Schedule of Events | Symposia

Naturalistic movie-viewing paradigms reveal distinct patterns of cognitive-linguistic processing across clinical populations

Poster Session D - Monday, March 31, 2025, 8:00 – 10:00 am EDT, Back Bay Ballroom/Republic Ballroom

Manuel Marte¹ (mjmarte@bu.edu), Bryce Gillis², Rowan Faris³, Colin Galvin⁴, Laura Rigolo⁵, Yanmei Tie⁶, Swathi Kiran⁷, Einat Liebenthal⁸; ¹Center for Brain Recovery, Boston University, ²Institute for Technology in Psychiatry, McLean Hospital, Harvard Medical School, ³Department of Neurosurgery, Brigham and Women’s Hospital, Harvard Medical School

Background: While standardized assessments remain diagnostic standards, naturalistic paradigms that mirror real-world processing may provide valuable complementary information. We investigated whether a movie-viewing paradigm combining continuous affect ratings with contextual language tasks could differentiate between healthy controls (HC, n=50), persons with aphasia (AP, n=30), and individuals with mild cognitive impairment (MCI, n=17). Methods: Participants watched movie clips while providing continuous valence ratings (-4 to +4) and completed post-clip comprehension and antonym generation tasks. We analyzed valence rating deviation patterns (Root Mean Squared Z-scores) and complexity across timescales (multiscale sample entropy), and language performance, to assess their diagnostic utility using LASSO classification models. Results: HC outperformed both clinical groups on language tasks (p < 0.05), with MCI showing an intermediate performance. AP showed greater valence rating deviation (p ≥ 0.004) and lower rating complexity at longer timescales (p < 0.001). Classification models distinguished AP from HC (AUC = 0.928) and the combined patient groups from HC (AUC = 0.859), with weaker classification of HC versus MCI (AUC = 0.721). Poor performance on both language tasks with high deviation of valence rating most strongly predicted AP classification, whereas poor antonym generation performance with reduced valence rating complexity most strongly predicted MCI classification. Conclusions: Naturalistic movie-viewing revealed distinct patterns of cognitive-linguistic processing across clinical populations. The combination of movie-based language performance and valence rating measures provided robust diagnostic utility, particularly in distinguishing marked (AP) and subtle (MCI) differences from HC, suggesting its potential as an ecologically valid tool that could inform targeted interventions.

Topic Area: EMOTION & SOCIAL: Emotion-cognition interactions