Somato-Cognitive Action Network in Focal Dystonia

Abstract

Background: The central pathology causing idiopathic focal dystonia remains unclear. The recently identified somato-cognitive action network (SCAN) has been implicated.

Objective: We tested whether the effector-agnostic SCAN may constitute a central pathology shared across dystonia subtypes, whereas the effector-specific regions in the primary sensorimotor cortex may show distinct functional changes specific to the dystonic body part.

Methods: We collected functional magnetic resonance imaging (MRI) from patients with focal dystonia (laryngeal dystonia [LD], N = 24; focal hand dystonia [FHD], N = 18) and healthy control participants (N = 21). Regions of interest were selected a priori within the basal ganglia-thalamo-cortical and cerebello-thalamo-cortical sensorimotor pathways. We investigated dystonia-dependent resting-state connectivity changes: between SCAN and related cortical regions, between cortical and noncortical regions, and among noncortical regions. Cortical network boundaries were individualized based on resting-state data. Separately, individualized hand and mouth/larynx regions were also generated from task-based MRI (finger-tapping and phonation, respectively) for comparison.

Results: Both focal dystonia subtypes showed significant functional changes (P = 0.048 for LD, P = 0.017 for FHD) compared to controls, driven by SCAN's higher functional connectivity to task-based mouth/larynx region and concomitantly lower connectivity to the cingulo-opercular network. No significant subcortical or cerebellar changes were observed when LD and FHD were modeled as independent groups. However, exploratory analysis combining LD and FHD suggested a dystonia-dependent asynchronization between SCAN and sensorimotor cerebellum (P = 0.010) that may indicate a pathological rather than compensatory process.

Conclusions: We demonstrate that SCAN is uniquely associated with focal dystonia dysfunction beyond the dystonic effector regions, offering insights into pathophysiology and treatments. © 2025 The Author(s). Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Keywords: cerebellum; connectivity, precision functional mapping; disinhibition; fMRI; motor cortex; task‐specific dystonia.

FIG. 1

Overview of cortical–subcortical‐cerebellar pathways in sensorimotor control. The recently revealed somato‐cognitive action network (SCAN) may be involved in focal dystonia pathophysiology. (A) Simplified basal ganglia‐thalamo‐cortical and cerebello‐thalamo‐cortical pathways are shown with the relevant subcortical structures and cerebellum. The effector‐specific (hand, foot, mouth) regions are interdigitated by the SCAN in the primary motor cortex (bilaterally, with only the left hemisphere labeled). (B) Hypothesized functional connectivity (in black arrows) that may demonstrate different SCAN versus effector region involvement in focal dystonia affecting different body parts. We also explored connectivity within subcortical‐cerebellar circuitry. AMN, action‐mode network; Caud, caudate; CON, cingulo‐opercular network; FC, functional connectivity; GPe, globus pallidus externus; GPi, globus pallidus internus; Putm, putamen; smCB, sensorimotor cerebellum; STN, subthalamic nucleus; Thal, thalamus. Created in BioRender. Wang, Y. (2025) https://BioRender.com/z11g018. [Color figure can be viewed at wileyonlinelibrary.com]

