The channel selection of the brain signal for stroke rehabilitation training has been featured in the cover story
of the top international journal IEEE Transactions on Neural Systems and Rehabilitation Engineering (2011.12)
Stroke is a cerebrovascular accident with high
disability and mortality rates. As stated in
the Hong Kong Hospital Authority Statistical
Report 2005-2010, the number of annual
stroke admissions to public hospitals has been
increasing from 11,062 cases in 1981 to
25,614 in 2010. It is expected that the burden
of stroke, giving rise to difficulties in motor
disabilities and affecting the daily life of stroke
victims, is likely to escalate substantially in
Hong Kong in the future.
Effective motor recovery after stroke depends
on intensive voluntary practice of the paretic
limbs. However, existing rehabilitation
products do not train or enable patients to
identify voluntary intention, thus making
it difficult for patients to relearn how to
reconnect signals and to control their
paralyzed limbs with their brains. The new
interactive control strategy in a robotic system
manages to breakthrough existing limitations.
It has been developed for the Brain Training
Device, and when used in conjunction with
the robotic hand system, it can guide stroke
patients to relearn the reconnection between
their brains and the paralyzed limbs. Since
it can detect brainwaves, it can control the
movement of limbs and even a robotic hand.
Voluntary planning and practicing during
therapeutic training are keys to successful
post-stroke rehabilitation. Electromyography
(EMG) and electroencephalography (EEG)
are bioelectrical signals generated by the
muscle and the cerebral cortex of a brain
respectively. These two signals are directly
related to the voluntary motor contributions
from the central nervous system (EEG,
motor imagery, i.e., thinking and planning
of a physical task) and from the peripheral
neuromuscular system (EMG, motor effort, i.e.,
execution of a physical task). In this project,
we investigated using EMG and EEG for
intention-driven robotic training in post-stroke
hand-wrist rehabilitation. Our results showed
that using 12 EEG electrodes could detect
voluntary intention from stroke patients with
high accuracy for the 20-session rehabilitation
training. The clinical studies also showed
the EMG interactive robotic training had
better motor functional recovery than passive
training on stroke patients after 20 sessions
of training. This project developed a new
stroke rehabilitation training program by using
stroke patient voluntary bioelectrical signals
to facilitate better motor function recovery.
The findings of this brain control algorithm
have been published as the cover story in
top international journal IEEE Transactions
on Neural Systems
and Rehabilitation
Engineering (2011.12).