Abstract
Objective:
In participants with traumatic brain injury (TBI) and peer controls, examine 1) differences in negative attributions (interpret ambiguous behaviors negatively); 2) cognitive and emotional factors associated with negative attributions; 3) negative attribution associations with anger responses, life satisfaction, and participation.
Setting:
Two TBI outpatient rehabilitation centers.
Participants:
Participants with complicated mild to severe TBI (n=105) and peer controls (n=105).
Design:
Cross-sectional survey study.
Main Measures:
Hypothetical scenarios describing ambiguous behaviors were used to assess situational anger and attributions of intent, hostility, and blame. Executive functioning, perspective taking, emotion perception and social inference, alexithymia, aggression, anxiety, depression, participation, and life satisfaction were also assessed.
Results:
Compared to peer controls, participants with TBI rated behaviors significantly more intentional, hostile, and blameworthy. Regression models explained a significant amount of attribution variance (25–43%). Aggression was a significant predictor in all models; social inference was also a significant predictor of intent and hostility attributions. Negative attributions were associated with anger responses and lower life satisfaction.
Conclusion:
People with TBI who have higher trait aggression and poor social inferencing skills may be prone to negative interpretations of people’s ambiguous actions. Negative attributions and social inferencing skills should be considered when treating anger problems after TBI.
Keywords: Anger, aggression, attributions, brain injury, social inference, cognition
INTRODUCTION
Problems with anger and aggression are prevalent after traumatic brain injury (TBI)1 and are often accompanied by adverse psychosocial outcomes (e.g., relationship challenges and social isolation,2–4 caregiver burden,5 incarceration,6 poor community re-integration7–10). Mechanisms contributing to post-TBI anger and aggression remain largely unknown. A better understanding of individual factors related to anger and aggression is necessary to guide development of assessments and interventions that target post-TBI anger.
Anger and aggression problems after TBI have recently been explored in the context of negative attributions. Negative attributions are judgments regarding the intent, hostility, and blame about others’ actions, and theory states that these attributions drive anger responses. Empirical support for this relationship has been widely illustrated in non-TBI populations.11–13 More recently, preliminary research in participants with TBI also showed support for this theory.14 Using hypothetical scenarios, this work showed that when persons with TBI judged others’ behaviors as more intentional, hostile, and blameworthy, they were more likely to respond with anger.14 While this attribution-anger association is typical, it becomes clinically problematic when attributions are more severe than warranted; these disparate judgments are referred to as negative attribution bias (NAB).13 A preliminary study on NAB compared attribution ratings in participants with TBI (n=46) to peer controls (n=49).15 On average, the TBI group reported significantly more severe attributions of intent, hostility, and blame compared to peer controls, suggesting people with TBI might be at risk for negative attribution bias.
If NAB is to be treated, it is important to understand related risk factors. Studies in non-TBI samples, including participants with psychiatric and mental health disorders,16–19 criminals/ inmates,20 and children with conduct disorder,11,12 found that people with higher trait aggression made more severe attributions than those with lower trait aggression. Another study in the general population also found that more extreme attributions were associated with anxiety and poor social inferencing (ability to infer others’ thoughts, beliefs, intentions), but was unrelated to attention, working memory, and reasoning.21 A preliminary investigation of these relations in a TBI sample suggested negative attributions were associated with executive functioning, perspective taking, trait aggression, trait anxiety, and alexithymia (impaired emotional processing).(unpublished data). However, these findings were on a small sample (n=46) not fully powered for these analytic comparisons. Further research is needed to identify more definitive answers regarding these associations.
The current study builds on our past research and is based on the model shown in Figure 1. Study aims were to: 1) compare differences in negative attributions (i.e., intent, hostility, blame) between participants with and without TBI in a larger sample, with a focus on responses to ambiguous scenarios,14,15 2) examine associations of cognitive and emotional factors with negative attributions; and 3) examine associations of negative attributions with an emotional response (i.e., anger) and outcomes on life satisfaction and participation. Ambiguous scenarios were the primary focus of the current study since biases are believed to be more prominent in situations when others’ motivations are unclear. We anticipated findings would indicate negative attribution bias in the TBI group; that cognitive and emotional factors would be related to negative attributions and would account for a significant portion of the negative attribution variance. It was also postulated that negative attributions would be related to anger responses, lower life satisfaction, and less participation.
Figure 1:
Conceptual model of negative attributions
METHODS
Participants
This study included 105 adults with TBI who were recruited through letters to past and current patients and participants in existing databases and registries at one of two rehabilitation centers, or through local support groups and online newsletter ads affiliated at one of the centers. A group of 105 peer controls were recruited from among friends or family members of the participants with TBI, or through online ads/ social media, flyers around the hospitals and local community clubs and organizations. They were frequency matched to participants with TBI on age and gender. While our recruitment materials indicated that our study aimed to understand anger and aggression after brain injury, it was made clear that we were recruiting people with and without anger problems to help us understand why some people have anger issues and others do not. The sample size was based on an overarching study that had a wider scope of interest than the current study, which required 210 participants (105 TBI and 105 peer controls) to examine up to 21 variables in a multiple regression model using the rule of thumb of 10 participants per variable in the model.
