Network dynamics engaged in the modulation of motor behavior in stroke patients

Abstract

Stroke patients with motor deficits typically feature enhanced neural activity in several cortical areas when moving their affected hand. However, also healthy subjects may show higher levels of neural activity in tasks with higher motor demands. Therefore, the question arises to what extent stroke-related overactivity reflects performance-level-associated recruitment of neural resources rather than stroke-induced neural reorganization. We here investigated which areas in the lesioned brain enable the flexible adaption to varying motor demands compared to healthy subjects. Accordingly, eleven well-recovered left-hemispheric chronic stroke patients were scanned using functional magnetic resonance imaging. Motor system activity was assessed for fist closures at increasing movement frequencies performed with the affected/right or unaffected/left hand. In patients, an increasing movement rate of the affected hand was associated with stronger neural activity in ipsilesional/left primary motor cortex (M1) but unlike in healthy controls also in contralesional/right dorsolateral premotor cortex (PMd) and contralesional/right superior parietal lobule (SPL). Connectivity analyses using dynamic causal modeling revealed stronger coupling of right SPL onto affected/left M1 in patients but not in controls when moving the affected/right hand independent of the movement speed. Furthermore, coupling of right SPL was positively coupled with the "active" ipsilesional/left M1 when stroke patients moved their affected/right hand with increasing movement frequency. In summary, these findings are compatible with a supportive role of right SPL with respect to motor function of the paretic hand in the reorganized brain.

Keywords: dynamic causal modeling; effective connectivity; movement frequency; reorganization; superior parietal lobe.

Figure 1

Behavioral results: differences between stroke patients and healthy controls (n = 2 × 11). Patients featured both reduced maximum fingertapping frequencies and reduced maximum grip strength of their stroke‐affected (right) hand compared to healthy controls

Figure 2

(a) Neural activity for visually paced fist closures (main effect “hand”) and the parametric modulation of “frequency” (n = 2 × 11; p < .05, FWE‐corrected on the cluster level; cluster forming threshold p < .001, uncorrected). For the main effect “hand,” a network of cortical and subcortical areas is activated comprising the bilateral V1 and extrastriate cortex, the contralateral M1, bilateral SMA, bilateral ventral premotor cortex (PMv) and bilateral PMd along the precentral sulcus, somatosensory regions (S1, S2), and the thalamus. For the parametric modulation of “frequency,” the contralateral M1 and in the right ventral premotor region are activated. (b) Differences between stroke patients and healthy controls (n = 2 × 11; p < .001, uncorrected). Patients show a stronger frequency‐dependent increase within contralesional superior parietal lobule (SPL) when moving their affected hand as compared to healthy controls [Color figure can be viewed at http://wileyonlinelibrary.com]

Figure 3

(a) Endogenous connectivity in healthy controls that is largely symmetrically organized across both hemispheres (DCM A‐matrix; n = 11; p < .05, FDR‐corrected). (b) Differences between stroke patients and healthy controls (DCM A‐matrix; n = 2 × 11; p < .05, FDR‐corrected). Neural coupling strength exerted from SPL within the unaffected hemisphere upon affected M1 was significantly stronger in patients as compared to healthy controls. Green arrows = positive coupling rates; red arrows = negative coupling rates. The width of each arrow corresponds to the coupling strength. For mean coupling parameters and p values (one‐sample t test against zero), see Supporting Information, Table 2 [Color figure can be viewed at http://wileyonlinelibrary.com]

Figure 4

(a) Modulatory effects on effective connectivity (top = main effect “hand,” bottom = parametric modulation “frequency”) during right‐ and left‐hand fist closures in healthy controls (DCM B‐matrix; n = 11; p < .05, FDR‐corrected). Neural activity within the contralateral “active” M1 was especially mediated by stronger coupling with SMA, PMd, and SPL. (b) Differences between stroke patients and healthy controls (top = main effect “hand,” bottom = parametric modulation “frequency”) (DCM B‐matrix; n = 2 × 11; p < 0.05, FDR‐corrected). The patients' group shows a stronger excitatory influence from ipsilateral/contralesional SPL upon affected M1 as compared to control subjects. Patients further showed an increase in the “frequency‐dependent modulatory influence on the connection from ipsilateral/contralesional SPL upon affected M1 as compared to control subjects.” Green arrows = positive coupling rate; red arrows = negative coupling rates. The width of each arrow corresponds to the coupling strength. For mean coupling parameters and p values (one‐sample t test against zero), see Supporting Information, Table 3 [Color figure can be viewed at http://wileyonlinelibrary.com]