Probing the imaginable: Shared and Distinct Processes between Imagery and Perception across semantic domains

Our ability to visualize things outside our field of view feels similar to our visual perception. Yet, perception arises from real-world sensory experiences, while imagery depends on mental representations. This raises important questions about whether imagery and perception involve distinct mechanisms, how similar latent spaces support semantic representations in imagery and perception, and the extent to which one’s imagery vividness and performance measures relate. To answer these questions, we employed a computerized battery prompting visual perception and visual mental imagery across five semantic domains: Colors, Faces, Letters, Maps, and Shapes. Seventy-five participants with typical imagery (based on their scores obtained from the Vividness of Visual Imagery Questionnaire, VVIQ) completed this battery. On each trial, we recorded two imagery performance measures, accuracy and response times (RTs), and two subjective measures, vividness for the imagery and confidence for the perceptual trials. Two analyses suggested the existence of domain-specific processes that differentiate imagery from perception: (1) there is a difference in both performance measures among some domains (Color, Face, Map) but not others (Letters, Shapes); (2) the logistic regression models predicted domain-specific performance. Participants’ accuracy and RTs for imagery and perception were highly correlated, suggesting shared processes exist across these two modalities. Further, unsupervised clustering analysis revealed a latent space of four clusters, fewer than the 10 theoretically possible clusters (2 modalities × 5 domains). Last, we found that trial-by-trial vividness better predicted imagery performance than off-task VVIQ, indicating its more accurate reflection of an individual’s imagery performance.