Analysis of fNIRS data in the prefrontal cortex by blind separation into independent components

The purpose in this paper is to reduced noise of hemodynamic response including physiological noises and motion artifacts. We measured neuronal activation coupled hemodynamic response using functional near infrared spectroscopy (fNIRS) from five subjects during arithmetical task in the prefrontal cortex. Blind separation into independent components (ICs) as good technique for signal processing is used to reduce these kinds of noise. We are estimating the source signals without knowing the parameters of mixing processes from measuring of hemodynamic response. Estimate source is assumed from the components are statistically independent and better hemodynamic response is derived after get ICs. To evaluate the result, general linear model (GLM) with t-test is used and increasing of t-values has been achieved.