Clinical recovery after stroke can be significant and has been attributed to plastic reorganization and recruitment of novel areas previously not engaged in a given task. As equivocal results have been reported in studies using single imaging or electrophysiological methods, here we applied an integrative multimodal approach to a group of well-recovered chronic stroke patients (n = 11; aged 50-81 years) with left capsular lesions. Focal activation during recovered hand movements was assessed with EEG spectral analysis and H2(15)O-PET with EMG monitoring, cortico-cortical connectivity with EEG coherence analysis (cortico-cortical coherence) and corticospinal connectivity with transcranial magnetic stimulation (TMS). As seen from comparisons with age-matched controls, our patients showed enhanced recruitment of the lateral premotor cortex of the lesioned hemisphere [Brodmann area (BA) 6], lateral premotor and to a lesser extent primary sensorimotor and parietal cortex of the contralesional hemisphere (CON-H; BA 4 and superior parietal lobule) and left cerebellum (patients versus controls, Z > 3.09). EEG coherence analysis showed that after stroke cortico-cortical connections were reduced in the stroke hemisphere but relatively increased in the CON-H (ANOVA, contrast analysis, P <0.05), suggesting a shift of functional connectivity towards the CON-H. Nevertheless, fast conducting corticospinal transmission originated exclusively from the lesioned hemisphere. No direct ipsilateral motor evoked potentials (MEPs) could be elicited with TMS over the contralesional primary motor cortex (iM1) in stroke patients. We conclude that (i) effective recovery is based on enhanced utilization of ipsi- and contralesional resources, (ii) basic corticospinal commands arise from the lesioned hemisphere without recruitment of ('latent') uncrossed corticospinal tract fibres and (iii) increased contralesional activity probably facilitates control of recovered motor function by operating at a higher-order processing level, similar to but not identical with the extended network concerned with complex movements in healthy subjects.
