
Spectral clustering-based resting-state network detection approach for functional near-infrared spectroscopy
Author(s) -
Lian Duan,
Xiaoqin Mai
Publication year - 2020
Publication title -
biomedical optics express
Language(s) - Uncategorized
Resource type - Journals
SCImago Journal Rank - 1.362
H-Index - 86
ISSN - 2156-7085
DOI - 10.1364/boe.387919
Subject(s) - cluster analysis , functional near infrared spectroscopy , computer science , pattern recognition (psychology) , independent component analysis , artificial intelligence , resting state fmri , correlation , spectral clustering , consistency (knowledge bases) , selection (genetic algorithm) , biological system , mathematics , cognition , geometry , neuroscience , biology , prefrontal cortex
In recent years, studying the resting-state network (RSN) by using functional near-infrared spectroscopy (fNIRS) has received increased attention. The previous resting-state fNIRS studies mainly adopted the seed-based correlation and the independent component analysis to detect RSN. However, these methods have several inherent problems. For example, the seed-based correlation method relies on seed region selection and neglects the interactions among multiple regions. The ICA method usually relies on manual component selection, which requires rich experience from the experimenter. In the present study, we developed a new approach for fNIRS-RSN detection based on spectral clustering. It consists of two steps. First, it calculates the individual-level partition of the fNIRS measurement region by using spectral clustering with an automatically determined cluster number. Second, the individual-level partitioning results are further clustered. Those clusters with high group consistency are determined as RSN clusters. We validated the method by using simulated data and in vivo fNIRS data. The results showed that the proposed method was effective and robust for fNIRS-RSN detection.