A multimodal dataset of photoplethysmography and continuous behavioral responses to ASMR and nature videos
Researchers have introduced REST-ASMR, a new multimodal dataset designed to capture physiological and behavioral responses to ASMR and nature videos. The dataset includes photoplethysmography (PPG) data, synchronized audiovisual stimuli, and subjective annotations from 34 participants. A Bidirectional Long-Short Term Memory model was able to predict subjective ASMR states and classify ASMR versus nature videos with high accuracy. AI
IMPACT Provides a new resource for affective computing and multimodal research, potentially enabling more personalized relaxation models.