Evaluation of mental fatigue in human-computer interaction-analysis using feature parameters extracted from event-related potential

This study made an attempt to evaluate mental fatigue induced during a VDT task using feature parameters extracted from event-related potential (P300). Since the peak of the grand averaged P300 waveform is not clear, it is sometimes difficult to detect the amplitude and the latency. The removal of the noisy EEG waveform based on the cross correlation between the grand average waveform and the individual waveforms was found to be effective for making the waveform clear. The parameter extraction methods using a principal component analysis or temporal changes of cross correlation between the grand average and the individual waveforms were used to evaluate mental fatigue. As a result, P300b component and the standard deviation of the time lag that corresponded to the maximum cross correlation between the grand averaged waveform and the individual waveforms were found to reflect some aspects of mental fatigue(the decrease of the cognitive information processing function).