Stroke is the leading cause of disability in many parts of the world. Impairment of hand function is common in stroke and can have a devastating impact on the daily lives of stroke survivors. Modern rehabilitation strategies for stroke include the use of robot-aided therapies and brain-computer interfaces (BCI). Robot-aided stroke therapies for the hand revolve around the use of external assistive exoskeletons and devices to enable stroke survivors to perform functional tasks, such as gesturing, and pick-and-place tasks. These therapies may rely on patients using their residual electromyographic signals to elicit assistance from the exoskeletons as a form of control. However, this excludes patients with severely impaired muscle function. The need for stroke rehabilitation without the use of the patients’ residual electromyographic signals is the main impetus of research in BCI-based Motor Imagery (BCI-MI). Motor imagery is the mental rehearsal of physical movement tasks without any overt physical movements. Repeated practice of motor imagery can trigger neuroplasticity in the brain, leading to the formation of new neural pathways in the brain, and aiding in recovery from stroke. Several
authors have employed the use of BCI-MI with either visual or mechanical feedback elements in preliminary studies and have found that the additional sensory feedback to be more efficacious than simply the use of BCI-MI [1-3]. Hence, we sought to investigate the clinical benefits of concomitant MI and robot-assisted physical practice by coupling an electroencephalography (EEG)-based BCI-assisted motor imagery stroke rehabilitation system with a soft robotic glove.