Indexing metadata

The effectiveness of brain-computer interface (BCI) control and the success of imagination of movement of the upper and lower extremities were evaluated by the accuracy of recognition of EEG signals (classification accuracy) when imagining movements of the hands, feet and locomotion during 10-day training of 10 volunteers. Averaged data of all the volunteers revealed, that, on the first day of training, the classification accuracy is higher when imagining locomotion than foot movements, on the second day – hands than locomotion, on the fifth day – feet than hands. The average values of classification accuracy when imagining movements of the hands and feet increase by the 3rd day of training, further changes are specific depending on which movement is imagined. When learning the imagination of locomotion, the accuracy of classification does not significantly change. An assessment of the dynamics of individual changes in the accuracy of classification according to linear trends showed that in three participants, training led to an increase in the accuracy of classification (of the hand movements and locomotion – in one subject, of feet – in two subjects); in other three participants – to decrease (of the movements of the hands and locomotion – in one subject, of the locomotion – in the second subject, of feet – in the third). The four participants, as well as the sample average, had no significant changes. The results are discussed in terms of changes in the activity of brain structures during learning and depending on the type of imaginary movements.