关键词: 生物运动, 情绪知觉, 面部表情, 自闭症, 精神分裂症

Abstract: Understanding others’ emotions requires extracting meaningful social information from diverse visual signals. Yet research on emotion perception has predominantly focused on facial expressions, leaving evidence on emotion perception from biological motion (BM) less fully integrated into broader accounts of social cognition. This review addresses two interrelated questions: How is emotional information conveyed by BM processed, and to what extent do atypical responses to BM reflect difficulties specific to movement processing versus alterations in more general mechanisms of social emotion perception? To address these questions, this review integrates research on BM emotion processing across perceptual mechanisms, neural systems, and subclinical and clinical populations.
The first line of evidence concerns the organization of BM emotion processing and its relationship to facial expression perception. Behavioral and pupillometric findings indicate that local motion and global configuration of BM make distinguishable contributions. Intact happy and sad BM elicit distinct patterns of pupil modulation relative to neutral BM. For scrambled displays preserving local motion, emotion-related pupil modulation emerges earlier, with both happy and sad BM eliciting larger responses than neutral BM but no reliable difference between the two emotional conditions. Comparisons across discrimination, detection, and passive viewing tasks further show that processing advantages for particular emotions depend on task requirements and stimulus characteristics, with happiness advantages frequently observed but anger advantages and null effects also reported. Neural evidence identifies the posterior superior temporal sulcus (STS) as a key node within a distributed network involving action observation and mentalizing regions. Associations between BM and facial emotion recognition, together with bidirectional adaptation aftereffects across these stimulus classes, further support partially shared mechanisms beyond low-level visual encoding, although they do not establish identical underlying neural representations.
The second line of evidence concerns the heterogeneity of atypical processing across subclinical and clinical populations. Elevated autistic traits and autism spectrum disorder have been linked to reduced sensitivity to BM emotion and attenuated physiological or neural modulation, although overt performance varies with individual abilities and task demands. Cognitive-perceptual schizotypal traits are associated with difficulties interpreting subtle emotional cues, whereas schizophrenia can involve pronounced difficulties recognizing fear and broader impairments that vary with symptom dimensions. High trait anxiety does not consistently reduce recognition accuracy for emotional BM, but may alter the relative processing of emotional and neutral movement. Depression has been associated with negative interpretive biases, whereas studies of bipolar disorder and Alzheimer’s disease have reported broader recognition difficulties. Dissociations between BM and facial emotion recognition in mild cognitive impairment further suggest that shared emotional mechanisms coexist with stimulus-specific vulnerabilities. These comparisons raise the possibility that superficially similar behavioral difficulties may arise from different combinations of reduced perceptual sensitivity, biased attribution, and altered integration of social information.
Drawing on these findings, we propose a hierarchical framework of social emotion perception that accommodates both stimulus-specific perceptual operations and partially shared emotional representations. Within BM processing, local kinematics may support rapid, relatively coarse discrimination between emotional and neutral movement, whereas integrating these cues with global configuration may enable finer differentiation between emotions. Across stimulus types, BM and facial expressions initially rely on perceptual operations sensitive to their distinct visual features, while the emotional information extracted from these signals may converge on partially shared representations at higher processing levels. From this perspective, atypical emotion processing may reflect difficulties in encoding particular social cues, integrating perceptual information, or representing emotional meaning across signals. Rather than assigning each population to a single impaired stage, this framework generates testable hypotheses about how these processes contribute to different patterns of atypical performance.
Taken together, the reviewed evidence outlines a hierarchical framework of BM emotion processing that connects perceptual encoding, shared emotional representations, and clinical heterogeneity, providing a more integrated account of how emotions are perceived from BM and how this ability may be disrupted. This framework may help refine theories of social emotion perception and guide research into the potential utility of BM-based measures in clinical assessment and intervention evaluation.

Key words: biological motion, emotion perception, facial expressions, autism spectrum disorder, schizophrenia