Estimation of directional coupling between cortical areas using Near-Infrared Spectroscopy (NIRS)

Chang-Hwan Im,Seungduk Lee,Beop-Min Kim,Dalkwon Koh,Young-Jin Jung,Do-Won Kim

doi:10.1364/oe.18.005730

Abstract

This study invesitigated the feasibility of measuring directional coupling between cortical areas with near-infrared spectroscopy (NIRS). Cerebral hemodynamic responses were recorded at the primary somatosensory cortex (S1), secondary somatosensory cortex (S2), and primary motor cortex (M1) regions of the rat barrel cortex during electrical stimulation of rat whiskers. Deoxyhemoglobin concentration changes were calculated from NIRS recordings and the Granger causality based on the multivariate autoregressive (MVAR) model was used to estimate the effective causal connectivity among S1, S2, and M1. The estimated causality patterns of seven rats showed consistent unidirectional coupling between the somatosensory areas and the motor areas (S1 and S2-->M1), which coincided well with our hypothesis because the rats' motor function was completely anesthetized. Our preliminary results suggest that cortico-cortical directional coupling can be successfully investigated with NIRS.

Highlights

Classical functional neuroimaging studies have focused on the functional mapping of brain areas, but they could provide only a limited amount of information on the underlying neuronal processes
We evaluated the Granger causality based on the multivariate autoregressive (MVAR) model for each of the trials
According to Schlögl’s study [38], the accuracy of the MVAR estimation is somewhat dependent on the number of time samples and the variance of the prediction error becomes smaller than 10% of the signal variance if the number of samples is larger than 70

Summary

Introduction

Classical functional neuroimaging studies have focused on the functional mapping of brain areas, but they could provide only a limited amount of information on the underlying neuronal processes. The interactions between different cortical areas, usually referred to as functional connectivity, can be measured by linear or nonlinear analyses of time series extracted from noninvasive brain imaging techniques such as functional magnetic resonance imaging (fMRI), high-resolution electroencephalography (EEG), electrocorticography (ECoG), and magnetoencephalography (MEG). Various algorithms to measure the directional coupling between different brain areas have been proposed and applied to both hemodynamics-based (fMRI) and electrophysiology-based (EEG, MEG, and ECoG) modalities. MVAR-based modeling has been extensively applied to functional neuroimaging studies based on fMRI [6,7], EEG [8], MEG [9], and ECoG [10]. The MVAR-modeling-based causality analysis is being applied to clinical neuroscience as well as to cognitive neuroscience and several articles in these fields have already been published [15,16]

Methods

Results

Conclusion