Increasing User Functionality of an Auditory P3 Brain-Computer Interface for Functional Electrical Stimulation Application

Abstract

A brain-computer interface (BCI) provides a hands-free means of controlling electrical devices by using signals derived directly from brain activity. Its aim is to provide an additional voluntary channel of output. A P3 BCI utilizes the fact that various stimuli may provide a detection response in the brain's electrical activity to act as a control signal. The aim of this research is to explore different stimulus paradigms in an attempt to develop an accurate, efficient and readily applicable P3 BCI for real task applications. The increased amplitude of a target P3 determines the extent to which it may be detected and thus its efficiency as a signal controller in a P3 BCI. Six different experimental paradigms were explored for feasibility and sustainable applicability. Principal component analysis (PCA) and independent component analysis (ICA) were used to pre-process the data to increase computational efficiency before a linear support vector machine (SVM) was used for categorization. The experimental procedures for single trial detection produced excellent results for visual and auditory stimuli. Visual proved slightly superior overall, but the auditory paradigms were sufficient for real applications. Increasing user functionality decreased the accuracy of the results. It should be noted that accuracies of over 90% were obtained in some instances. Salient results suggest increasing the number of varying stimuli causes minimal differences in speed categorization. The added benefit of a threestimulus paradigm as opposed to the traditional paradigm is highlighted by its increased user functionality for applications such as functional electrical stimulation (FES). Additionally auditory BCIs do not require the user to avert their visual attention away from the task at hand and are thus more practical in a real environment. Coupled with the proposed threestimulus procedure, the P3 BCI's capability is vastly improved for people suffering from neurological disorders.