Wpływ redukcji liczby elektrod w systemie BCI na ocenę aktywności elektrycznej mózgu

W artykule zaprezentowano wyniki analiz dotyczących aktywności elektrycznej mózgu ukierunkowanych na możliwość redukcji liczby elektrod w badaniu EEG wykonywanym na potrzeby asynchronicznego interfejsu mózg-komputer (BCI). Stosowne obliczenia potwierdzają zasadność wyboru zestawu 8 elektrod (tzn. F3, T7, C3, Cp1, C4, T8, F4 i Cz) w systemie BCI, wykorzystującym wyspecjalizowany wzmacniacz EEG firmy g.tec, skonstruowanym w IETiSIP Politechniki Warszawskiej.

The subject of the paper is evaluation of the brain electrical activity associated with imagining some specific motor actions for the needs of asynchronous Brain-Computer Interface (BCI) [1-4]. These analysis, called EEG inverse problems, can be useful among others to optimize the number and placement of electrodes. Dedicated calculations were carried out using the algorithm sLORETA (Section 3) [5-13]. The basis of the BCI interface is the ability to detect differences between the considered classes of tasks. In the case of asynchronous interfaces, the evaluation of brain activity in the frequency domain provides much more conclusive information than the time-domain analysis. These indicate that, although the best conditions for synchronous neuronal activity are in the range of delta waves (up to 4 Hz), the biggest differences between the compared classes are apparent in the alpha band (8-12 Hz) in the central parts of the cortex (Section 5; pic. 2,3). Moreover, the performed calculations show no significant difference in the location of the brain activity sources for the results obtained using the set of 32 electrodes and after the fourfold reduction in the number of electrodes. Thus, they confirm the relevance of the set of 8 electrodes (i.e. F3, T7, C3, CP1, C4, T8, F4, and Cz) in the BCI system constructed and used in the Department of Information and Measuring Systems of the Warsaw University of Technology (Sections 6,7; Fig. 5; Tab. 2).