Encoding of common grasps in the resting human brain

Introduction: An abundant literature shows similarities between task-driven and intrinsic (low frequency fluctuations(LFFs)/ resting state) driven activity. But how exactly do task-related functional connectivity (FC) of LFFs get modulated, and what is the significance of this information? Newest resting state literature suggests that the brain at rest preserves representations of the statistical regularities of the body and the natural environment (Livne, 2020; Pezzulo et al., 2021; Betti et al., 2021). For example, recent findings suggest a higher similarity between resting state and task-evoked fluctuation patterns and patterns elicited by common hand movements as opposed to uncommon (Livne, 2020). Accordingly, in a previous work, we show that the resting somatomotor cortex retains the topography of the hand and that that effect is mediated by use (El Rassi, 2022). Here we test the hypothesis that watching motor execution performed in a common/naturalistic way has a similar FC architecture to that of resting state LFFs. We expect that visual task related modulations of LFF FC will be more pronounced than when observing nonnaturalistic actions. This is based on literature showing that action observation excites similar networks as action execution, and that that excitation is facilitated when observing naturalistic movements (Lago, 2010). Here we take it one step forward and investigate visual task modulations with respect to resting state architecture. In particular we hypothesize that 1) functional connectivity architecture is different when observing common hand grasps vs uncommon, and 2) modulations related to watching common hand grasps are more correlated to resting state fluctuations that code regularities of the environment, as opposed to viewing uncommon hand grasps. Methods: We collected 15 minutes of resting state fMRI data and then asked our participants to watch videos of common and uncommon (tilted shoulder, elbow and fingers) hand grasps of 26 subjects, organized in blocks (30 s blocks, including 10 videos each). All data was preprocessed using SPM, and subject-specific whole brain ROI to ROI LFF functional connectivity matrices were generated for the resting state block and each category, using Conn toolbox. We used the parcellation of Conn toolbox of the whole brain ending up with 32 main nodes with 496 connections. We first performed a t-test contrasting common and uncommon FC matrices. We then computed for each subject the correlation between FC patterns at rest and in each of our 2 conditions. Finally, we performed a t-test between the correlations of rest and task. Results: Results: We found 1) significant differences in connectivity between common and uncommon conditions in 48 nodes (p<0.05, uncorrected) and a general increase of functional connectivity when watching uncommon vs common movements. Of interest, the biggest clusters with FC differences were between the visual system and the sensorimotor and dorsal attention networks, 2) a higher correlation of FC between rest and common movement viewing as opposed to uncommon (p=0.002).