Participants with TBI had to be at least six months post-injury and met at least one of the following criteria for complicated mild to severe TBI22: Glasgow Coma Scale Score <13 at time of injury (n=59), post-traumatic amnesia ≥24 hours (n=29), loss of consciousness ≥30 minutes (n=32), or CT scan showing intracranial abnormality (n=98). Participants with TBI were excluded if they had a premorbid neurological disorder (e.g., stroke, Parkinson’s Disease) or major psychiatric disorder that could impact social cognition (e.g. schizophrenia; bipolar disorder), or if they were ever diagnosed with a developmental disability (e.g. autism). Peer controls were excluded if they had any of these disorders or history of an acquired brain injury at any timepoint in their life. All participants were 18 or older; were free of receptive or expressive language impairments; had adequate comprehension (Discourse Comprehension Test; DCT)23 determined at screening; spoke fluent English. See Table 1 for descriptive statistics of demographics and injury information of the sample.
Table 1:
Demographics of TBI and Peer Control Participants
| Variable | TBI group (N = 105) | Peer Control group (N = 105) | P-value |
|---|---|---|---|
| Age [Mean (SD)] | 39.84 (13.5) | 40.58 (15.2) | 0.709 |
| Education level, years [Mean (SD)] | 14.15 (2.0) | 15.01(2.5) | 0.007 |
| Race, n (%) | 0.939 | ||
| Asian | 2 (1.9%) | 2 (1.9%) | |
| Black | 19 (18.1%) | 23 (21.9%) | |
| Native American | 1 (1%) | 1 (1%) | |
| Other | 2 (1.9%) | 3 (2.9%) | |
| White | 81 (77.1%) | 76 (72.4%) | |
| Ethnic group, n (%) | 0.347 | ||
| Not Hispanic or Latino | 93 (88.6%) | 97 (92.4%) | |
| Hispanic or Latino | 12 (11.4%) | 8 (7.6%) | |
| Sex, n (%) | 0.679 | ||
| Female | 45 (42.9%) | 48 (45.7%) | |
| Male | 60 (57.1%) | 57 (54.3%) | |
| Marital status, n (%) | 0.012 | ||
| Married | 31 (29.5%) | 53 (50%) | |
| Domestic partner | 2 (1.9%) | 2 (1.9%) | |
| Relationship | 13 (12.4%) | 8 (7.6%) | |
| Single | 40 (38.1%) | 35 (33.3%) | |
| Divorced | 19 (18.1%) | 7 (6.8%) | |
| Age at the time of injury [Mean (SD)] years | 31.9 (13.8) | ||
| Years since injury [Median (min,max)] | 5.1 (0.5,40.0) | ||
| Post traumatic amnesia, n (%) | |||
| < 1 hr | 10 (9.5%) | ||
| >1hr, but <24 hrs | 5 (4.8%) | ||
| 1–6 days | 12 (11.4%) | ||
| 7–13 days | 12 (11.4%) | ||
| 14–20 days | 9 (8.6%) | ||
| 21–29 days | 13 (12.4%) | ||
| 30–59 days | 21 (20.0%) | ||
| >60 days | 23 (21.9%) | ||
| Cause of injury, n (%) | |||
| Vehicular | 48 (45.7%) | ||
| Fall | 20 (19.1%) | ||
| Assault | 9 (8.6%) | ||
| Other | 28 (26.7%) | ||
| Loss of consciousness, n (%) | |||
| < 30 min | 29 (30.9%) | ||
| >30 min, but <24 hrs | 18 (19.2%) | ||
| 1–6 days | 16 (17.0%) | ||
| 7–13 days | 7 (7.5%) | ||
| 14–20 days | 7 (7.5%) | ||
| 21–29 days | 3 (3.2%) | ||
| 30–59 days | 11 (11.7%) | ||
| >60 days | 3 (3.2%) |
Measures
Epps’ Hypothetical Scenarios: Attribution and Anger Ratings13
These 21 written and narrated scenarios, presented via computer, describe character’s behaviors as benign, ambiguous, or hostile (7 per condition). Participants rated each scenario for extent to which they anticipated feeling angry, as well as how intentional, hostile, and blameworthy they perceived the character’s behavior to be. Ratings were on 9-point Likert scale. Results focused specifically on the ratings for the ambiguous scenarios. These scenarios were used in our prior research with persons with TBI.14
Stroop Color-word interference test24,25
This is a timed test of selective attention/susceptibility to interference and inhibitory control that requires naming of colors of ink that words are written in, when the word represents a color name that contrasts with the ink color. Participants must inhibit the reading of the color word and name the color of the ink.
Controlled Oral Word Association Test (COWAT; F-A-S)26
The COWAT is a measure of verbal fluency that requires participants to generate as many words as they can think of, in one minute, that start with a specific letter. There are three letter trials and the total score is the number of words recalled across the three trials.
Category Fluency27
Category fluency measures verbal conceptual fluency by having participants generate the names of as many animals as they can think of in one minute, followed by two other trials requiring generation of fruits and vegetables, respectively. The total score is the number of examples generated across the three trials.
