Page 153 - Handbook of Biomechatronics

150 Domen Novak

close the hand while the SSVEP was used to move the arm to different loca-
tions (Horki et al., 2011). Again, however, the system was not suitable for use
with prosthetic arms due to its lack of mobility and inaccurate response.
If the use of a BCI with truly mobile prosthetic limbs is desired, we
should instead turn to the ECoG and intracortical electrodes, which provide
sufficient signal quality for continuous control of a prosthetic arm via move-
ment decoding (rather than simple classification). This was demonstrated in
multiple studies where intracortical electrodes were surgically implanted
into people with tetraplegia and used to control an advanced robotic arm
with multiple degrees of freedom (Hochberg et al., 2012; Collinger et al.,
2013). The studies found that, after training, people with tetraplegia could
use the intracortical BCI to effectively perform reach-and-grasp motions.
While the arm in these studies was stationary, future studies could attach
it to the body of an amputee and use it as a prosthesis since the BCI did
not depend on any external stimuli. However, the need for intracortical
electrodes may limit the adoption of this technology, as many amputees
may prefer to use simpler prostheses rather than undergo brain surgery.

2.4 Restoration of Limb Function After Spinal Cord Injury

While the previous section demonstrated the use of BCIs for control of arti-
ficial limbs, a similar principle could be used by people with spinal cord
injury, who still have all their limbs but have lost the nerves connecting
the brain and the limb. In the past, restoration of limb function in people
with spinal cord injury was frequently done with functional electrical stim-
ulation, where the remaining muscles were artificially stimulated in a coor-
dinated pattern (generated by, e.g., a finite state machine) in order to move
the limbs (Ho et al., 2014). However, such electrical stimulation frequently
results in unnatural and/or unstable motion patterns (e.g., “robotic” gait).
A more natural alternative would be to use a BCI to guide functional elec-
trical stimulation of the limb, thus achieving more intuitive and stable con-
trol than could be achieved with an artificial control system. The same
approach could also be used with other assistive devices such as exoskeletons.
As with artificial limbs, such BCI-guided restoration of limb function
mainly relies on invasive systems to achieve the necessary signal quality.
A proof-of-concept BCI system that used intracortical electrodes to control
an implanted functional electric stimulator was recently presented by
Ajiboye et al. (2017) for reaching and grasping motions in tetraplegia.
463days after device implantation, the single study participant was able to