How does categorization reshape the representational geometry of visual cortex?

Le résumé fourni par la source

When we look for something of a particular category (e.g. a lost pair of keys), we differentially attend to some features of objects at the expense of others. What kind of influence does this selective attention have on the neural representation of stimulus features? One hypothesis is that categorization causes us to attentionally weight whole stimulus dimensions. Another is that categorization modulates perception resulting in greater local discriminability between stimuli spanning the categorical divide. Previous findings favoring one or other of these hypotheses have tended to focus on different brain regions (e.g. medial temporal lobe vs early visual cortex) without directly comparing the two alternatives or assessing how the effects of selective attention unfold over time. We conducted a simultaneous fMRI-EEG (3T and 256 channel) study in which participants (N = 15, two sessions each) carried out two categorization tasks of simple stimuli that varied in their orientation and spatial frequency. We then isolated a variety of regions of interest (ROIs), focusing on early and high-level visual areas in the ventral visual stream, as well as areas associated with conceptualization and memory, including the anterior and medial temporal lobes. Representational similarity analysis was used to compare the neural dissimilarities in these ROIs to the EEG signal, as well as a large range of categorization models of behavior that varied in the form of selective attention. We found that different attentional weighting models better captured the effect of categorization task with respect to behavior and spatiotemporal characteristics of the neural responses, both within and between participants.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.