schiehser, D. M. . . . Twamley, E. W. (2015). Neuropsychological perfor- mance in treatment-seeking Operation Enduring Freedom/ Operation Iraqi Freedom veterans with a history of mild trau- matic brain injury. Journal of Clinical and Experimental Neuro- psychology, 37(4), 379–388.

Jorgensen, M., & Togher, L. (2009). Narrative after traumatic brain injury: A comparison of monologic and jointly-produced discourse. Brain Injury, 23(9), 727–740.

Jotwani, V., & Harmon, K. G. (2010). Postconcussion syndrome in athletes. Current Sports Medicine Reports, 9(1), 21.

Kennedy, M. R. T., & Nawrocki, M. D. (2003). Delayed predic- tive accuracy of narrative recall after traumatic brain injury: Salience and explicitness. Journal of Speech, Language, and Hearing Research, 46(1), 98–112.

Kontos, A. P., Elbin, R. J., Kotwal, R. S., Lutz, R. H., Kane, S., Benson, P. J. . . . Collins, M. W. (2015). The effects of combat- related mild traumatic brain injury (mTBI): Does blast mTBI history matter? The Journal of Trauma and Acute Care Surgery, 79(4 Suppl. 2), S146.

Lange, R. T., Brickell, T. A., French, L. M., Merritt, V. C., Bhagwat, A., Pancholi, S., & Iverson, G. L. (2012). Neuropsychological outcome from uncomplicated mild, complicated mild, and mod- erate traumatic brain injury in US military personnel. Archives of Clinical Neuropsychology: The Official Journal of the National Academy of Neuropsychologists, 27(5), 480.

Lezak, M. D., Howieson, D. B., Bigler, E. D., & Tranel, D. (2012). Neuropsychological Assessment (5th ed.). New York, NY: Oxford University Press.

MacGregor, A. J., Morrison, R. H., Quinn, K. H., & Garlarneau, M. R. (2011). Repeated concussion among U.S. military per- sonnel during Operation Iraqi Freedom. Journal of Rehabilita- tion Research & Development, 48(10), 1269–1278.

Malec, J. F., Brown, A. W., Leibson, C. L., Flaada, J. T., Mandrekar, J. N., Diehl, N. N., & Perkins, P. K. (2007). The Mayo Classifi- cation System for Traumatic Brain Injury Severity. Journal of Neurotrauma, 24(9), 1417–1424.

Marini, A., Galetto, V., Zampieri, E., Vorano, L., Zettin, M., & Carlomagno, S. (2011). Narrative language in traumatic brain injury. Neuropsychologia, 49(10), 2904–2910.

Marini, A., Zettin, M., & Galetto, V. (2014). Cognitive correlates of narrative impairment in moderate traumatic brain injury. Neuropsychologia, 64, 282–288.

Matsuoka, K., Kotani, I., & Yamasato, M. (2012). Correct infor- mation unit analysis for determining the characteristics of nar- rative discourse in individuals with chronic traumatic brain injury. Brain Injury, 26(13–14), 1723–1730.

Miotto, E. C., Cinalli, F. Z., Serrao, V. T., Benute, G. G., Lucia, M. C. S., & Scaff, M. (2010). Cognitive deficits in patients with mild to moderate traumatic brain injury. Arquivos de Neuro-Psiquiatria, 68, 862–868.

Miotto, E. C., Savage, C. R., Evans, J. J., Wilson, B. A., Martin, M. G. M., Balardin, J. B., . . . Amaro Junior, E. (2012). Semantic strategy training increases memory performance and brain

activity in patients with prefrontal cortex lesions. Clinical Neurology and Neurosurgery, 115(3), 309–316.

Nelson, D. L., Kitto, K., Galea, D., McEvoy, C. L., & Bruza, P. D. (2013). How activation, entanglement, and searching a semantic network contribute to event memory. Memory & Cognition, 41(6), 797–819.

Nelson, N. W., Hoelzle, J. B., McGuire, K. A., Ferrier-Auerbach, A. G., Charlesworth, M. J., & Sponheim, S. R. (2010). Evalua- tion context impacts neuropsychological performance of OEF/OIF veterans with reported combat-related concussion. Archives of Clinical Neuropsychology, 25, 713–723.

Peach, R. K., & Coelho, C. A. (2016). Linking inter- and intra- sentential processes for narrative production following traumatic brain injury: Implications for a model of discourse processing. Neuropsychologia, 80, 157–164.

Ptak, R., van der Linden, M., & Schnider, A. (2010). Cognitive rehabilitation of episodic memory disorders: From theory to practice. Frontiers in Human Neuroscience, 4, 57.

Rigg, J. L., & Mooney, S. R. (2011). Concussions and the military: Issues specific to service members. PM & R: The Journal of Injury, Function, and Rehabilitation, 3(10 Suppl. 2), S380.

Reynolds, C. (1997). Forward and backward memory span should not be combined for clinical analysis. Archives of Clinical Neuropsychology, 12(1), 29–40.

Rosen, V. M., & Engle, R. W. (1997). Forward and backward serial recall. Intelligence, 25(1), 37–47.

Ryan, P. B., Lee-Wilk, T., Kok, B. C., & Wilk, J. E. (2011). Inter- disciplinary rehabilitation of mild TBI and PTSD: A case report. Brain Injury, 25(10), 1019–1025.

St Clair-Thompson, H., & Allen, L. (2013). Are forward and back- ward recall the same? A dual-task study of digit recall. Memory & Cognition, 41(4), 519–532.

Strain, J., Womack, K., Didehbani, N., Spence, J., Conover, H., Hart, J., . . . Cullum, C. (2015). Imaging correlates of memory and concussion history in retired national football league athletes. JAMA Neurology, 72(7), 773–780.

Stuss, D. T. (2011). Functions of the frontal lobes: Relation to ex- ecutive functions. Journal of the International Neuropsychology Society, 17(5), 759–765.

Tapper, A., Gonzalez, D., Roy, E., & Niechwiej-Szwedo, E. (2017). Executive function deficits in team sport athletes with a his- tory of concussion revealed by a visual-auditory dual task par- adigm. Journal of Sports Sciences, 35(3), 231–240.

Tayim, F. M., Flashman, L. A., Wright, M. J., Roth, R. M., & McAllister, T. W. (2016). Recovery of episodic memory sub- processes in mild and complicated mild traumatic brain injury at 1 and 12 months post injury. Journal of Clinical and Experi- mental Neuropsychology, 38(9), 1005–1014.

Terrio, H., Brenner, L. A., Ivins, B. J., Cho, J. M., Helmick, K., Schwab, K., . . . Warden, D. (2009). Traumatic brain injury screening: Preliminary findings in a US army brigade combat team. Journal of Head Trauma Rehabilitation, 24, 14–23.

Verfaellie, M., Lafleche, G., Spiro, A., & Bousquet, K. (2014). Neuropsychological outcomes in OEF/OIF veterans with self-report of blast exposure: Associations with mental health, but not MTBI. Neuropsychology, 28(3), 337–346.

Wambaugh, J. L., Mauszycki, S., Cameron, R., Wright, S., & Nessler, C. (2013). Semantic feature analysis: Incorporating typicality treatment and mediating strategy training to pro- mote generalization. American Journal of Speech-Language Pathology, 22(2), S334–S369.

Wechsler, D. (2009). Wechsler Memory Scales–Fourth Edition. San Antonio, TX: Pearson.

494 American Journal of Speech-Language Pathology • Vol. 27 • 485–494 • March 2018

Copyright of American Journal of Speech-Language Pathology is the property of American Speech-Language-Hearing Association and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.