EEG channel selection is a technique that tries to fetch maximum information from the brain with less number of channels for a specific cognitive task. Though higher number of channels provide more information, the complexity in placing the electrodes and duration of recording is high. In this context proposed here is selection of suitable channel for different cognitive tasks. Brain signals are recorded using 16 channel EEG from group of volunteers by using Control Oral Word Association test, Motor imagery, Stroop test, Trial making test and Rest. SNR based algorithm is used to find 8 optimum channels. Phase locking value (PLV) computed and grouped based on the results of channel selection algorithms. Mahalanobis distance (MD) is used to find the closeness between the groups. MD values between the groups of optimal channels is lesser when compared to other channels. This means the selected channels are functionally related that indicates the active brain region for specific cognitive task is only selected as optimal channels. The extension of this study could be helpful in development of handheld or portable device.