New Tools for the Assessment of Social Competence in Traumatic Brain Injury

Background: Patients with non-penetrating traumatic brain injuries (TBI) are at high risk for damage to ventral prefrontal brain regions, due to the brain’s acceleration into the bony ridges of the anterior portions of the skull. Current neuropsychological assessments of these patients focus mainly on the assessment of so-called “executive functions,” which are associated with dorsolateral prefrontal regions. Ventral prefrontal pathology is more likely to disrupt social and emotional functions, but assessments of these abilities using objective measurements that require patients to demonstrate their competence are rare. Mounting evidence suggests that chronic social and emotional deficits are common in TBI, and that these difficulties result in significant functional impairments post-injury, making clear the need to develop and use objective assessment tools during clinical neuropsychological assessments. Methods: In the current project, the Global Interpersonal Skills Test (GIST), a questionnaire measure of social skills with both a self-rated and an informant-rated version, was developed, along with three novel or adapted performance-based measures of social and emotional functioning. The first novel measure was the Assessments of Social Context (ASC), a video-based task examining comprehension of social context using non-verbal cues, including the identification of emotions, intensions and positive or negative attitudes of one person towards another. The second task was the Awareness of Interoception Test (AIT), a heartbeat detection paradigm adapted from previous literature that measures participants’ sensitivity to their cardiac function, which is implicated in basic emotional functioning. The third task was the Social Interpretations Task (SIT), an animation-based task adapted from previous literature examining participants’ ability to apply a social interpretation to stimuli that are not inherently social in nature. All novel tasks were developed and piloted with healthy undergraduates. A group of patients with moderate to severe TBI (N = 16), and a group of non-brain damaged controls (N = 16) underwent neuropsychological testing. Standard neuropsychological measures including the Wechsler Test of Adult Reading, Digit Span and Digit Symbol Coding from the Wechsler Adult Intelligence Scale, phonemic (FAS) and semantic fluency (Animals) from the Verbal Fluency Test, and the Trail Making Test, as well as the novel measures to both patients with TBI and controls were administered. The following predictions were made: vi 1) The novel performance-based measures of social abilities would be more sensitive to the presence of moderate and severe TBI than the standard neuropsychological measures of cognition; 2) On the informant version of the GIST, but not the self-report version, patients would have scores than controls; 3) Patients’ performance on the ASC would be less accurate than controls’, and this difference would be associated with real-world social skills, measured by the informant-version of the GIST; 4) Patients’ AIT performance would be less accurate than controls’, and this would be related to their self-rated emotional changes; 5) Patients’ SIT performance would be less accurate than controls’, and this difference would be associated with real-world social skills, again measured by the informant-version of the GIST. Findings: Findings were consistent with these hypotheses: 1) A statistically significant logistic regression revealed that a Social composite variable comprising the ASC, AIT and SIT was more sensitive to the presence of TBI (β = 9.59, p < .05) than a Cognitive composite variable comprising Digit Symbol Coding, Trails B completion time, and Phonemic Fluency (β = 0.006, p = .466). 2) A multivariate analysis of variance (MANOVA) revealed that informants of patients gave lower GIST scores to patients than the informants of controls (F(1,28) = 22.2, p < 0.0001), whereas there were no differences between groups on the self-rated version of the GIST (F(1,28) = .35, p = .56); 3) Patients’ ASC performance was significantly poorer than controls’ on a MANOVA (F(1,31) = 21.7, p < .0001), and ASC total scores were significantly correlated with GIST informant scores, using Spearman’s rank-order correlations (ρ(31) = .624, p < .0001). 4) Patients’ AIT performance was significantly poorer than controls’ using an independent samples t test (t(13) = 1.43, p < 0.005), and qualitative investigation of subjective reports of emotional change among patients suggested a potential relationship between emotional changes and AIT performance. vii 5) Patients’ SIT performance was significantly poorer than controls on an independent samples t test (t(30) = -2.12, p < 0.05), and SIT scores were significantly correlated with GIST informant scores, using Spearman’s rank-order correlations (ρ(31) = .460, p < .0001). Interpretation: This research represents a preliminary step in the development of clinically useful measures of social and emotional difficulties following TBI. Given the small sample size of the patient group, and the presence of co-morbid difficulties among some of the participants in this research, further testing of these measures in larger, more homogeneous samples would strengthen the current results, as would using a comparison group of people with milder TBI, rather than neuro-typical controls. The complexity of social behaviour requires that the current measures be further validated against other real-world assessments of social ability, and that assessments of other aspects of social behaviour be conducted. Nonetheless, the measures described here are a promising start to supplementing the neuropsychological toolkit in an area that requires further development at the present time