Interpersonal Reactivity Index (IRI)- Perspective-Taking subscale28,29
The IRI is an empathy measure with four subscales: perspective-taking, fantasy, empathic concern, and personal distress. Participants rate the extent to which each of 28 empathic behaviors describes them. Only results from the perspective-taking subscale were used for our analyses per the focus of outlined aims. Scores for each subtest range from 0–28. The IRI has been used regularly in the TBI population.30–32 It can distinguish participants with and without neurological disorders and has substantial test-retest reliability and internal reliability.28,29
The Awareness of Social Inference Test (TASIT)33
The TASIT measures affect recognition and social inferences with short video scenarios divided into three subtests. In the Emotional Evaluation Test (EET) subtest, actors nonverbally express six different emotions, which participants must identify. In the Social Inference-Minimal (SI-M) subtest, actors depict social exchanges that are either sincere or sarcastic with minimal context, whereas the Social Inference-Enriched (SI-E) subtest shows interpersonal interactions that are sarcastic exchanges or lies with more context for making inferences. In the SI-M and SI-E subtests, participants make inferences about what the characters are doing, saying, thinking, and feeling. The TASIT has been found to have good construct validity.33
Buss-Perry Aggression Questionnaire (AQ)34
This standardized measure evaluates overall aggression, and includes subscales for anger, hostile thoughts, physical aggression, and verbal aggression. It is comprised of 34 statements, rated on a 5-point scale for the extent to which each is characteristic of the participant. Total T scores derived from age and gender norms were used for analyses. The AQ is a psychometrically sound measure34 that has been frequently used in many populations.35–39
State Trait Anxiety Inventory (STAI)40
The STAI is a self-report measure of state and trait anxiety. Our analyses focused on trait anxiety subtest scores. Participants rate each of 20 statements, on a 4-point scale, regarding the extent to which they have the stated feelings, . Scores range from 20–80. Higher scores indicated higher anxiety. The STAI has good concurrent validity with other anxiety measures41 and is commonly used in the TBI population.5,42,43
The Patient Health Questionnaire-9 (PHQ-9)44
The PHQ-9 assesses depression through nine self-report questions. A total score is calculated to determine depression severity. Higher scores indicate greater depression. This measure had good psychometric properties and has been validated in the TBI population.45
Toronto Alexithymia Scale (TAS-20)46
This 20-item self-report measure uses three subscales to assess participants’ ability to a) identify their own emotions; b) describe their feelings; and c) engage in externally-oriented thinking. Scores range from 20–100 and higher scores are indicative of more severe alexithymia. The TAS-20 has good validity and reliability metrics47 and is a widely accepted measure for alexithymia, including in people with TBI.
Participation Assessment with Combined Tools-Objective (PART-O)48
The PART-O is a standardized 17 item measure developed to evaluate participation in the community with respect to three participation domains: Productivity, Social Relations, and Out and About.49 Each item is scored on a 0 to 5 scale.
Satisfaction with Life Scale (SWLS)50
This self-report questionnaire consists of 5 statements for which the participants are asked to rate their agreement regarding life satisfaction on a 7 point Likert scale. The SWLS has been demonstrated to have good reliability and validity.50,51
Procedures
Interested potential participants from Indiana University/Rehabilitation Hospital of Indiana and TIRR Memorial Hermann were consented prior to testing. After consent, participants were administered a Demographic and Relevant Medical History Questionnaire, followed by a comprehension assessment to determine if they met the reading comprehension criteria (DCT=75% correct).23 Participants who met inclusion criteria completed the remainder of assessments. Aside from one participant, participants completed the battery of assessments within a single session, which lasted between 3–4 hours.
DATA ANALYSES
Analyses were conducted in SAS Version 9.4 (Cary, NC). Demographics were compared between the TBI group and peer controls using two-sample t-tests or Chi-square tests. Two-sample t-tests were used to compare negative attributions between participants with and without TBI using the Hochberg step-up method to control for multiple comparisons. Pearson correlations were computed to assess associations between negative attributions and cognitive and emotional factors. Correlations were categorized as very weak (0.00 – 0.19), weak (0.20 – 0.39), moderate (0.40 – 0.59), strong (0.60 – 0.79), and very strong (0.80 – 1.00). Block-stepwise hierarchical regression was used to assess the proportion of variance in negative attributions explained by demographic (block 1), cognitive factors (block 2), and emotional factors (block 3). Variables were included in these models if their bivariate correlation with a given negative attribution was significant at the 0.05 level. Multicollinearity was considered present if (a) the Condition Index was > 30 and, (b) two or more variance proportions were ≥ .50. A two-sample t-test was used to compare anger between participants with and without TBI. Linear regression was used to estimate the proportion of variance in anger explained by the three negative attributions.
RESULTS
Differences in Negative Attributions between Participants with TBI and Peer Controls
Compared to peer controls, participants with TBI rated ambiguous scenarios as significantly more intentional (p=.001), hostile (p=.001), and blameworthy (p=.003). (Table 2)
Table 2:
Mean differences in negative attributions to ambiguous scenarios between TBI group and peer controls
| Variable | TBI group (N = 105) | Peer control group (N = 105) | Cohen’s d | Hochberg Adjusted P-value |
|---|---|---|---|---|
| Attributions of Intent | 5.06 (1.5) | 4.43 (1.2) | 0.46 | 0.001 |
| Attributions of Hostility | 4.57 (1.7) | 3.83 (1.4) | 0.48 | 0.001 |
| Attributions of Blame | 5.73 (1.7) | 5.03 (1.5) | 0.44 | 0.003 |
Relations of Cognitive and Emotional Factors with Negative attribution After TBI
Bivariate relationships between negative attributions and cognitive and emotional factors are presented in Table 3. All cognitive factors had significant correlations with negative attributions except for Emotional Evaluation Test (non-significant). Near moderate negative correlations were found for hostile attributions and letter fluency, category fluency, and social inference – minimal. Moderate negative correlations were found for hostile attributions and social inference – enriched. All emotional factors significantly correlated positively with negative attributions, but most correlations were in the weak range. Only aggression had moderate positive correlations with negative attributions.
