The aim of the study was to examine whether neural state of spared motor and premotor cortices captured before a therapy predicts therapy-related motor gains in chronic subcortical stroke.Design
Ten survivors, presenting chronic moderate upper limb impairment, underwent proton magnetic resonance spectroscopy, magnetic resonance imaging, clinical, and kinematics assessments before a 4-wk impairment-oriented training. Clinical/kinematics assessments were repeated after therapy, and motor gain was defined as positive values of clinical upper limb/elbow motion changes and negative values of trunk motion changes. Candidate predictors were N-acetylaspartate-neuronal marker, glutamate-glutamine-indicator of glutamatergic neurotransmission, and myo-inositol-glial marker, measured bilaterally within the upper limb territory in motor and premotor (premotor cortex, supplementary motor area) cortices. Traditional predictors (age, stroke length, pre-therapy upper limb clinical impairment, infarct volume) were also investigated.Results
Poor motor gain was associated with lower glutamate-glutamine levels in ipsilesional primary motor cortex and premotor cortex (r = 0.77, P = 0.01 and r = 0.78, P = 0.008, respectively), lower N-acetylaspartate in ipsilesional premotor cortex (r = 0.69, P = 0.02), higher glutamate-glutamine in contralesional primary motor cortex (r = −0.68, P = 0.03), and lower glutamate-glutamine in contralesional supplementary motor area (r = 0.64, P = 0.04). These predictors outperformed myo-inositol metrics and traditional predictors (P ≈ 0.05–1.0).Conclusions
Glutamatergic state of bilateral motor and premotor cortices and neuronal state of ipsilesional premotor cortex may be important for predicting motor outcome in the context of a restorative therapy.