Novel application of Multi-Channel Electroencephalography (EEG)

Abstract

With the development of technology in recording Electroencephalogram (EEG), there are many ongoing researches focusing on the relationship between EEG and our brain functionalities.

This final year project aimed to identify the connection between 'Visual Recognition', 'Memory recalling' and Event-Related Potential (ERPs). We have developed 2 experiments. Experiment 1 was the direct implementation of the 'Oddball Paradigm'. In the experiment 2, different known person ('Target') and unknown person ('Non-Target') images were displayed to subjects and subjects were requested to categorize whether they know the person or not.

The result shows that under experiment 1, EEG Channel P7 & P8 have recorded a significant ERPs P300 when the subject views a 'Target Stimuli' as P7 & P8 are responsible for visual image processing. Under the controlled experiments, by comparing the subject's average target to non-target P300 ERPs waveforms, the subject's 'categorization' process could be identified. In experiment 2, it could be shown that when the subject successfully categorizes person in the image based on their memories, there elicited a large average P300 ERPs elicited. In dealing with the data analysis, I have also applied the machine learning technology k-Nearest Neighbor classifier to do the classification.