Mirror therapy (MT) has been proposed to promote motor recovery post-stroke through activation of mirror neuron system, recruitment of ipsilateral motor pathways, or/and increasing attention toward the affected limb. However, neuroimaging evidence for these mechanisms is still lacking. To uncover the underlying mechanisms, we designed a randomized controlled study and used a voxel-based whole-brain analysis of resting-state fMRI to explore the brain reorganizations induced by MT. Thirty-five stroke patients were randomized to an MT group (n = 16) and a conventional therapy (CT) group (n = 19) for a 4-week intervention. Before and after the intervention, the Fugl-Meyer Assessment Upper Limb subscale (FMA-UL) and resting-state fMRI were collected. A healthy cohort (n = 16) was established for fMRI comparison. The changes in fractional amplitude of low-frequency fluctuation (fALFF) and seed-based functional connectivity were analyzed to investigate the impact of intervention. Results showed that greater FMA-UL improvement in the MT group was associated with the compensatory increase of fALFF in the contralesional precentral gyrus (M1) region and the re-establishment of functional connectivity between the bilateral M1 regions, which facilitate motor signals transmission via the ipsilateral motor pathways from the ipsilesional M1, contralesional M1, to the affected limb. A step-wise linear regression model revealed these two brain reorganization patterns collaboratively contributed to FMA-UL improvement. In conclusion, MT achieved motor rehabilitation primarily by recruitment of the ipsilateral motor pathways. Trial Registration Information: http://www.chictr.org.cn. Unique Identifier. ChiCTR-INR-17013644, submitted on December 2, 2017.
