fNIRS Data from the Prospective of Network Theory

Source abstract

We analyze visibility networks constructed from signals captured from functional NearInfrared Spectroscopy Data Acquisition and Pre-processing technology (fNIRS) making use of properties of network theory with the aim to characterize the network properties of fNIRS visibility networks. The fNIRS technology is used to capture the brain activity of dyads of two persons by measuring the oxyhemoglobin (HbO) level during a task called “MapTask”. Our approach consists in three consecutive steps: (i) firstly, we employe a sliding window technique to segment fNIRS signals; (ii) secondly, we convert the HbO signals in each sliding window to visibility networks; (iii) thirdly, we employ network properties such as diameter, clustering coefficient, assortativity, transitivity and density across different cerebral time windows. Furthermore, we investigate the degree distribution of the nodes in the networks and it is observed that they follow a power- law distribution as the length of the signal increased suggesting scale- free characteristics.