We wished to characterize changes in function of the executive motor system individually and to correlate these with the degree of recovery from motor disability following stroke. Six male stroke patients underwent perfusion positron emission tomography scanning while undertaking a repetitive, graded force exertion task with the hand affected by stroke, and at rest. We correlated regional cerebral blood flow (rCBF) and force exerted, and compared the distribution of blood flow in all active states with that at rest. The statistical parametric maps of task-related flow changes that were generated were coregistered with anatomical magnetic resonance imaging scans. A comparison of force exertion with rest (categorical comparison) in the group as a whole showed fewer activations than found in normal subjects in both the infarcted and unaffected hemispheres, indicating disruption of the entire motor network. We found a polynomial correlation between synaptic activity (as indexed by blood flow change) and force exerted in 3 patients in the contralateral primary sensorimotor cortex, distinctly different from the logarithmic curve in normal volunteers. For the group there was an initial steep rise of relative rCBF at lower exerted forces, a plateau, and then a second steep rCBF increase at a force of approximately 50% of maximal voluntary contraction (MVC). The ipsilateral ventral posterior supplementary motor area (pSMA) and parietal areas showed correlated activity with force exerted that was not found in normal subjects. Coregistered functional and anatomical images in individuals indicated considerable intersubject variability in the patterns of activation of the contralateral primary motor cortex, pSMA, anterior cingulate cortex, dorsolateral prefrontal cortex, anterior opercular cortex, and parietal cortex. Compensatory actuation of cortical areas that comprise components of the sensorimotor system not normally activated by the task used favours the concept that recruitment of preexisting corticocortical and possibly parallel corticospinal pathways plays a prominent role in functional reorganization. Ipsilateral force-correlated rCBF changes were sometimes present in dorsolateral premotor, insular and parietal cortex consistent with bilateral representation of movements in these motor-associated areas. Quantitative analysis of the rCBF-force relationship provides clearer evidence for large-scale functional reorganization than simple task-control comparisons. The binominal relationship between rCBF and force suggests increased synaptic activity at approximately 50% MVC, a level at which many stroke patients experience an increased sense of effort.