Abstract:
BCI technology connects the brain to external devices. P300 wave extraction from electroencephalography (EEG) recordings is a popular BCI protocol. P300 waves are event-related potentials with a 300-ms latency after rare stimuli. This paper improved P300-based BCIs using convolutional neural networks (CNNs). P3CNET, a novel BCI classifier, improved P300 classification accuracy of the best state-of-the-art classifier. We also examined BCI system usability-enhancing pre-processing and training options. The optimal signal interval for EEG data pre-processing improved classification accuracy. We examined the minimum number of calibration sessions to balance accuracy and speed. To improve deep learning architecture explainability, we analyzed the saliency maps of the input EEG signal leading to a correct P300 classification. Eliminating less informative electrode channels did not improve accuracy. All methods and explorations were tested on two CNN classifiers, proving their generalizability. Finally, we demonstrated the benefits of transfer learning on other P300 datasets using the novel architecture. BCI practitioners can use the architectures and practical suggestions to improve its effectiveness.
Note: Please discuss with our team before submitting this abstract to the college. This Abstract or Synopsis varies based on student project requirements.
Did you like this final year project?
To download this project Code with thesis report and project training... Click Here