Table 3:
Correlation of Negative Attributions with Cognitive and Emotional Factors
| Variable | COWAT letter fluency | MOANS category fluency | Color word T score | Perspective Taking | TASIT EET | TASIT SI-M | TASIT SI-E | Alexithymia | Depression | Aggression scale total | State trait anxiety scale total |
|---|---|---|---|---|---|---|---|---|---|---|---|
| Attributions of intent | −0.22 (0.02) |
−0.26 (0.008) |
−0.23 (0.02) |
−0.18 (0.07) |
0.16 (0.10) |
−0.22 (0.02) |
−0.34 (<0.001) |
0.29 (0.002) |
0.19 (0.05) |
0.46 (<0.001) |
0.29 (0.003) |
| Attributions of hostility | −0.38 (<0.001) |
−0.35 (<0.001) |
−0.25 (0.01) |
−0.10 (0.30) |
0.19 (0.06) |
−0.37 (<0.001) |
−0.47 (<0.001) |
0.28 (0.004) |
0.23 (0.02) |
0.43 (<0.001) |
0.26 (0.007) |
| Attributions of blame | −0.23 (0.02) |
−0.24 (0.01) |
−0.09 (0.36) |
−0.23 (0.02) |
0.14 (0.16) |
−0.22 (0.0242) |
−0.29 (0.003) |
0.29 (0.003) |
0.22 (0.03) |
0.41 (<0.001) |
0.28 (0.004) |
Multivariable Models
Only demographics found to be significantly associated with negative attributions were included in the models. For the intent model, race (black) was entered. For the hostility model, years of education, race (black) ethnicity (Hispanic), and Retired or Disabled employment status were entered. No demographic variables were significantly related to Blame, and therefore not entered into that model. No injury related variables were related to negative attributions and therefore not included in the models. TASIT’s Emotional Evaluation Test was not related to any of the negative attributions so was removed from all multivariable models. In addition, multicollinearity was observed between depression and anxiety. Depression was excluded from the models and anxiety retained, as the correlation of attributions with anxiety was higher than with depression and, a priori anxiety was expected to have a stronger relationship with negative attribution than depression based on findings from previous negative attribution research.21 Perspective taking was only related to blame, so it was included in that model only. All other variables met the inclusion criteria for entry into the multivariable models.
Reported in Table 4, hierarchal regression show that demographic factors explained little of the variation in negative attributions except for hostility (18%). The addition of cognitive factors and emotional factors significantly increased the variation explained for all negative attributions, ranging between 17–19% for cognitive factors and 6–9% for emotional factors.
Table 4:
Hierarchical Modeling Results for Negative Attributions
| Intent | Hostility | Blame | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Step | R2 | ΔR2 | F (df1,df2) | p-value | R2 | ΔR2 | F (df1,df2) | p-value | R2 | ΔR2 | F (df1,df2) | p-value |
| Block 1: Demo-graphics | 6% | - | - | - | 18% | - | - | - | 0% | |||
| Block 2: Cognitive Factors | 23% | 17% | 3.64 (6,97) | 0.003 | 35% | 17% | 4.06 (6,94) | 0.001 | 19% | 19% | 3.78 (6,98) | 0.002 |
| Block 3: Emotional Factors | 32% | 9% | 6.38 (2,95) | 0.003 | 43% | 8% | 6.82 (2,92) | 0.002 | 25% | 6% | 4.17 (2,96) | 0.02 |
Final model results are presented in Table 5. For attributions of intent and hostility models, aggression and TASIT social inference – enriched variables were the only significant predictor, though Hispanic ethnicity was marginally significant for the hostility model. Aggression was the only significant predictor of blame attributions. Adjusting for all other variables in the models, higher aggression levels were associated with higher negative attribution levels. Also, TASIT higher social inference –enriched levels was associated with lower intent and hostility levels. Participants of Hispanic descent had higher levels of hostility.
Table 5:
Regressions Models for Negative Attributions
| Intent | Hostility | Blame | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Variable | B(SEβ) | t(1) | p-value | B(SEβ) | t(1) | p-value | B(SEβ) | t(1) | p-value |
| Intercept | 4.67(1.50) | 3.11 | 0.002 | 6.41(2.06) | 3.12 | 0.002 | 6.14 (2.03) | 3.02 | 0.003 |
| Education Level | - | - | - | −0.02(0.08) | −0.27 | 0.79 | - | - | - |
| Black | 0.07(0.37) | 0.18 | 0.85 | 0.16(0.41) | 0.39 | 0.70 | - | - | - |
| Hispanic | - | - | - | 0.84(0.46) | 1.82 | 0.07 | - | - | - |
| Retired/Disabled/Unemploy | - | - | - | −0.18(0.33) | −0.55 | 0.58 | - | - | - |
| Animal fluency | 0(0.05) | 0.00 | 0.99 | −0.08(0.06) | −1.40 | 0.17 | −0.06(0.06) | −0.95 | 0.35 |
| Letter fluency | −0.03(0.04) | −0.82 | 0.41 | −0.05(0.04) | −1.34 | 0.18 | −0.03(0.05) | −0.60 | 0.55 |
| Stroop T score | −0.02(0.02) | −1.08 | 0.28 | −0.01(0.02) | −0.40 | 0.69 | - | - | - |
| Perspective Taking | - | - | - | - | - | - | −0.03(0.03) | −1.08 | 0.28 |
| Social inference (minimal) | 0.02(0.03) | 0.67 | 0.50 | 0.01(0.04) | 0.38 | 0.70 | 0.01(0.04) | 0.34 | 0.74 |
| Social inference (enriched) | −0.06(0.03) | −2.04 | 0.04 | −0.08(0.03) | −2.24 | 0.03 | −0.05(0.04) | −1.54 | 0.13 |
| Alexithymia | 0.01(0.01) | 0.62 | 0.54 | 0(0.01) | 0.17 | 0.86 | 0(0.02) | 0.18 | 0.86 |
| Anxiety | −0.01(0.01) | −0.93 | 0.35 | −0.02(0.02) | −1.14 | 0.26 | −0.01(0.02) | −0.44 | 0.66 |
| Aggression | 0.07(0.02) | 3.44 | 0.001 | 0.08(0.02) | 3.60 | <0.001 | 0.07(0.02) | 2.68 | 0.009 |
Relations of Negative Attributions with Anger Responses to Characters’ Actions
Anger was significantly higher in the TBI group compared to the peer control group on average (Mean (SD) = 5.61 (1.5) vs 5.10 (1.4), d = 0.36, p = 0.01). The (adjusted) variance in anger explained by intent, hostility, and blame attributions was 57%. All negative attributions were strongly correlated (negatively) with anger (Table 6).
