https://www.mdu.se/

Objective: Neurofeedback training using motor imagery-based brain-computer interfaces (MI-BCIs) show promising but variable outcomes in chronic stroke rehabilitation. However, little is known on the dual task effects of feedback when simultaneously performing a mental task. This becomes especially relevant for cognitively challenged patient groups. This study aims to investigate instantaneous effects of continual visual feedback on neurophysiological activity while stroke patients practiced MI during BCI training.

Approach: Fourteen chronic stroke patients participated in a 10-week long intervention, with 4 weeks of intense MI-BCI training. During each MI training trial, patients were instructed to mentally imagine either opening or closing their paretic hand while being provided continually updated visual neurofeedback only during the last part of the trial (after 4 seconds). We analysed patient-specific spatial and temporal dynamics of peak frequency power related to MI with and without feedback, and analysed the correlation of changes to blink rate, and clinical motor- and cognitive abilities.   

Main results: Individual MI-related peak frequency of event-related desynchronisation (ERD) across sensorimotor areas was mainly found within the Alpha band. Two peak ERD patient profiles were observed: dominant ipsilesional (43\% of patients) or contralesional ERD (57\% of patients). Independent of ERD profile, we show weakened ERD in anticipation of visual feedback, followed by a restrengthening shortly after feedback onset, not significantly different from the prior no feedback phase. However, bilateral occipital decrease in individual peak frequency suggests increased information processing shortly after feedback onset. Furthermore, ERD was weakened during the last part of the trial (with feedback). This reduction was correlated to increased blink rate, and we show a tendency of higher cognitive ability for patients with greater reduction in ERD during the end of the trial. These results indicate that visual feedback does not interfere, nor significantly help the practice of MI during BCI training. 

Significance: This study provides insight in the temporal and spatial dynamics of peak frequency ERD in response to visual feedback while stroke patients practice MI during BCI training.  We believe the results will contribute to building a solid foundation of knowledge and understanding for the informed development of BCI technology in stroke rehabilitation.