L. D. Bark, K. A. Saltykov, S. A. Kozhukhov, M. A. Kulikov
Our basic goal is to investigate the topography of ERP components using brain-computer interface (BCI) with symbolic and cyrillic matrix. 28 channels EEG in response to matrix where letters and symbols were contained was carried out in 18 experiments made on 9 healthy subjects. After EEG processing visual event related potential (ERPs) was obtained. Amplitudes and peak latencies of N1, P3a, P3b ERP components was analysed.
To decrease of individual differences between the subjects we integrated joint zones in the ERPs, each of them containing a group of topographically closed electrodes. By this way the ERP components can be performed as combination of nine numbers. To detect extent of similarity of ERP recorded from different subjects multidimensional non-metric scaling was performed. Matching the diagrams allowed us to separate out stable distributions of responses parameters in 9 regions (patterns). There are ERP components with the same activity pattern for significant and non-significant letters and symbols as it was shown for P3a. The other ERP components were different for different kinds of stimuli (N1, P3b).
According to our findings, we can claim that BCI hardware-software devices, developed for a BCI with an alphabetic matrix, can be successfully applied for a BCI with a symbolical matrix. Taking into consideration the features of this stimulation while developing classification systems is necessary, though.