Universal and Culture-Tuned Neural Codes for Vocal Emotion: An fMRI MVPA Study Using Japanese and Canadian Voices

Abstract Understanding how the human brain decodes nonverbal emotional voices is essential for elucidating the neural basis of social cognition and for advancing research on affective dysfunction in psychiatric disorders. Using functional MRI and multivariate pattern analysis (MVPA), we investigated both universal and culture-tuned neural mechanisms underlying emotional voice perception. Forty-one healthy adults (22 Asian, 19 Western participants) listened to five emotions—angry, sad, neutral, happy, and pleased—expressed by Japanese (Tokyo Affective Voices, TAV) and Canadian (Montreal Affective Voices, MAV) actors. Whole-brain MVPA revealed robust decoding in bilateral superior temporal gyri, inferior and middle frontal gyri, anterior cingulate cortex, and insula. Sadness showed the most stable representation, whereas pleasure engaged medial prefrontal and cingulate regions, reflecting internally oriented processing of positive affect. Modest cultural effects emerged, with higher decoding for neutral and happy TAV stimuli. These findings suggest a framework for objectively characterizing emotional voice processing across cultural contexts. Such an approach may provide a basis for future translational research aimed at informing more objective approaches to the study of social and affective functioning, while further validation will be required.