Assessing memory for emotions separately from emotion recognition after traumatic brain injury

Adults with traumatic brain injury (TBI) often show deficits in recognition of facial emotion, but their ability to remember emotions is poorly understood. Furthermore, there are no practicable tasks that measure this ability. This study examined the construct of memory for emotions using a novel Facial Recognition and Memory for Emotion (FRAME) test.

Participants were 53 adults with complicated mild-to-severe TBI and a comparison group of 64 neurologically healthy adults. The FRAME and a neuropsychological battery were administered to participants. Analyses included zero-order and partial correlations, as well as group comparisons. A series of hierarchical logistic regressions evaluated the incremental utility of the FRAME in distinguishing adults with and without TBI.

Adults with TBI performed worse than healthy participants across FRAME indices. Processing speed was the strongest correlate of both emotion recognition and memory for emotion. The FRAME demonstrated a pattern of correlations with cognitive tests supporting convergent and discriminant validity of the concept that memory for emotion is distinct from the simple perception of it. Hierarchical logistic regression models showed that memory for emotion accounted for unique variance in group membership beyond emotion recognition accuracy, memory for nonemotional faces, and verbal delayed recall.

Support was found for the construct validity of a novel performance-based assessment measure of recognition and memory for facial displays of emotion. We conclude that memory for facial emotions represents a unique aspect of social cognition, distinct from accurate recognition of facial emotions. (PsycInfo Database Record (c) 2025 APA, all rights reserved).