Cross-Subject Commonality of Emotion Representations in Dorsal Motion-Sensitive Areas

• Published in: Frontiers in neuroscience Vol. 14; p. 567797
• Main Authors: Liang, Yin, Liu, Baolin
• Format: Journal Article
• Language: English
• Published: Lausanne Frontiers Research Foundation 14.10.2020; Frontiers Media S.A
• Subjects: Brain; Classification; Cognitive ability; cross-subject decoding; emotion perception; Emotions; Experiments; Functional magnetic resonance imaging; Motion detection; motion-sensitive areas; multivariate pattern analysis; Nervous system; Neuroscience; Perception; Statistical analysis
• ISSN: 1662-453X; 1662-4548; 1662-453X
• DOI: 10.3389/fnins.2020.567797

Emotion perception is a crucial question in cognitive neuroscience and the underlying neural substrates have been the subject of intense study. One of our previous studies demonstrated that motion-sensitive areas are involved in the perception of facial expressions. However, it remains unclear whether emotions perceived from whole-person stimuli can be decoded from the motion-sensitive areas. In addition, if emotions are represented in the motion-sensitive areas, we may further ask whether the representations of emotions in the motion-sensitive areas can be shared across individual subjects. To address these questions, this study collected neural images while participants viewed emotions (joy, anger, and fear) from videos of whole-person expressions (contained both face and body parts) in a block-design functional magnetic resonance imaging (fMRI) experiment. Multivariate pattern analysis (MVPA) was conducted to explore the emotion decoding performance in individual-defined dorsal motion-sensitive regions of interest (ROIs). Results revealed that emotions could be successfully decoded from motion-sensitive ROIs with statistically significant classification accuracies for three emotions as well as positive versus negative emotions. Moreover, results from the cross-subject classification analysis showed that a person’s emotion representation could be robustly predicted by others’ emotion representations in motion-sensitive areas. Together, these results reveal that emotions are represented in dorsal motion-sensitive areas and that the representation of emotions is consistent across subjects. Our findings provide new evidence of the involvement of motion-sensitive areas in the emotion decoding, and further suggest that there exists a common emotion code in the motion-sensitive areas across individual subjects.