| Resumo: | This aim of this project was to study the correlation between brain activity and limb motion. This was done by developing a BCI system for collecting mocap data and EEG data simultaneously from users through a game. A Unity application was developed as the core of the system while integrating an OptiTrack motion tracking system and Emotiv EEG headset with it to record mocap data and EEG data respectively. Mocap data was collected in 2 coordinate planes and EEG data was collected for 14 EEG channels. The data was collected in sets known as trials and differentiated through the direction of limb motion. An additional dimension of limb motion was obtained using differing game speeds. 6 subjects participated in the experiment and had their mocap data and EEG data collected. Offline correlation analysis was performed in MATLAB to classify trials into their direction or speed. Classification of fast trials, slow trials, left trials and right trials resulted in an average accuracy of 78%, 70%, 87% and 92% respectively. Regression analysis was also performed to attempt to estimate different mocap data types based on EEG features.
|
|---|