Hand exoskeleton dexterity achieved by shared control with a semi-invasive brain-computer interface

Objective. Successful developments in brain-computer interfaces (BCIs) have led to increased expectations to restore skilled finger movements in patients suffering from temporary, progressive, or accidental paralysis. Recent studies have demonstrated that individual finger movements can be accurately decoded from electrocorticography (ECoG) recordings, with electrodes placed on the cortical convexity, to control a hand prosthetic. However, what has only been sparsely addressed is the control of an exoskeleton fitted to the paralyzed hand albeit such patients outnumber the hand amputees. Hand exoskeletons controlled by EEG-based BCIs have been used in stroke rehabilitation but fall short in providing accurate finger movement assistance primarily due to the limited quality of the EEG signal. Furthermore, as the exoskeleton’s controller parameters are kept fixed, one is disregarding the possibility of the subject losing muscle control due to a progressive disorder or regaining control during stroke rehabilitation both of which are marked by altered brain activity patterns. This study aims to extend brain-controlled hand exoskeleton research with shared control principles, shared between the decoded ECoG recordings, the spared performed muscular activity, and the exoskeleton controller. Approach. First, this study will develop and test an ECoG decoder of individual and coordinated finger movements. Then, a shared controller will be designed to arbitrate between exoskeleton-based assistance and patient-intended movement. Finally, the proposed control strategy will be implemented on a Gloreha Sinfonia hand exoskeleton and tested on epilepsy patients with acute ECoG implants in both offline and online settings. Expected results. The proposed control architecture is expected to provide fluent, reliable, and dexterous control of an ECoG-actuated hand exoskeleton. Significance. The results of this study are expected to support patients in regaining finger dexterity in their paralyzed hand, and in this way increase their engagement in daily living activities.