Detection of P300 Signals Obtained By Using Auditory Paradigm Consisting Musical Notes

P300 signals are typical signals those are observed on electroencephalogram (EEG), recording of electrical activity of brain, which has a positive rise, approximately 300 milliseconds after encountering a stimulus at an unexpected time. Studies based on P300 are commonly done using visual stimuli. In this study, classification was done using musical notes as auditory stimuli with two labels, "Includes P300" and "Not Includes P300". It is observed that P300 signals have low signal-to-noise ratio on each single data. Feature vectors are calculated as sums of variable number of signal groups to increase the performance of classification. Accuracy, sensitivity, specificity as performance metrics are analyzed to see the impacts of group size and channels used on classification performance. Classification accuracy of 96.67% is obtained on a single channel as the average of 3 participants.