Effective channels in classification and functional connectivity pattern of prefrontal cortex by functional near infrared spectroscopy signals

In this paper, we apply support vector machine (SVM) based classification of functional near-infrared spectroscopy (fNIRS) which is non-invasive monitoring of human brain function by measuring the changes in the concentration of oxyhemoglobin and deoxyhemoglobin. Data collected from 11 healthy volunteers and 16 schizophrenia subjects. Signals were first preprocessed and decomposed by using discrete wavelet transform DWT to eliminate systemic physiological interference. A preliminary analysis based on Genetic Algorithm (GA) favored eight channels of the reconstructed fNIRS signals for further analysis. Energy in these 8 reconstructed signals was computed and used for classification of signals. SVM based classifier was employed to diagnosis schizophrenia. The results show the promising classification accuracy of nearly 84% in detection of schizophrenia from healthy subjects. The major finding of this study is that selected channels were able to identify differences in functional connectivity patterns of prefrontal cortex (PFC) elicited by Stroop task.