Functional near infrared spectroscopy based congitive task classification using support vector machines

The present study analyzes brain hemodynamic concentration of frontal cortex during four cognitive mental tasks. The analysis procedure consists of three sequential steps. First, the strong brain activation regions have been investigated thoroughly from all subjects in order to find a proper electrode location that generates important brain stimuli. Second, a feature extraction method that is based on wavelet transforms and denoising technique for extraction of important task-relevant features. Finally, support vector machines have been using in the classification of mental tasks with wavelet input coefficients. By applying the methodology for 4-subjects in average we achieved 92 % classification rates. However, the results depend on the type of the task that subject were performing. It is expect that the proposed method can be a basic technology for brain-computer interface by combining wavelets with support vector machines.