Abstract
Motor impaired patients performing repetitive motor tasks often reveal large single-trial performance variations. Based on a data-driven framework, we extracted robust oscillatory brain states from pre-trial intervals, which are predictive for the upcoming motor performance on the level of single trials. Based on the brain state estimate, i.e. whether the brain state predicts a good or bad upcoming performance, we implemented a novel gating strategy for the start of trials by selecting specifically suitable or unsuitable trial starting time points. In a pilot study with four chronic stroke patients with hand motor impairments, we conducted a total of 41 sessions. After few initial calibration sessions, patients completed approximately 15 hours of effective hand motor training during eight online sessions using the gating strategy. Patients' reaction times were significantly reduced for suitable trials compared to unsuitable trials and shorter overall trial durations under suitable states were found in two patients. Overall, this successful proof-of-concept pilot study motivates to transfer this closed-loop training framework to a clinical study and to other application fields, such as cognitive rehabilitation, sport sciences or systems neuroscience.
Highlights
Machine learning methods allow for the single-trial decoding of brain recordings like the electroencephalogram (EEG) to drive real-time applications [1] in brain-computer interface (BCI) systems
We analyzed if brain state independent vertical cursor position at the go-cue could potentially suffice to explain observed reaction time changes
Despite potentially patient-related challenges, we found clear evidence that the online gating strategy succeeded in separating the average oscillatory power between conditions and that the proposed protocol manipulated single-trial reaction times in all four chronic stroke patients: Trials started during suitable brain states resulted in improved motor performance metrics compared to trials started during unsuitable brain states
Summary
Machine learning methods allow for the single-trial decoding of brain recordings like the electroencephalogram (EEG) to drive real-time applications [1] in brain-computer interface (BCI) systems. BCI were suggested to extract information about background brain states [4], [5], [6]. This is closely related to the research field of passive BCIs [7] where the user’s brain state is used as an additional input modality for a technical system. Focusing on the field of post-stroke motor rehabilitation, a variety of BCI systems have been proposed and their efficacy—as well as the efficiency compared to non BCIsupported baseline methods—is still under intense investigation [8], [9], [10], [11].
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