Table 6:
Correlations between Negative Attributions and Anger, Participation, and Satisfaction with Life
| Variable | Anger | Social Relations | Productivity | Out and About | Satisfaction with life |
|---|---|---|---|---|---|
| Attributions of intent | 0.69 (<.001) |
−0.07 (0.48) |
−0.16 (0.10) |
−0.05 (0.61) |
−0.24 (0.01) |
| Attributions of hostility | 0.70 (<.001) |
−0.11 (0.26) |
−0.26 (0.007) |
−0.10 (0.35) |
−0.21 (0.03) |
| Attributions of blame | 0.71 (<.001) |
−0.02 (0.81) |
−0.12 (0.21) |
−0.10 (0.29) |
−0.22 (0.02) |
Relations of Negative Attributions with Participation and Life Satisfaction
Negative attributions were significantly but weakly (negatively) associated with life satisfaction (Table 6). The only significant correlation between negative attributions and PART-O domains was attribution of hostility with Productivity.
DISCUSSION
Similar to prior work, this study showed that persons with TBI are more inclined to interpret other people’s behaviors more negatively than uninjured peers (NAB). Compared to peer controls, persons with TBI attributed greater intent, hostility, and blame to characters’ ambiguous behaviors, as well as reported more anger in response to ambiguous scenarios. New to this study was the focus on ambiguous scenarios, which are believed to facilitate determination of NAB due to unclear reasons for behaviors. Results also indicated that the more negative the attributions, the angrier participants reported feeling. In our previous study, attributions of hostile intent and blame accounted for 62% of the adjusted anger variance.15 A similar adjusted variance was seen in the current study (57%). This further validates that for some people with TBI, the frequency and intensity of anger is partially due to skewed interpretations of others’ actions. Of course, not every person with TBI who experiences anger will have NAB, as there are many factors that could contribute to post-TBI anger. However, our findings suggest that NAB contributes to post-TBI anger and could be an important target for treatment.
Another goal of this study was to understand the cognitive and emotional factors that contribute to negative attributions after TBI. This study suggests that the more difficulty people had in making social inferences in enriched situations, the more likely they were to attribute negative intent and hostility in ambiguous situations. This is consistent with findings from research conducted in healthy persons which also found that hostility ratings to ambiguous situations were related to social inferencing.21 Additionally, results from the current study showed that executive function, including verbal conceptual fluency and response inhibition, did not contribute to negative attributions when accounting for all variables with significant bivariate correlations. These findings indicate that traditional neuropsychological tests may not be the best means of assessing the cognitive abilities that are important for interpreting social behavior, and thus predicting likelihood of social success after TBI. Adding tests of social inference to traditional cognitive test batteries can be helpful for assessing potential for negatively interpreting others behavior, and these results can be used to target treatment goals. The Epps scenarios used in this study appear to be sensitive in their capacity to observe attribution differences between TBI and peer controls.
Trait aggression is the one emotional factor that contributed significantly to negative attributions of intent, hostility, and blame, after controlling for other factors. This is consistent with findings from non-TBI studies, which also found negative attribution bias was commonly associated with trait aggression in peer controls.12,13,52 It is intuitive that persons who have an aggressive character trait may be suspicious of others and may attribute negative characteristics to others. However, it is also possible that long-standing negative attribution bias in persons with TBI may contribute to chronic trait aggression tendencies. This is an important distinction, as the latter possibility would emphasize the importance of early intervention to avoid the development of an aggressive trait. Since the current analyses cannot determine cause and effect, these possibilities should be investigated in future research.
Another notable finding of this study was the weak, yet significant, association between negative attributions and decreased life satisfaction. It was expected that people who were inclined to make more negative attributions would have a lower satisfaction with life because of the challenges it would present to relationships and community participation. However, the PART-O productivity was the only participation domain to be correlated with attributions of hostility, and that correlation was weak. This was an unexpected finding. However, questions in the PART-O regarding social relations revolve around frequency of social interactions, not the quality of those interactions. Frequency of social interactions may be impacted less by negative attributions than is relationship quality. Additionally, participation is impacted by many factors, including environmental characteristics (e.g., accessibility, attitudes) which may be stronger contributors to participation.
The relationship between Hispanic ethnicity and greater attributions of hostility in ambiguous situations emphasizes the importance of considering cultural factors when investigating negative attribution bias. For Hispanics, it may be factors other than TBI that contribute to negative attribution bias. It is possible that cultural factors may predispose someone of Hispanic ethnicity to be more suspicious of others’ motives. Future research should compare attribution bias in Hispanics with and without TBI, as well as investigate predictors of attribution bias in this ethnic subgroup.
