Abstract Neural Representations of Category Membership beyond Information Coding Stimulus or Response

Abstract

For decades, researchers have debated whether mental representations are symbolic or grounded in sensory inputs and motor programs. Certainly, aspects of mental representations are grounded. However, does the brain also contain abstract concept representations that mediate between perception and action in a flexible manner not tied to the details of sensory inputs and motor programs? Such conceptual pointers would be useful when concepts remain constant despite changes in appearance and associated actions. We evaluated whether human participants acquire such representations using fMRI. Participants completed a probabilistic concept learning task in which sensory, motor, and category variables were not perfectly coupled or entirely independent, making it possible to observe evidence for abstract representations or purely grounded representations. To assess how the learned concept structure is represented in the brain, we examined brain regions implicated in flexible cognition (e.g., pFC and parietal cortex) that are most likely to encode an abstract representation removed from sensory–motor details. We also examined sensory–motor regions that might encode grounded sensory–motor-based representations tuned for categorization. Using a cognitive model to estimate participants' category rule and multivariate pattern analysis of fMRI data, we found the left pFC and MT coded for category in the absence of information coding for stimulus or response. Because category was based on the stimulus, finding an abstract representation of category was not inevitable. Our results suggest that certain brain areas support categorization behavior by constructing concept representations in a format akin to a symbol that differs from stimulus–motor codes.

Geometry of neural computation unifies working memory and planning

“…Future experiments can record neural activity patterns from human participants or animals performing CDL tasks, to test whether internal representations are governed by contingency 27,43–45 . Such an experiment could also help localize representational heterogeneity among cortical regions during the CDL task.…”

Adaptive coding of stimulus information in human frontoparietal cortex during visual classification

“…Crucially, while ventral visual cortex represents the full stimulus space, i.e. information about both categories and individual exemplars 28,35 , higher-level brain regions such as the prefrontal cortex have been reported to carry little information about individual exemplars in both humans 37 and non-human primates 26,36 . These findings demonstrate a tuning of "perceptual" representations towards the task goal, successful categorization.…”

“…It has been shown that dlPFC and parietal cortex in nonhuman primates only represent behaviorally relevant stimulus categories, but carry little information about individual exemplars within these categories (Freedman et al, 2003;Freedman & Assad, 2006). For humans, category representations in dlPFC have been shown to be highly abstract, and independent of low-level stimulus features (Mok & Love, 2021), and this reflects the tuning of stimulus representations towards the task goal, i.e. successful classification.…”

Neurodynamical Computing at the Information Boundaries of Intelligent Systems

“…1a). Conceptual knowledge and remote memories are more robustly accessible than their specific and recent counterparts, due to more extensive consolidation of synaptic traces from hippocampal circuits to the medial temporal lobe and frontal areas [132][133][134]. Collectively, these properties suggest that the cortical-hippocampal memory graph is not a strict hierarchy, but more like a distributed, multilevel arrangement of specialist to generalist neurons.…”