Improving classification performance of four class FNIRS-BCI using Mel Frequency Cepstral Coefficients (MFCC)

Abstract

Experimentation and analysis of Functional near-infrared spectroscopy (fNIRS) in Brain-Computer Interface (BCI) has increasingly been studied as a communication possibility for patients who are severely paralyzed. This study has applied this technique to distinguish brain activities during four different mental tasks. These tasks include Mental Arithmetic (MA), Motor Imagery of Left-Hand (LHMI) and Right-Hand (RHMI) and Rest. fNIRS data used is from an open access dataset of 29 individuals which was collected by Continuous-wave imaging system (NIR Scout). In this research Data integration is performed before the data is preprocessed. Usual preprocessing is done using Butterworth filter to minimize or eliminate any unwanted signal distortion. After that an extensive signal analysis is done in which six different statistical features (Signal Mean (SM), Skewness (SK), Kurtosis (KR), Standard Deviation (SD), Signal Peak (SP), and Signal Variance (SV)) are obtained in the time domain and 13 Mel Frequency Cepstral Coefficients (MFCC) features are obtained from the frequency domain. As per literature review, MFCC has never been used as feature towards classification of fNIRS signal, which is a novel contribution towards this study. Separate Classification analysis is performed on each domain features. We were able to compare, differentiate and distinguish the brain signal activities captured while performing four different tasks using three different classifiers i.e. Linear Discriminant Analysis (LDA), Support Vector Machine (SVM) and K Nearest Neighbor (KNN). The average classification accuracy of 90.54% is achieved from K Nearest Neighbors (KNN) using the time domain features and accuracy achieved from Support Vector Machine (SVM) using the frequency domain features is 95.7%. Comparison with benchmark study shows the efficiency of MFCC as suitable features for improved classification accuracy.