Future studies should explore the potential of treating post-TBI negative attribution-related anger with interventions that target social inferencing. A published case study in two individuals who had a severe TBI and severe aggression described a group perspective-teaching intervention using role play and Gestalt psychology perspective switching techniques.53 The goal was to help patients generate benign reasons for others’ actions. Aggression levels in both patients dropped to within normal limits after the intervention. The approach of teaching perspective taking to make more accurate and benign social inference is consistent with our findings and future studies should continue to explore this type of intervention to reduce anger and negative attribution bias after TBI. Another viable approach that should be considered is cognitive behavioral therapy (CBT),54 specifically to reframe attributions of intent, hostility and blame.
Limitations
The groups significantly differed in years of education and marital status. Regarding education, it was only associated with hostile attributions, and was not a significant factor in the regression model. This study did not examine the relations of attributions with marital status. Future studies should explore if higher rates of divorce and lower rates of marriage after TBI are related to participants’ attributions. Also, socioeconomic status was not included in this study; however, related to socioeconomic status is employment, which was added to the regression model for hostility and was not significant. Nevertheless, employment is not equivalent to socioeconomic status and should be examined in future research studying negative attributions. Executive functioning testing was limited, and therefore it would be beneficial to explore negative attribution relations with a broader range of executive functioning skills in future studies. It should be noted that pilot data that informed cognitive assessments for this larger study did explore relations with the Tower of Hanoi and did find significant associations. The measures used to assess anger in the current study were self-report and were thus subject to potential bias based on participants’ willingness to admit to feelings of anger. More objective measures of emotional response (e.g., Galvanic skin response, heart rate) may be helpful for future studies. Also, the experience of anger and self-report of how one would respond in a situation is not necessarily indicative of behavior. One can feel angry in response to a situation, but not act on that anger. Future research should consider using role play to investigate actual responses in socially ambiguous situations. Perhaps virtual reality technology could be used to simulate scenarios that could achieve more ecologically valid responses while also maintaining systematic control over the environment. Finally, our sample was primarily chronic (average of 5 years post-injury), and the results cannot be generalized to an acute sample.
CONCLUSION
In response to socially ambiguous behaviors, persons with TBI are more likely than peer controls to attribute negative intent, hostility, and blame to others’ actions, and attributions are associated with increased anger. This bias appears to be impacted by trait aggression and social inferencing skills. Social inferencing and related skills, such as perspective taking, are malleable treatment targets. In conclusion negative attributions should be clinically assessed and treated as part of TBI rehabilition.53,55
Funding:
National Institute on Disability, Independent Living, and Rehabilitation Research, Field Initiated Program (Grant #90IF00-95-01-00)
Footnotes
Data access:
The raw data supporting the conclusions of this manuscript will be made available by the authors, without undue reservation, to any qualified researcher.
Conflict of Interest
The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.
Contributor Information
Dawn Neumann, Associate Professor, Indiana University School of Medicine, Department of Physical Medicine and Rehabilitation, Indianapolis, IN; Research Director, Rehabilitation Hospital of Indiana.
Angelle M. Sander, Associate Professor and Director, Division of Clinical Neuropsychology and Rehabilitation Psychology, Baylor College of Medicine and Harris Health System, Department of Physical Medicine and Rehabilitation, Houston, TX; Senior Scientist and Director of the Brain Injury Research Center, TIRR Memorial Hermann.
Susan M Perkins, Indiana University School of Medicine, Biostatistics Department, Indianapolis, IN.
Surya Sruthi Bhamidipalli, Indiana University School of Medicine, Biostatistics Department, Indianapolis, IN.
Flora M Hammond, Nila Covalt Professor and Chair, Physical Medicine and Rehabilitation, Indiana University School of Medicine, Indianapolis, IN; Chief of Medical Affairs, Rehabilitation Hospital of Indiana.
References
- 1.Neumann D, Malec JF, Hammond FM. The Relations of Self-Reported Aggression to Alexithymia, Depression, and Anxiety After Traumatic Brain Injury. The Journal of Head Trauma Rehabilitation. 2016. doi: 10.1097/HTR.0000000000000261. [DOI] [PubMed] [Google Scholar]
- 2.Lezak MD. Psychological implications of traumatic brain damage for the patient’s family. Rehabilitation Psychology. 1987;31(4):241–250. [Google Scholar]
