Two philosophies: replace versus restore
In Volumes I and II you met BCIs that decode intention to drive a cursor or a robotic arm — they replace a lost output channel and are meant to be used forever. A rehabilitation BCI has a different goal. It does not try to bypass the injury permanently; it tries to help the surviving nervous system relearn to do the job itself — restorative versus assistive BCI. The flagship indication is stroke, where a partly-spared corticospinal pathway must be coaxed back into service.
The Hebbian principle, made literal
The engine runs on Hebb's rule: neurons that fire together wire together — Hebbian plasticity induction, the substrate of activity-dependent plasticity. A rehab BCI turns that slogan into an engineering specification: manufacture precisely-timed co-activation between an intention signal and its intended target, over and over, until the connection between them strengthens.
The Hebbian rate rule: the change in a connection weight grows with the product of pre- and post-synaptic activity. A rehab BCI's whole job is to make r_pre (the patient's intention) and r_post (the target's activation) large at the same moment.
The closed loop as therapy
The minimal rehabilitation loop detects intention — for example the event-related desynchronization of the sensorimotor mu rhythm during a motor imagery attempt — and delivers contingent feedback that moves the limb: BCI-controlled functional electrical stimulation or contingent proprioceptive feedback. This closes the sensorimotor loop the injury broke, pairing the cortical command with the sensory consequence — closed-loop rehabilitation.
An honest snapshot
What is genuinely shown: several randomized trials and meta-analyses report that BCI-based training improves upper-limb motor outcomes in chronic stroke more than control. But the trials are small, protocols are heterogeneous, effect sizes are modest, and the overall evidence is graded low-to-moderate — see outcome measures and sham control and blinding.
This honest framing carries through the whole track: the mechanism — plasticity — is plausible and demonstrable at the level of physiology, while durable functional benefit at scale is promising but not settled. Keep that distinction sharp; it is the difference between a real science and a hopeful story.