Portal de Periódicos da CAPES

Detection of resting state functional connectivity using partial correlation analysis: A study using multi-distance and whole-head probe near-infrared spectroscopy

2016; Elsevier BV; Volume: 142; Linguagem: Inglês

10.1016/j.neuroimage.2016.08.011

1095-9572

Eisuke Sakakibara, Fumitaka Homae, Shingo Kawasaki, Yukika Nishimura, Ryu Takizawa, Shinsuke Koike, Akihide Kinoshita, Hanako Sakurada, Mika Yamagishi, Fumichika Nishimura, Akane Yoshikawa, Aya Inai, M. Nishioka, Yosuke Eriguchi, Jun Matsuoka, Yoshihiro Satomura, Naohiro Okada, Chihiro Kakiuchi, Tsuyoshi Araki, Chiemi Kan, Maki Umeda, Akihito Shimazu, Minako Uga, Ippeita Dan, Hideki Hashimoto, Norito Kawakami, Kiyoto Kasai,

Heart Rate Variability and Autonomic Control

Multichannel near-infrared spectroscopy (NIRS) is a functional neuroimaging modality that enables easy-to-use and noninvasive measurement of changes in blood oxygenation levels. We developed a clinically-applicable method for estimating resting state functional connectivity (RSFC) with NIRS using a partial correlation analysis to reduce the influence of extraneural components. Using a multi-distance probe arrangement NIRS, we measured resting state brain activity for 8 min in 17 healthy participants. Independent component analysis was used to extract shallow and deep signals from the original NIRS data. Pearson's correlation calculated from original signals was significantly higher than that calculated from deep signals, while partial correlation calculated from original signals was comparable to that calculated from deep (cerebral-tissue) signals alone. To further test the validity of our method, we also measured 8 min of resting state brain activity using a whole-head NIRS arrangement consisting of 17 cortical regions in 80 healthy participants. Significant RSFC between neighboring, interhemispheric homologous, and some distant ipsilateral brain region pairs was revealed. Additionally, females exhibited higher RSFC between interhemispheric occipital region-pairs, in addition to higher connectivity between some ipsilateral pairs in the left hemisphere, when compared to males. The combined results of the two component experiments indicate that partial correlation analysis is effective in reducing the influence of extracerebral signals, and that NIRS is able to detect well-described resting state networks and sex-related differences in RSFC.