- 3.Lezak MD, O’Brien KP. Longitudinal study of emotional, social, and physical changes after traumatic brain injury. Journal of Learning Disabilities. 1988;21(8):456. [DOI] [PubMed] [Google Scholar]
- 4.Oddy M, Coughlan T, Tyerman A, Jenkins D. Social-Adjustment after Closed Head-Injury - a Further Follow-up 7 Years after Injury. Journal of Neurology Neurosurgery and Psychiatry. 1985;48(6):564–568. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Marsh N, Kersel D, Havill J, Sleigh J Caregiver burden at 6 months following severe traumatic brain injury. Brain Injury. 1998;12(3):225–238. [DOI] [PubMed] [Google Scholar]
- 6.Slaughter B, Fann JR, Ehde D. Traumatic brain injury in a county jail population: prevalence, neuropsychological functioning and psychiatric disorders. Brain Injury. 2003;17(9):731–741. [DOI] [PubMed] [Google Scholar]
- 7.Winkler D, Unsworth C, and Sloan S Factors that lead to successful community integration following severe traumatic brain injury. Journal of Head Trauma Rehabilitation. 2006;21(1):8–21. [DOI] [PubMed] [Google Scholar]
- 8.Felmingham KL, Baguley IJ, Crooks J. A comparison of acute and postdischarge predictors of employment 2 years after traumatic brain injury. Archives of Physical Medicine and Rehabilitation. 2001;82(4):435–439. [DOI] [PubMed] [Google Scholar]
- 9.Harris JA. The relationship between aggression and employment integrity. Journal of Business and Psychology. 1997;12(1):39–44. [Google Scholar]
- 10.Sansone RA, Leung JS, Wiederman MW. Aggressive behavior and employment histories in patients from an internal medicine outpatient clinic. Comprehensive psychiatry. 2012. [DOI] [PubMed] [Google Scholar]
- 11.De Castro BO, Veerman JW, Koops W, Bosch JD, Monshouwer HJ. Hostile Attribution of Intent and Aggressive Behavior: A Meta Analysis. Child Development. 2002;73(3):916–934. [DOI] [PubMed] [Google Scholar]
- 12.Dodge KA. Translational science in action: Hostile attributional style and the development of aggressive behavior problems. Development and psychopathology. 2006;18(03):791–814. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Epps J, Kendall PC. Hostile attributional bias in adults. Cognitive Therapy and Research. 1995;19(2):159–178. [Google Scholar]
- 14.Neumann D, Malec JF, Hammond FM. The association of negative attributions with irritation and anger after brain injury. Rehabilitation psychology. 2015;60(2):155. [DOI] [PubMed] [Google Scholar]
- 15.Neumann D, Malec JF, Hammond FM. Negative Attribution Bias and Anger After Traumatic Brain Injury. Journal of Head Trauma Rehabilitation. 2017; 10.1097/HTR.0000000000000259(32(3)):197–204 [DOI] [PubMed] [Google Scholar]
- 16.An SK, Kang JI, Park JY, Kim KR, Lee SY, Lee E. Attribution bias in ultra-high risk for psychosis and first-episode schizophrenia. Schizophrenia research. 2010;118(1):54–61. [DOI] [PubMed] [Google Scholar]
- 17.McNiel DE, Eisner JP, Binder RL. The relationship between aggressive attributional style and violence by psychiatric patients. Journal of Consulting and Clinical Psychology. 2003;71(2):399. [DOI] [PubMed] [Google Scholar]
- 18.Waldheter EJ, Jones NT, Johnson ER, Penn DL. Utility of social cognition and insight in the prediction of inpatient violence among individuals with a severe mental illness. The Journal of nervous and mental disease. 2005;193(9):609–618. [DOI] [PubMed] [Google Scholar]
- 19.Combs DR, Penn DL, Michael CO, et al. Perceptions of hostility by persons with and without persecutory delusions. Cognitive neuropsychiatry. 2009;14(1):30–52. [DOI] [PubMed] [Google Scholar]
- 20.Vitale JE, Newman JP, Serin RC, Bolt DM. Hostile attributions in incarcerated adult male offenders: An exploration of diverse pathways. Aggressive Behavior. 2005;31(2):99–115. [Google Scholar]
- 21.Jeon IH, Kim KR, Kim HH, et al. Attributional Style in Healthy Persons: Its Association with’Theory of Mind’Skills. Psychiatry investigation. 2013;10(1):34–40. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Malec JF, Brown AW, Leibson CL, et al. The Mayo classification system for traumatic brain injury severity. Journal of Neurotrauma. 2007;24(9):1417–1424. [DOI] [PubMed] [Google Scholar]
- 23.Brookshire R, Nicholas L. Discourse comprehension test. Tucson, AZ: Communication Skill Builders; 1993:105–108. [Google Scholar]
- 24.Stroop JR. Studies of interference in serial verbal reactions. Journal of experimental psychology. 1935;18(6):643. [Google Scholar]
- 25.Golden CJ, Freshwater SM. Stroop color and word test. 1978. [Google Scholar]
- 26.Spreen O, Strauss E. Controlled oral word association (word fluency). Spreen O, Strauss E A compendium of neuropsychological tests Oxford: Oxford Univ Pr; 1991:219–227. [Google Scholar]
- 27.Lucas JA, Ivnik RJ, Smith GE, et al. Mayo’s older Americans normative studies: category fluency norms. Journal of clinical and experimental neuropsychology. 1998;20(2):194–200. [DOI] [PubMed] [Google Scholar]
- 28.Davis MH. Measuring individual differences in empathy: Evidence for a multidimensional approach. Journal of Personality and Social Psychology. 1983;44(1):113. [Google Scholar]
- 29.Davis MH. A multidimensional approach to individual differences in empathy. JSAS Catalog of Selected Documents in Psychology. 1980;10(85):85. [Google Scholar]
- 30.Neumann D, Babbage DR, Zupan B, Willer B. A randomized controlled trial of emotion recognition training after traumatic brain injury. The Journal of Head Trauma Rehabilitation. 2015;30(3):E12–23. [DOI] [PubMed] [Google Scholar]
