[Effectiveness of "brain-computer" interfaces with biofeedback in the rehabilitation of cognitive impairment after a stroke].

Objective: Comparison of the effectiveness of two "brain-computer" interface (BCI) software complexes using biofeedback (BF) and standard therapy in restoring cognitive functions after a stroke.

Material And Methods: Eighty-nine stroke patients were examined. Neuropsychological testing was carried out using the Montreal Cognitive Assessment Scale (MoCA), the Tracking test, the Wechsler subtest 9 Kohs block design test, the Schulte tables, the Memorization of 10 Words test (according to A.R. Luria). Using the simple randomization method, three groups were formed: the main group (37), the comparison group (33) and the control group (19). In Group 1, sessions were conducted with BCI+BF based on the rhythm P300; in Group 2, with BCI+BF based on the mu-rhythm of electroencephalography (EEG), Group 3 received standard therapy.

Results: An increase in the total MoCA score was reported in all three groups. The results in Groups 1 and 2 were comparable, exceeding those in Group 3 (=0.199, <0.001, =0.037). The effectiveness in Group 1 did not depend on the baseline MoCA score, exceeding the indicators in Group 3; in Group 2, the advantage over Group 3 was with a baseline MoCA of at least 22. According to the Schulte tables and the Tracking test, comparable statistically significant changes were obtained in Groups 1 and 2; no statistically significant change was reported in the control group. The Kohs block design test showed a more statistically significant change in the main group. The Memorization of 10 Words test by A.R. Luria also showed a more consistent improvement in mnestic disorders in the main group.

Conclusion: The effectiveness of BCI+BF exceeded standard therapy for post-stroke cognitive impairment. The advantage of IMC+BFB used in the main group over IMC+BFB in the comparison group was noted, which was due to a decrease in the effectiveness of the latter with a baseline MoCA score of less than 22 points, lower performance in the Memorizing 10 Words test and the Kohs block design test.