Sensitivity and specificity of abstraction using gist reasoning measure in adults with traumatic brain injury

Abstraction, a unique aspect of human reasoning, is affected by TBI. Lack of sensitive metrics to characterize abstraction in TBI recovery may impede detection of cognitive deficits and impact daily function. Recent findings found lower performance in abstracting meaning from complex information, referred to as gist reasoning, in adults with TBI. This study extends those findings to (a) compare abstract thinking between adults with TBI and controls on gist reasoning, similarities, and proverbs and (b) examine and compare sensitivity and specificity of abstraction measures. Participants included adults with moderate to severe chronic TBI (n = 30) and controls (n = 40), all between 25 and 55 years. ANOVA examined group differences on abstract thinking performance. A 10-fold cross-validation analysis examined sensitivity and specificity of each measure, and all possible combinations of measures. The TBI group performed significantly lower on gist reasoning and similarities. Results demonstrated higher sensitivity of gist reasoning (84.7%) as compared to similarities and proverbs. Specificity of gist reasoning (71.1%) was second highest, with proverbs multiple choice being the highest (85.2%). Combination of measures did not yield significant gains in sensitivity beyond gist reasoning alone. Gist reasoning could provide an additional diagnostic tool to detect impaired abstraction deficits in adults with chronic TBI. Furthermore, gist reasoning performance has implications to guide cognitive training.