Exploring the functional organisation of category-selective cortex in perception and imagery

Abstract

How the brain organises responses during visual perception and visual imagery is a central question in cognitive neuroscience. One seemingly fundamental organising principle that structures neural responses during both processes is category selectivity, defined as the preferential activation of a brain region to a particular stimulus category. During visual perception, a process shaped by feedforward and feedback signals, domain-specific responses are seen for categories including, but not limited to, objects, people, and places. During visual imagery, a process dominated by feedback signals, category-selective responses fall in very close proximity to those seen during perception. In this thesis I explore the functional organisation of category selectivity during both perception and imagery.

First, informed by a neuropsychological case study, I explored hemispheric dominance in shape perception to inform possible hemispheric asymmetries in object-selective processing. I used transcranial magnetic stimulation (TMS) to disrupt neural activity in shape-responsive cortex whilst participants completed a psychophysically-calibrated shape discrimination task. TMS was applied over left or right hemisphere targets, defined using functional magnetic resonance imaging (fMRI). TMS data revealed no significant differences in performance between stimulating the left experimental versus control (V5/MT+) site and right experimental versus control site. However, exploratory analysis of experimental target site conditions showed reduced discrimination sensitivity during right relative to left hemisphere stimulation.

This reduction might be explained by hemispheric asymmetries in underlying processing of local and global features tapped into by the shape discrimination task. These data provide preliminary support for a right hemisphere dominance in early stages of object processing in occipitotemporal cortex, although require replication.

Next, I capitalised on the individual differences in visual imagery vividness to extend the exploration of category-selective processing to the imagery domain.

A subset of the population report very poor to no voluntary visual imagery – a phenomenon referred to as aphantasia. I used fMRI and electroencephalography (EEG) to explore perceptual and imagery responses to personally familiar people and places in aphantasics and compared these to existing data from typical imagers.

In the spatial domain, despite a similar global topography of responses, there were important local reductions in category selectivity in aphantasics compared to controls in ventral temporal and medial parietal cortex during visual imagery. This reduction appears indicative of reduced differentiation between category-specific imagery responses. Further, a weakening of negative responses to non-preferred stimulus categories (i.e., people in place-preferring regions) seemed to underpin this reduced selectivity, possibly indicating insufficient inhibition of inappropriate activation. In the temporal domain, aphantasics showed an early reduction in category-level decoding accuracy during imagery relative to controls, possibly revealing a critical time period during which the strength of category representations is central to the ability to experience visual imagery. In contrast, perceptual responses did not appear to show meaningful differences between the groups in either the spatial or temporal domain.

These experiments provide significant insight into the functional organisation of category selectivity during perception and imagery. Important hemispheric asymmetries may exist during the perception of objects suggestive of differences in object feature processing between hemispheres. During imagery, the degree of category selectivity in ventral temporal and medial parietal cortex and the strength of category-specific representations across time appears central to the subjective experience of visual imagery. The mechanisms responsible for shaping category-selective responses during perception and imagery appear somewhat dissociated, owing to the specific alteration of category selectivity in aphantasics during visual imagery attempts.