Physical Medicine

BCI-controlled ankle robot modestly outperforms robot-only training for post-stroke lower-limb recovery

In subacute stroke, 2 weeks of BCI-controlled ankle robot training improved lower-limb motor outcomes beyond robot training alone. Both groups improved dorsiflexion AROM, dorsiflexor strength, FMA-LE, Berg, and FAC, but FMA-LE gains were significantly larger with BCI assistance. The BCI group also showed greater calf spasticity reduction on the Modified Ashworth Scale, suggesting added benefit for tone management. qEEG changes with BCI—reduced slow-wave power and interhemispheric asymmetry, increased alpha power and connectivity—support training-induced neuroplastic reorganization. 1

Objective post-stroke sleep trajectories strongly predict 12-month neurological recovery

Actigraphy across acute, recovery, and chronic phases identified four distinct sleep quality trajectories in 306 stroke patients. Consistently poor sleep and long sleep with reduced efficiency and worsening fragmentation markedly increased odds of poor 12‑month mRS outcomes. Compared with consistently good sleepers, these unfavorable trajectories carried roughly 5–7‑fold higher risk of poor neurological recovery. Findings support routine objective sleep assessment after stroke and proactive, individualized sleep management as part of rehabilitation. 2

TWIST tool retains strong discrimination for time to independent walking after stroke

The TWIST tool uses age, knee extension strength, and Berg Balance at 1 week to predict time to independent walking up to 26 weeks. In this temporal external validation of 89 non-walking stroke inpatients, discrimination remained good at all timepoints (C-statistic >0.8). Calibration was optimistic for some TWIST scores, particularly 0 and 2 at 16–26 weeks and 3 at all timepoints, prompting refined interpretations. TWIST now offers clinicians a validated, simple framework for early walking prognostication and goal-setting using routine measures. 3