Researchers here report on the use of a combination of a brain-computer interface and functional electrical stimulation of muscles to bypass damage leading to paralysis of the hand, allowing some degree of restored function. The approach was demonstrated in non-human primates in which nerves connecting the hand to the brain were damaged via surgery. Competition in approaches to the problem of nervous system damage is a good thing, but one would hope that this class of application of brain-computer interface is largely made irrelevant by future advances in regenerative medicine.
Paralysis following stroke is a leading cause of long-term motor disability. Brain machine interfaces (BMIs) can transform cortical activity into control signals for an external device, such as a robotic arm or computer cursor, and may provide a solution for restoring lost function. Bypassing the damaged pathway using brain-controlled functional electrical stimulation (FES) to regain volitional control of the paralysed limb is promising for restoring lost motor function. Brain-controlled FES works as an "artificial" neural pathway by creating a causal relationship between brain activity and an evoked limb movement. However, subjects may be required to learn a novel causal input-output relationship to control the paralysed limb.
Disruption of descending pathways, as can result from stroke, results in a lost connection between the brain and target muscles. Functional recovery in such a situation is characterised by substantial reorganisation in the structure and function of the damaged brain. Thus, our nervous system shows remarkable flexibility to adapt to novel neuromotor mappings. How the brain incorporates a novel "artificial" neural pathway into volitional limb control within the surviving cortical areas remains largely unclear.
In the present study, we generated a model of chronic hemiparalysis in the extremities caused by subcortical stroke in monkeys. We then employed an artificial cortico-muscular connection (ACMC) to connect the preserved cortical areas to muscles beyond the damaged site. Specific neural oscillations in the cortical area were detected contingent to the input and converted into electrical stimulation delivered to the muscles in real time. We demonstrated that, despite damage to subcortical areas, a flexible change in the neural oscillations controlling the ACMC was observed in a targeted manner throughout an extensive sensorimotor area. Thus, monkeys that experienced a subcortical stroke could rapidly learn to regain lost volitional control of a paralysed hand.