Constraint-induced movement therapy after ischemic stroke: From cortical reorganization to molecular plasticity

Abstract

Stroke recovery is a dynamic, multilevel process involving structural, functional, and molecular reorganization of surviving neural circuits. Constraint-induced movement therapy (CIMT) has emerged as one of the most effective rehabilitation strategies for improving upper-limb motor function after stroke by overcoming learned nonuse and promoting activity-dependent plasticity. This review integrates clinical and experimental evidence to provide a comprehensive mechanistic framework of CIMT-induced recovery. At the systems level, motor recovery is governed by peri-infarct plasticity, diaschisis reversal, and behavioral adaptation within a temporally defined plasticity window. Experimental studies using middle cerebral artery occlusion (MCAO) models demonstrate that CIMT enhances synaptic remodeling, neurogenesis, angiogenesis, and neurovascular adaptation in peri-infarct regions. At the molecular level, CIMT engages coordinated signaling pathways, including brain-derived neurotrophic factor (BDNF)–TrkB signaling, synaptic structural proteins (GAP-43, synaptophysin, PSD-95), modulation of axonal growth inhibition via the Nogo-A/ROCK pathway, and regenerative processes that support neural circuit reorganization. Comparative analysis indicates that CIMT primarily drives task-specific cortical plasticity, whereas treadmill exercise promotes global neuroplasticity through neurotrophic and neurovascular mechanisms. Combined rehabilitation strategies integrate these complementary effects, leading to synergistic recovery across multiple functional domains. From a translational perspective, optimal rehabilitation likely requires a multimodal framework integrating task-specific, aerobic, and neuromodulatory interventions. Future research should focus on identifying biomarkers of responsiveness and defining optimal timing, intensity, and sequencing of therapies to enable personalized rehabilitation strategies.