Stroke is the dominant clinical population for rehab device deployment in absolute volume. Recovery follows a stereotyped time-course with greatest plasticity in the first 90 days, plateauing thereafter — though long-term gains remain achievable with sustained dose. This chapter walks upper-limb and lower-limb modalities, FES integration, BCI-gated paradigms, and the standardised assessment battery.
Robotics (per Chapter 2), constraint-induced movement therapy (CIMT) per Taub + Uswatte + Pidikiti 1999, mirror therapy per Altschuler 1999, mental practice / motor imagery per Page 2001, and FES-cycling are the principal modalities. Modality choice tracks the Brunnstrom stage + Fugl-Meyer score: severe paresis benefits from FES + robotics + mental practice; moderate paresis from CIMT + robotics; mild paresis from CIMT + task-specific repetitive practice.
Body-weight-supported treadmill training (BWSTT), robotic gait trainers (Lokomat, G-EO, ReWalk in non-ambulatory settings), overground robotic exoskeletons, FES for foot-drop (per IEC 60601-2-10 + the H100 / H200 / Bioness L300 / WalkAide platform), and conventional gait training. Outcome battery: 6MWT, 10MWT, BBS, Functional Ambulation Category (FAC per Holden 1984).
FES per IEC 60601-2-10: pulse-width 200–500 µs, pulse-frequency 20–50 Hz, amplitude per-patient sensory-motor-pain threshold envelope. Cycling-mode FES drives quadriceps + hamstring + glute alternation against a fixed cadence. Gait-mode FES gates dorsiflexor stim on swing-phase detection (insole-based or IMU-based). EMG-triggered FES gates stim on patient-initiated muscle activity, preserving the volitional drive.
BCI-gated paradigms per Daly + Wolpaw 2008 + Soekadar + Birbaumer + Slutzky + Cohen 2015 add a motor-intent decoding layer ahead of the FES or robotic actuator. EEG headsets (Emotiv, OpenBCI, Brain Products, ANT Neuro) feed a motor-imagery classifier; the classifier output gates the per-trial stim or robotic motion. Closed-loop BCI rehab remains research-track; pivotal-evidence is accumulating but commercial deployment is constrained.
Fugl-Meyer Assessment (FMA-UE for upper extremity, FMA-LE for lower extremity), Action Research Arm Test (ARAT), Box-and-Block Test (Mathiowetz 1985), Wolf Motor Function Test, Modified Ashworth Scale (spasticity), Berg Balance Scale, and the Stroke Impact Scale (patient-reported) form the core battery. The American Heart Association / American Stroke Association guidelines specify the minimum reporting envelope per outcome study. Each measure has a documented minimal-clinically-important-difference (MCID) and a documented minimal-detectable-change (MDC) per the source psychometric literature; reporting a numerical change without naming the MCID it crossed (or did not cross) leaves the clinical significance undefined.