[Comparison of event-related potentials components characteristics obtained during stimulation of symbolical and...

In order to create a brain-computer interface (BCI) on the basis of paradigm P300 (the so-called Farwell-Donchin paradigm, FD) with a symbolical matrix used as stimuli, there were compared characteristics of event-related potentials (ERP) obtained from stimulation both by symbolical and alphabetical matrixes. The matrixes contained 6 x 6 signs (cyrillic letters or symbols-pictograms). Nine healthy adults were examined in 18 experiments during which 28 channel EEG were recorded while matrixes of two types (containing either cyrillic letters or symbols-pictograms) were used for stimulation. The obtained ERP data, i.e. amplitudes and peak latencies of the following components of ERP: N1, P3 with sub-components P3a and P3b, N4 were compared and analized for different types of the stimulation matrixes. Similar changes in amplitude or peak latency received from 7 or more out of 9 examinees were taken into consideration, matching the criteria of significance. It was discovered that for components P3a, P3b and N4 the amplitudes of ERP in response to a symbolic matrix were bigger than to a letter matrix, the opposite being true for component N1. Latent periods of ERP components were shorter for a symbolic matrix than for a letter matrix in case of components N1 and P3a, and longer in case of P3b and N4. In order to find out which zones of the brain react to stimulation the most, there was conducted a pair t-test (series of pair t-tests) to analize the topography of variety of ERP responses to different types of stimuli, and, through comparing the amplitudes of ERP components, a topographical map detailing the variety of responses to the different types of matrixes was obtained. The data about the differences were analized separately for each of 28 channels, then the (absolute magnitude t-test) were summed up algebraically for all the nine examinees. Thus, it was shown, for amplitudes of all the tested ERP components in the case of pair "significant-insignificant letters", the topography of the t-test is represented by two separate areas with distinct lateralization. In the case of pair "significant-insignificant symbols", the field of maximum responses streanch in croas direction. In the case of pair "significant letters-significant symbols" the topography is elaboratly orgenised.