Overlapping but distinct: Distal connectivity dissociates hand and tool processing networks

Abstract

The processes and organizational principles of information involved in object recognition have been a subject of intense debate. These research efforts led to the understanding that local computations and feedforward/feedback connections are essential to our representations and their organization. Recent data, however, has demonstrated that distal computations also play a role in how information is locally processed. Here we focus on how long-range connectivity and local functional organization of information are related, by exploring regions that show overlapping category-preferences for two categories and testing whether their connections are related with distal representations in a category-specific way. We used an approach that relates functional connectivity with distal areas to local voxel-wise category-preferences. Specifically, we focused on two areas that show an overlap in category-preferences for tools and hands -the inferior parietal lobule/anterior intraparietal sulcus (IPL/aIPS) and the posterior middle temporal gyrus/lateral occipital temporal cortex (pMTG/LOTC)and how connectivity from these two areas relate to voxel-wise category-preferences in two ventral temporal regions dedicated to the processing of tools and hands separatelythe left medial fusiform gyrus and the fusiform body area respectively -as well as across the brain. We show that the functional connections of the two overlap areas correlate with categorical preferences for each category independently. These results show that regions that process both tools and hands maintain object topography in a category-specific way.This potentially allows for a category-specific flow of information that is pertinent to computing object representations.

High-Level Representations in Human Occipito-Temporal Cortex Are Indexed by Distal Connectivity

“…Previous evidence shows that distal functional connectivity is correlated with task-based activation in OTC ( Chen et al, 2017 ; Amaral et al, 2021 ). It is therefore possible that most-connected voxels are effectively the same as those showing strongest local activation, and therefore most-connected and most-activated voxel sets might yield equivalent decoding performance.…”

“…Indeed, connectivity is a major constraint on the functional organization of cerebral cortex in general, such that the functional response of a given region is partially determined by the integration of relevant information shared via structural and functional connectivity to other brain regions (e.g., Garcea et al, 2019 ; Lee et al, 2019 ; Mahon and Caramazza, 2011 ; Sporns and Zwi, 2004 ; Sporns, 2014 ). More specifically, category-preferring OTC responses are functionally coupled with, and modulated by, distal regions that share the same category preference (e.g., tool responses in medial fusiform gyrus (MFus) are shaped by inferior parietal cortex; Amaral et al, 2021 ; Chen et al, 2017 ; Garcea et al, 2019 ; Lee et al, 2019 ); similarly, OTC responses for multiple visual categories (e.g., faces, objects, bodies, and places) can be reliably predicted from patterns of white matter connectivity to the wider brain ( Saygin et al, 2012 , 2016 ; Osher et al, 2016 ).…”

Tool heads prime saccades

“…This intimate link between tools and hands is reflected in the neural and cognitive overlap between how tools and hands are processed (e.g. 15 – 17 ). Some experiments show that hand grasping might be automatically prepared in the presence of graspable tools.…”

“…This sequential preparation of eye and hand programs in tool recognition is reflected in distinct organization of the neural pathways processing visual information about tools. It has been shown that the recognition of manipulable objects is achieved through interactions between the dorsal and ventral streams within the tool processing network 1 , 3 , 17 , 35 – 38 . Handles are usually recognized by their elongated but coarse shape, which does not require the processing of fine spatial details, as grasping usually entails a coarse preshaping of the hand 39 .…”

“…This seems likely, as manipulable object recognition takes place through parallel processing, indicating an intensive exchange of information between the ventral and the dorsal stream components in tool recognition for action (e.g. 1 , 3 , 17 , 35 – 38 , 46 , 47 ).…”