LAUSR

BACKGROUND Hybrid brain computer interfaces (BCIs) combining multiple brain imaging modalities have been proposed recently to boost the performance of single modality BCIs. NEW METHOD In this paper, we propose a novel motor imagery (MI) hybrid BCI that uses electrical brain activity recorded using Electroencephalography (EEG) as well as cerebral blood flow velocity measured using functional transcranial Doppler ultrasound (fTCD). Features derived from the power spectrum for both EEG and fTCD signals were calculated. Mutual information and linear support vector machines (SVM) were employed for feature selection and classification. RESULTS Using the EEG-fTCD combination, average accuracies of 88.33%, 89.48%, and 82.38% were achieved for right arm MI versus baseline, left arm MI versus baseline, and right arm MI versus left arm MI respectively. Compared to performance measures obtained using EEG only, the hybrid system provided significant improvement in terms of accuracy by 4.48%, 5.36%, and 4.76% respectively. In addition, average transmission rates of 4.17, 5.45, and 10.57 bits/min were achieved for right arm MI versus baseline, left arm MI versus baseline, and right arm MI versus left arm MI respectively. COMPARISON WITH EXISTING METHODS Compared to EEG-fNIRS hybrid BCIs in literature, we achieved similar or higher accuracies with shorter task duration. CONCLUSIONS The proposed hybrid system is a promising candidate for real-time BCI applications.