Motor execution and imagery: a comparison of the functional network based on ICA and hierarchical integration

Abstract

Neuroimaging studies have revealed that motor imagery (MI) shared similar neural substrates with motor execution (ME) though there are some differences in the activation pattern. Most previous studies generally focused on voxel-wise based analysis. However, the congruence and difference in functional brain network relevant to MI and ME task has been rarely investigated. In this study, independent component analysis (ICA) was applied to characterize the functional brain networks underlying MI and ME. Results shows that the brain networks underlying MI and ME shared similar brain regions consisted of supplementary motor area (SMA), contralateral primary sensorimotor area (M1/S1), striatum, bilateral premotor area (PMA), posterior parietal lobule (PPL), and cerebellum. However, the ME task induced stronger activities in SMA-proper, bilateral M1/S1 and cerebellum while the MI task produced greater activities in preSMA, right cerebellum, bilateral PMA, parietal cortex and striatum. These findings are in accordance with the model proposed by Hikosaka (2002) that includes the parietal–prefrontal cortical loops for a spatial sequence and the motor cortical loops for a motor sequence. Moreover, the functional connectivity within the MI/ME-relevant network was evaluated using hierarchical integration that can quantify the total amount of interaction within the network and further assess the information exchanges within/between sub-networks. Results of hierarchical integration further indicate that parietalprefrontal areas contributes more to the integration of MI network than that of ME network while motor cortical areas contributes more to the integration of ME network than that of MI network.