Emotion Categorization Across Modalities: Representational Similarity and Clustering Patterns from Visual, Auditory, and Audiovisual Stimuli

Emotion recognition is a fundamental aspect of social communication, yet most research has focused on visual emotion expressions, leaving less understood how auditory and audiovisual cues shape perceptual organization. The present study used representational similarity analysis (RSA) to examine how the structure of emotion recognition differs across modalities. Participants completed an emotion recognition task involving 14 emotions presented in auditory, visual, and audiovisual formats. Confusion matrices were computed for each modality and converted into representational dissimilarity matrices (RDMs). Hierarchical clustering was applied to RDMs using silhouette-based cutoffs to identify emotion clusters. Multidimensional scaling (MDS) was also conducted on confusion profiles to provide two-dimensional visualizations of emotion space. Results revealed that audiovisual recognition produced the highest degree of distinctness, with most emotions forming singleton clusters and only a few small groupings. Visual recognition showed more cohesive clusters organized largely by valence. In contrast, auditory recognition yielded the broadest multi-emotion groupings, where vocal cues blurred categorical boundaries and produced clusters that combined positive and negative states. Together, these findings demonstrate that the perceptual organization of emotion recognition varies by modality. Audiovisual cues enhanced distinctiveness, visual cues reinforced valence-based structure, and auditory cues fostered broader clusters across valence boundaries. By integrating RSA and MDS, this study highlights the importance of moving beyond accuracy to examine the representational structure of emotion perception across sensory modalities.