Diverse functional connectivity patterns of resting-state brain networks associated with good and poor hand outcomes following stroke

Abstract

Motor stroke has been characterized by disruptions in multiple large-scale functional brain networks. However, it remains unclear whether stroke patients with good hand outcomes show different connectivity profiles within and between networks from those with poor hand outcomes. In this cross-sectional study, we recruited 52 chronic subcortical stroke patients [illness duration (mean ± SD): 16 ± 16.2 months] and 52 healthy controls from the local hospital and community from June 2010 to August 2016. We first performed independent component analysis (ICA) on resting-state fMRI data to extract fifteen resting-state networks. Then, we compared the functional connectivity within and between networks across 52 healthy controls, 26 patients with a partially paralyzed hand (PPH), and 26 patients with a completely paralyzed hand (CPH). Compared to the patients with a PPH, the patients with a CPH showed increased connectivity in the contralesional sensorimotor cortex within the contralesional sensorimotor network; the increased connectivity was negatively correlated with the performance of the paretic hand. Moreover, the patients with a CPH, compared to those with a PPH, showed decreased strengths of connectivity between the ipsilesional sensorimotor network and both the dorsal sensorimotor network and ventral visual network; the decreased strengths of connectivity were positively correlated with the performance of the paretic hand. Collectively, our findings suggest that stroke patients with different hand outcomes show distinct functional reorganization patterns in large-scale brain networks. These findings shed light on the network-level neuromechanisms that help explain why stroke survivors in the chronic stage show different hand outcomes.

Keywords: Functional brain network; Functional connectivity; ICA; Independent component analysis; Resting-state functional magnetic resonance imaging; Stroke.

Fig. 1

Lesion overlap map for the partially paralyzed hand (PPH) group, the completely paralyzed hand (CPH) group and all stroke patients. The color bar indicates the number of patients having lesions in each voxel. Left indicates the ipsilesional hemisphere.

Fig. 2

Fifteen resting-state brain networks were identified using independent component analysis. The color bar represents the T values ranging from 8 to 20. Left indicates the ipsilesional hemisphere.

Fig. 3

Disrupted functional connectivity within networks among the three groups. (A) Disrupted within-network FC patterns between each pair of the PPH, CPH and controls, which are rendered in the 3D images (q < 0.01, FDR corrected). (B) The violin plot displays the intergroup differences in the zFC values in the contralesional SMC within the contralesional sensorimotor network among the PPH, CPH and controls. (C) The zFC values of the contralesional SMC are negatively correlated with the FMA-HW scores across all stroke patients (Bonferroni corrected). The color bars represent the T values from the intergroup comparisons. CPH, completely paralyzed hand; PPH, partially paralyzed hand; FMA-HW, Fugl-Meyer Assessment of Hand and Wrist; FC, functional connectivity; SMC, sensorimotor cortex. Left indicates the ipsilesional hemisphere.

Fig. 4

Disrupted functional connectivity between networks among the three groups. (A) Diverse functional network connectivity patterns between each pair of the PPH, CPH and controls. The solid and dotted lines represent decreased and increased between-network connectivity, respectively (q < 0.05, FDR corrected). (B) The violin plot displays the intergroup differences in connectivity between the iSMN and the vVN and between the iSMN and the dSMN among the PPH, CPH and controls. (C) The functional network connectivity of the iSMN with the vVN and the iSMN with the dSMN are positively correlated with the FMA-HW scores across all stroke patients (Bonferroni corrected, except for the correlation between iSMN-vVN connectivity and FMA-HW scores.). CPH, completely paralyzed hand; PPH, partially paralyzed hand; i/c/d/v SMN, ipsilesional/contralesional/dorsal/ventral sensorimotor network; i/c FPN, ipsilesional/contralesional frontoparietal network; AN, auditory network; v/d VN, ventral/dorsal visual network; FMA-HW, Fugl-Meyer Assessment of Hand and Wrist; ns, not significant. Left indicates the ipsilesional hemisphere. BrainNet Viewer was used for 3D surface visualization (www.nitrc.org/projects/bnv) (Xia et al., 2013).