MBB-AVR: a mind-brain-body dataset with dynamic affective states in immersive virtual reality

This is the authors' abstract. We don't add key points for this paper.

We present MindBrainBody-AffectiveVR (MBB-AVR), a multimodal dataset capturing continuous affective experience and both central and peripheral physiological responses during naturalistic emotional stimulation in immersive virtual reality. Forty-seven healthy adults (24 female, 22 male, 1 non-binary; age = 27.1 ± 5.4) viewed a 23-minute stereoscopic 360° video sequence, composed of computer-generated, language-free content designed to elicit high affective variability over time, presented via a head-mounted display (HTC Vive Pro Eye). While immersed, participants continuously rated their momentary valence and arousal using a previously validated rating tool with real-time visual feedback. Synchronized recordings include 64-channel EEG, ECG, respiration, photoplethysmography, galvanic skin response, and head and eye tracking with pupillometry. Participants additionally completed questionnaires including interoceptive sensibility and alexithymia. The dataset is organized following the Brain Imaging Data Structure (BIDS) and includes extracted low-level audio and video features of the stimulus. It may support research on the temporal dynamics of affect, multimodal physiological correlates of emotion, and affective computing. The dataset is available at https://gin.g-node.org/fourcade_a/mbb-avr.