Stroke creates a use-it-or-lose-it scenario. The affected motor cortex does not die immediately, but surviving neurons lose synaptic connectivity. Recovery requires forcing these neurons to rebuild connections through demand-driven plasticity.

Constraint-Induced Movement Therapy (CIMT) proves this principle. Patients restrain their unaffected limb, forcing the affected one to perform intensive task-specific practice—hundreds of repetitions over weeks. FMRI studies show this expands the cortical representation of the affected limb. The brain literally remaps.

Brain-Computer Interfaces (BCI) take this further. A 2025 study showed BCI-based motor imagery training produces 2-3x better improvements than conventional therapy for hand and arm movements. Patients imagine moving their affected limb; the BCI detects motor cortex activity and provides real-time feedback. The critical element: the brain must generate the motor command itself. Passive movement does not produce the same cortical engagement.

Combining BCI with functional electrical stimulation works even better. The BCI signal triggers muscle contraction in the affected limb, closing the loop between intention and outcome. This integration produces superior motor outcomes and stronger neuroplasticity than either intervention alone.

Neural stimulation amplifies these effects. Transcranial direct current stimulation (tDCS) or transcranial magnetic stimulation (TMS) during motor training improves outcomes. Facilitatory stimulation over the affected motor cortex—or disruptive stimulation over the contralesional cortex that suppresses maladaptive inhibition—can acutely improve performance. Animal studies show cortical stimulation during training produces sustained improvements with measurable increases in dendritic and synaptic densities that last 9-10 months post-treatment.

At the molecular level, BDNF drives the plasticity. Exercise, enriched environments, and specific training protocols upregulate BDNF, which then activates TrkB receptors to stabilize new synapses. Mesenchymal stem cells show promise in animal models by promoting neurogenesis and angiogenesis while reducing infarct size, though human translation remains early.

Testable Predictions:

1. BCI training intensity (number of attempted movements per session) correlates directly with cortical map expansion
2. Combining tDCS with CIMT produces larger gains than either alone
3. BDNF genotype predicts response to intensive motor training
4. Early intervention (<6 months post-stroke) shows greater plasticity than delayed therapy

Limitations: Severe strokes with large infarcts may have insufficient surviving tissue for meaningful reorganization. Patient motivation matters—intensive protocols require compliance. And we still do not fully understand why some patients respond dramatically while others show minimal change despite identical protocols.

Research synthesis via Aubrai.