Authors :
Presenting Author: Xiaoxu Fan, PhD – Baylor College of Medicine
Abhishek Tripathi, N / A – Rice UNiversity
Kelly Bijanki, PhD – Baylor College of Medicine
Rationale:
More than 60% of epilepsy patients exhibit a range of depressive comorbidities , yet little is known about how emotion processing is altered in epilepsy-related mood disorders. Methods:
Intracranial EEG (iEEG) recordings from epilepsy patients offer a unique opportunity to examine neural processing with high spatial and temporal resolution. Here, we leverage artificial intelligence (AI) to examine how epilepsy patients (N=42) process dynamic emotional information in naturalistic contexts. We analyzed iEEG data from epilepsy patients as they watched an audiovisual film. AI models extracted 48 dynamic facial emotions from the film, and encoding models evaluated neural representations of these features. Results:
We identified robust encoding of facial emotions in the dorsolateral prefrontal cortex (DLPFC), anterior superior temporal cortex (aSTC), and posterior superior temporal cortex (pSTC). Notably, DLPFC did not encode facial emotions in children, suggesting a developmental trajectory in affective processing. Conclusions:
These findings provide novel insights into how facial emotions are represented in epilepsy patients under naturalistic conditions. AI-driven emotion analysis offers an objective and scalable approach to studying emotion perception, paving the way for future investigations into how emotional processing is altered with increasing depression severity.
Funding: This work was supported by NIH R01-MH127006.