Network localization of pediatric lesion-induced dystonia

Abstract

Objective: Dystonia is a movement disorder defined by involuntary muscle contractions leading to abnormal postures or twisting and repetitive movements. Classically dystonia has been thought of as a disorder of the basal ganglia, but newer results in idiopathic dystonia and lesion-induced dystonia in adults point to broader motor network dysfunction spanning the basal ganglia, cerebellum, premotor cortex, sensorimotor, and frontoparietal regions. It is unclear whether a similar network is shared between different etiologies of pediatric lesion-induced dystonia.

Methods: Three cohorts of pediatric patients with lesion-induced dystonia were identified. The lesion etiologies included hypoxia, kernicterus, and stroke versus comparison subjects with acquired lesions not associated with dystonia. Multivariate lesion-symptom mapping and lesion network mapping were used to evaluate the anatomy and networks associated with dystonia.

Results: Multivariate lesion-symptom mapping showed that lesions of the putamen (stroke: r = 0.50, p <0.01; hypoxia, r = 0.64, p <0.001) and globus pallidus (kernicterus, r = 0.61, p <0.01) were associated with dystonia. Lesion network mapping using normative connectome data from healthy children demonstrated that these regional findings occurred within a common brain-wide network that involves the basal ganglia, anterior and medial cerebellum, and cortical regions that overlap the cingulo-opercular and somato-cognitive-action networks.

Interpretation: We interpret these findings as novel evidence for a unified dystonia brain network that involves the somato-cognitive-action network, which is involved in higher order coordination of movement. Elucidation of this network gives insight into the functional origins of dystonia and provides novel targets to investigate for therapeutic intervention.

Figure 1A-D.

Lesion masks of all participants in this study are shown, separated by group. The lesion masks of the 10 participants with HIE are shown in 1A; the region of peak overlap in the putamen contained 8 lesions. The overlap of kernicterus participants is shown in 1B. The maximum number of overlapping lesions (7) was located in the globus pallidus. The peak lesion overlap of the stroke participants was in the putamen (5), shown in 1C. The comparison group had a peak lesion overlap in the right frontal lobe (max overlap=13, MNI 30 41 0) (1D).

Figure 2A-D.

The lesion-symptom mapping analysis was first conducted with the entire sample (2A, r = 0.42, p < 0.001) and showed regions of peak association with dystonia in the putamen and globus pallidus. Lesion symptom maps for each etiology analyzed independently are also shown, including hypoxic ischemic encephalopathy (2B, r = 0.64, p < 0.001), kernicterus (2C, r = 0.61, p < 0.001), and stroke (2D, r = 0.5, p < 0.001). Voxels are scaled to values between 0 and 1, with greater values indicating a greater association with dystonia.

Figure 3.

Dystonia lesion network maps. A) The functional dystonia lesion network map for all etiologies combined is shown in red (scaled between 0 – 11.35) and the structural dystonia lesion network map is shown in blue (scaled between –11.35 and 0.0). Only regions reaching statistical significance after correction for multiple comparisons are shown. The voxel range is T = −9.4 – 12.33, p = 0.0001 for grey matter and T = 0 – 8.65, p = 0.0029 for white matter. B) The functional lesion network map has regional peaks that appear to correspond with the inter-effector nodes of the somato-cognitive action network (green outlined areas) and cingulo-opercular/action-mode network (yellow outlined areas).