Opposite gradients of mental imagery and perception in orbitofrontal and occipitotemporal cortex

978 

Abstract Submission 

Jianghao Liu¹, Laurent Cohen¹, Minye Zhan², Paolo Bartolomeo¹

¹Paris Brain Institute, Paris, France, ²Neurospin, Gif-sur-Yvette, France

Jianghao Liu  
Paris Brain Institute
Paris, France

Laurent Cohen  
Paris Brain Institute
Paris, France

Minye Zhan  
Neurospin
Gif-sur-Yvette, France

Paolo Bartolomeo  
Paris Brain Institute
Paris, France

Subjective visual experience can be shaped bottom-up by external reality or built top-down in visual mental imagery (1). Specifically, we previously found that mental imagery of faces and colors recruits the relevant domain-preferring ventral occipito-temporal (VOTC) cortical patches (2). We further discovered face- and color-preferring orbitofrontal cortex (OFC) patches in human participants, localized in a medial-lateral fashion similar to those of the VOTC (2).However, how does this OFC activity relate to VOTC activity? How do these patches sustain subjective visual experience through bottom-up and top-down processing? For example, individuals with congenital aphantasia show some visual cortex activation in the absence of subjective imagery experience. Does aphantasia impact neural activity in these patches?

We collected 7T fMRI data from 10 typical imagers and 10 aphantasic individuals while they performed mental imagery and perceptual tasks in five different domains of stimuli: faces, object colors, object shapes, words, and a map of France. After scanning, they also arranged experimental items either by semantic features, or by perceived visual features. In individual participants, we identified 4 VOTC face-preferring patches, 3 VOTC color-preferring patches, and 1 OFC patch each for face and color domains. Fig. 1A displays a representative typical imager. We investigated: activity magnitude, domain selectivity, neural representations (representational similarity analysis) compared to behavioral arrangements, and functional connectivity (psychophysiological interaction). We examined differences between patches through repeated-measures ANOVAs and conducted post-hoc gradient tests to identify prevailing trends across posterior to anterior patches, comparing their activity during imagery and perception, in typical imagers and in aphantasic individuals.

OFC face-preferring patches were consistently located lateral to the OFC color-preferring patches ( x-coordinates, Bayes factor = 12.68), similar to their spatial arrangement in VOTC. Across posterior to anterior VOTC and OFC patches during perception, activity amplitudes were highest at V1, decreased along the VOTC and OFC patches; during imagery however, the activity amplitude showed the opposite pattern, which increased from posterior to anterior patches. Functional connectivity also showed opposite flows between perception and imagery, with OFC patches serving as the apex of the top-down process. Although domain selectivity and the similarity of neural representations to behavioral arrangements increased from posterior to anterior patches, both during perception and imagery.Aphantasic individuals showed comparable activities in the VOTC patches but reduced OFC activities, including decreased visual color representation in OFC color patchs, reduced activity amplitude, and orbitofrontal-temporal connectivity in both modalities. See an illustrative figure of results in Fig.1B.

During visual perception and visual mental imagery of faces and colors, activity of domain-preferring cortical patches is not uniform but follows functional gradients across the ventral visual cortex and OFC cortex. The newly discovered domain-preferring OFC patches seem to be located close to the apex of top-down processing in mental imagery. Aphantasic individuals showed decreased OFC activity in mental imagery, perhaps linked to impaired metacognitive monitoring or read-out activities (3). Thus, aphantasia is associated with different patterns of activity in the OFC cortex, but not in the high-level visual cortex.

Imagery ¹

Activation (eg. BOLD task-fMRI) ²

Congenital

Consciousness

Cortex

FUNCTIONAL MRI

HIGH FIELD MR

Perception