- 31.Neumann D, Zupan B, Babbage DR, et al. Affect recognition, empathy, and dysosmia after traumatic brain injury. Archives of Physical Medicine and Rehabilitation. 2012;93(8):1414–1420. [DOI] [PubMed] [Google Scholar]
- 32.Neumann D, Zupan B, Malec J, Hammond F. Relationships between alexithymia, affect recognition, and empathy after traumatic brain injury. The Journal of Head Trauma Rehabilitation. 2014;29(1):E18–27. [DOI] [PubMed] [Google Scholar]
- 33.McDonald S, Flanagan S, Rollins J, Kinch J. TASIT: A new clinical tool for assessing social perception after traumatic brain injury. J Head Trauma Rehabil. 2003;18(3):219–238. [DOI] [PubMed] [Google Scholar]
- 34.Buss AH, Perry M. The aggression questionnaire. Journal of Personality and Social Psychology. 1992;63(3):452–459. [DOI] [PubMed] [Google Scholar]
- 35.Dyer KFW, Bell R, McCann J, Rauch R. Aggression after traumatic brain injury: Analysing socially desirable responses and the nature of aggressive traits. Brain Injury. 2006;20(11):1163–1173. [DOI] [PubMed] [Google Scholar]
- 36.Greve KW, Sherwin E, Stanford MS, Mathias C, Love J, Ramzinski P. Personality and neurocognitive correlates of impulsive aggression in long-term survivors of severe traumatic brain injury. Brain Injury. 2001;15(3):255–262. [DOI] [PubMed] [Google Scholar]
- 37.Holtzworth-Munroe A, Rehman U, Herron K. General and spouse-specific anger and hostility in subtypes of maritally violent men and nonviolent men*. Behavior Therapy. 2000;31(4):603–630. [Google Scholar]
- 38.Hoptman MJ, D’Angelo D, Catalano D, et al. Amygdalofrontal functional disconnectivity and aggression in schizophrenia. Schizophrenia Bulletin. 2009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Palmer EJ, Thakordas V. Relationship between bullying and scores on the Buss-Perry Aggression Questionnaire among imprisoned male offenders. Aggressive Behavior. 2005;31(1):56–66. [Google Scholar]
- 40.Spielberger CD, Gorsuch RL, and, L RE. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press; 1970. [Google Scholar]
- 41.Spielberger CD, Reheiser EC, Ritterband LM, Sydeman SJ, Unger KK. Assessment of Emotional States and Personality Traits: Measuring Psychological Vital Signs, Clinical Personality Assessment: Practical Approaches. In Butcher JN (Ed.): New York: Oxford University Press; 1995. [Google Scholar]
- 42.Curran CA, Ponsford JL, Crowe S. Coping strategies and emotional outcome following traumatic brain injury: A comparison with orthopedic patients. The Journal of Head Trauma Rehabilitation. 2000;15(6):1256–1274. [DOI] [PubMed] [Google Scholar]
- 43.Neumann D, Malec JF, Hammond FM. The relationship of negative attributions with irritation, anger, and trait aggression after brain injury. Journal of Head Trauma Rehabilitation. Under Review February 2014. [Google Scholar]
- 44.Kroenke K, Spitzer RL, Williams JBW. The PHQ-9. Journal of General Internal Medicine. 2001;16(9):606–613. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Fann JR, Bombardier CH, Dikmen S, et al. Validity of the Patient Health Questionnaire-9 in assessing depression following traumatic brain injury. The Journal of Head Trauma Rehabilitation. 2005;20(6):501–511. [DOI] [PubMed] [Google Scholar]
- 46.Bagby RM, Parker JDA, Taylor GJ. The 20-Item Toronto-Alexithymia-Scale .1. Item Selection and Cross-Validation of the Factor Structure. Journal of Psychosomatic Research. 1994;38(1):23–32. [DOI] [PubMed] [Google Scholar]
- 47.Bagby RM, Taylor GJ, Parker JDA. The 20-Item Toronto-Alexithymia-Scale .2. Convergent, Discriminant, and Concurrent Validity. Journal of Psychosomatic Research. 1994;38(1):33–40. [DOI] [PubMed] [Google Scholar]
- 48.Whiteneck GG, Dijkers MP, Heinemann AW, et al. Development of the Participation Assessment with Recombined Tools–Objective for use after traumatic brain injury. Archives of physical medicine and rehabilitation. 2011;92(4):542–551. [DOI] [PubMed] [Google Scholar]
- 49.Bogner JA, Whiteneck GG, Corrigan JD, Lai J-S, Dijkers MP, Heinemann AW. Comparison of scoring methods for the Participation Assessment with Recombined Tools–Objective. Archives of physical medicine and rehabilitation. 2011;92(4):552–563. [DOI] [PubMed] [Google Scholar]
- 50.Diener E, Emmons RA, Larsen RJ, Griffin S. The satisfaction with life scale. Journal of personality assessment. 1985;49(1):71–75. [DOI] [PubMed] [Google Scholar]
- 51.Pavot W, Diener E. Review of the satisfaction with life scale. Psychological assessment. 1993;5(2):164. [Google Scholar]
- 52.De Castro BO, Veerman JW, Koops W, Bosch JD, Monshouwer HJ. Hostile attribution of intent and aggressive behavior: A meta‐analysis. Child development. 2002;73(3):916–934. [DOI] [PubMed] [Google Scholar]
- 53.Winegardner J, Keohane C, Prince La, Neumann D. Perspective training to treat anger problems after brain injury: Two case studies. NeuroRehabilitation. 2016;39(1):153–162. [DOI] [PubMed] [Google Scholar]
- 54.Beck R, Fernandez E. Cognitive-behavioral therapy in the treatment of anger: A meta-analysis. Cognitive Therapy and Research. 1998;22(1):63–74. [Google Scholar]
- 55.Westerhof-Evers HJ, Visser-Keizer AC, Fasotti L, et al. Effectiveness of a Treatment for Impairments in Social Cognition and Emotion Regulation (T-ScEmo) After Traumatic Brain Injury: A Randomized Controlled Trial. The Journal of head trauma rehabilitation. 2017;32(5):296–307. [DOI] [PubMed] [Google Scholar]