The main objective is to high-light the P300 potential on certain electroencephalographic signals. P300 occurs at a relatively well defined time in relation to a stimulus and it represents a signal with a specified band frequency.
Method: The electroencephalographic (EEG) was recorded with 4 wet electrodes by means of g.MOBIlab+ module, a g.tec acquisition system. The multiresolution wavelet transform was chosen to extract the P300 potential from the EEG signal because it provides information on both time and frequency domains.
Results: The multiresolution wavelet transform decomposes the signal in sub-bands and it helps to highlight the P300 potential. The spectrum of the P300 potential is around 3Hz. For the multiresolution wavelet decomposition this corresponds to coefficients of approximation of order 4 according to 0 to 60 Hz band of the original EEG signal. The representation of these coefficients emphasizes a better detection of P300 potential then in the original signal.
Conclusion: It is shown to be a more appropriate method than the direct analysis of the signal because it works with lower dimensional signals. This method of detection of the P300 potential can be used successfully in the implementation of a Brain Computer Interface (BCI).