Resting state functional connectivity based on principal component transformation of cortical fMRI measurements

Functional brain connectivity on the basis of fMRI time series analysis is a promising research endeavor in the study of the brain in its normal state as well as under different pathologies and neurological disorders. This study introduces a new approach to constructing rest-state connectivity networks interconnection with less amount of need to a-priori assumption and without setting any specific threshold. These network topologies are shown to reflect well the fMRI measurements. This data-driven solution at constructing fMRI-based connectivity networks considers the brain as a network of networks, and defines smallest sub-network as the regions of interest made from structural segmentation of cortical areas of the brain. Principal components (PC) of these defined subnetworks are used to gauge patterns of interconnections in the hierarchy of brain networks based on a geometrical concept. Experimental evaluations were conducted on resting state fMRI recordings of a group of healthy subjects. Results of this study support the assertion that resting state networks and default mode networks can be potentially derived without the need of either thresholding or a-priori considerations.