Domain selectivity in the parahippocampal gyrus is predicted by the same structural connectivity patterns in blind and sighted individuals

Abstract

Human ventral occipital temporal cortex contains clusters of neurons that show domain-preferring responses during visual perception. Recent studies have reported that some of these clusters show surprisingly similar domain selectivity in congenitally blind participants performing nonvisual tasks.Animportant open question is whether these functional similarities are driven by similar innate connections in blind and sighted groups. Here we addressed this question focusing on the parahippocampal gyrus (PHG), a region that is selective for large objects and scenes. Based on the assumption that patterns of long-range connectivity shape local computation, we examined whether domain selectivityinPHGisdrivenbysimilarstructuralconnectivitypatternsinthetwopopulations.Multipleregressionmodelswerebuilttopredictthe selectivity ofPHGvoxels for large human-made objects from white matter(WM)connectivity patterns in both groups. These models were then tested using independent data from participants with similar visual experience (two sighted groups) and using data from participants with different visual experience (blind and sighted groups). Strikingly, theWM-basedpredictions between blind and sighted groups were as successfulaspredictionsbetweentwoindependentsightedgroups. Thatis,thefunctional selectivity for large objects ofaPHGvoxelinablindparticipant could be accurately predicted by its WM pattern using the connection-to-function model built from the sighted group data, and vice versa. Regions that significantly predicted PHG selectivity were located in temporal and frontal cortices in both sighted and blind populations. These results show that the large-scale network driving domain selectivity in PHG is independent of vision.