PROJECT TITLE :

Grouped Automatic Relevance Determination and Its Application in Channel Selection for P300 BCIs

ABSTRACT:

During the development of a brain-computer interface, it is useful to use information in multiple electrode signals. But, a little channel subset is favored for not only machine learning feasibility, but conjointly practicality in business and clinical BCI applications. An embedded channel choice approach based mostly on grouped automatic relevance determination is proposed. The proposed Gaussian conjugate cluster-sparse previous and the embedded nature of the involved Bayesian linear model enable simultaneous channel selection and have classification. Moreover, with the marginal chance (evidence) maximization technique, hyper-parameters that verify the sparsity of the model are directly estimated from the training set, avoiding time-consuming cross-validation. Experiments are conducted on P30zero speller BCIs. The results for both public and in-house datasets show that the channels selected by our techniques yield competitive classification performance with the state-of-the-art and are biologically relevant to P300.