Layer-specific cortical signatures uncover a sensory origin of post-stroke motor dysfunction

The cerebral cortex computes through a laminar microcircuit in which superficial layers integrate cortico cortical input and deep layers issue corticospinal output. How focal injury disrupts this input output architecture has been characterized in animal models, but has never been resolved in the human brain, leaving it unknown whether stroke degrades the motor cortex uniformly or dissociates its computational compartments. We used ultra high field 7T functional MRI to resolve activation across cortical depth in the primary motor cortex (M1) hand knob during finger tapping (12 stroke patients/14 controls), and related layer-specific signals to comprehensive motor assessments through principal component analysis, correlation, and cross validated predictive modeling. In the ipsilesional hemisphere, both the input (L2/3) and output (L5) layers showed reduced activation during affected-hand tapping movement in patients compared to controls, indicating a combined failure of sensorimotor integration and corticospinal output. In the contralesional hemisphere, the two compartments dissociated and tracked distinct behavioral processes: superficial (L1) activation varied with global motor outcome, whereas deep output layer (L5/L6) activation tracked specifically with affected hand grip force. Cross validated modeling confirmed this double dissociation: superficial activation predicted global outcome, deep activation predicted grip force, and no ipsilesional layer predicted either. These findings provide the first human evidence that stroke does not disrupt the motor cortex uniformly but reorganizes at the level of individual laminar compartments, with the superficial input stage and deep output stage indexing separable aspects of motor dysfunction. Beyond linking cortical microcircuit models derived from animals to the organization of impairment in humans, layer resolved imaging reveals prognostic information inaccessible to whole region measures, pointing toward laminar signatures that could help stratify motor deficits and guide targeted rehabilitation.