Inferring Emotion in Social Stories by People with Traumatic Brain Injury

Abstract

1) Compare performance of people with TBI to a normative sample on an assessment task designed to measure emotion inference ability (EIST); 2) Determine the relationship between facial affect recognition and emotion inference abilities for people with TBI. Quasi-experimental design. The EIST consists of social stories relevant to adults. Participants need to use available contextual language cues to indicate how characters in the stories were feeling. Community clinics in several sites. 177 participants with moderate to severe TBI; 40 participants without TBI. Participants with TBI had to demonstrate adequate language comprehension on the Discourse Comprehension Test (DCT) to participate. Not applicable 1) Emotion Inference in Stories Test (EIST); Facial affect recognition (DANVA2). Participants with TBI scored significantly lower than the normative group on the EIST, F(1, 214)=47.49, p < .001. Examining our TBI sample only, facial affect recognition was significantly correlated with emotional inference from stories, r=.40, p < .001 – people who had lower scores on the facial affect recognition task tended to also have more difficulty on the EIST task. Despite adequate discourse comprehension abilities, people with TBI had significant difficulty using contextual language cues to infer the emotions of characters in the social stories. Performance was associated with facial affect recognition, suggesting there is overlap in skills needed to infer emotions from faces and contextual language. Although further development is needed, the results suggest that the EIST task is a feasible tool for identifying emotion inference difficulties in people with TBI.