FIG. 2

Surface and volumetric regions of interest (ROI). The affected effector regions (hand for focal hand dystonia [FHD], mouth/larynx for laryngeal dystonia [LD]) were defined from task and resting‐state scans for comparison. Group atlases were used for subcortical and cerebellar volumetric ROIs. (A) Task hand and task vocal ROIs are the peak activation regions (yellow) from finger‐tapping and phonation tasks, respectively, in an individualized manner (see Figs. S1 and S2). Task hand ROI on each hemisphere was separately generated from contralateral index finger‐tapping and combined, whereas bilateral task vocal ROI was from phonating/i:/. An a priori mask with pre‐ and postcentral gyrus (blue and tortoise outline) and central sulcus (red outline) was used to limit the task‐based effector ROI to the primary sensorimotor cortex (Destrieux et al., 2010). (B) The resting‐state parcellation algorithm divided the cortex into 18 nonoverlapping functional networks per hemisphere, which included rest hand, rest mouth, SCAN, CON(AMN), and (rest) foot (Gordon et al., 2017, 2023). This iterative parcellation algorithm adjusted individual network boundaries based on k‐means clustering (Fig. S3). All surface ROIs were on the fsaverage6 surface. (C) Subcortical volumetric ROIs were obtained from atlases (Pauli et al., 2018, Tian et al., 2020) in MNI152 space and resampled to an isotropic voxel size of 2 mm. (D) Sensorimotor part of the cerebellum was obtained (Buckner et al., 2011) in MNI152 space and resampled to an isotropic voxel size of 2 mm. AMN, action‐mode network; Caud, caudate; CON, cingulo‐opercular network; GPe, globus pallidus externus; GPi, globus pallidus internus; Putm, putamen; SCAN, somato‐cognitive action network; smCB, sensorimotor cerebellum; STN, subthalamic nucleus; THa, anterior thalamus; THp, posterior thalamus. [Color figure can be viewed at wileyonlinelibrary.com]

FIG. 3

Individualized cortical regions of interest (ROIs) and linear model results. (A) Individualized cortical ROIs are shown on the fsaverage6 surface, with the color scale indicating percentage overlap for all subjects in our study after quality control (N = 57 for task hand, N = 56 for task vocal, N = 61 for resting‐state parcellation). Foot ROI was included as an active control region. (B) Cortical functional connectivity (FC) changes involving SCAN were tested with both task‐derived and resting‐state ROIs. With task‐based effector ROI (left), SCAN demonstrated a significant interaction effect for both laryngeal dystonia (LD) and focal hand dystonia (FHD) groups compared to control (P = 0.048, P = 0.017, respectively). This interaction was driven by SCAN's higher connectivity with task vocal ROI and lower connectivity with CON (ie, AMN) in focal dystonia groups. No significant dystonia‐dependent effects were found with resting‐state effector ROIs (right). (C) We tested if cortical effector regions and/or SCAN showed FC changes with closely synapsed basal ganglia and cerebellar circuitry. Task‐based and resting‐state cortical ROIs were included in separate linear models. We did not find significant cortico‐subcortical or cortical‐cerebellar connections related to LD or FHD. For all linear model results, raw FC values after r‐to‐z transform were plotted with an overlay of estimated marginal mean (EMM), which accounts for covariates (age, sex, symptom duration, and mean relative motion during resting‐state scan). All error bars indicate 95% confidence interval of the EMM. AMN, action‐mode network; Caud, caudate; CON: cingulo‐opercular network; GPe, globus pallidus externus; GPi, globus pallidus internus; Putm, putamen; SCAN: somato‐cognitive action network; smCB, sensorimotor cerebellum; STN, subthalamic nucleus; THa, anterior thalamus; THp, posterior thalamus.

FIG. 4

Exploratory investigation of the functional connectivity between SCAN and sensorimotor cerebellum in focal dystonia. (A) The dystonia group combining laryngeal dystonia (LD) and focal hand dystonia (FHD) showed significantly different FC (P = 0.010) compared to controls. Specifically, there was a weak negative FC between SCAN and sensorimotor cerebellum in control subjects, which diminished to around zero in the combined dystonia group. This may suggest an asynchronization of the cortico‐cerebellar sensorimotor connection with a loss of tonic inhibition at rest in focal dystonia. (B) Dystonia severity was negatively correlated with symptom duration (P = 0.007). Severity measures of dystonia (VHI and ADDS) were normalized (mean‐centered and divided by the standard deviation) to allow being pooled for correlation tests. Higher normalized severity measure indicates greater symptoms. (C) There was a trend toward positive correlation of SCAN‐cerebellum FC against severity (P = 0.083) and (D) a trend toward negative correlation against symptom duration (P = 0.073). ADDS, Arm Dystonia Disability Scale; EMM, estimated marginal mean; FC, functional connectivity; SCAN, somato‐cognitive action network; smCB, sensorimotor cerebellum; VHI, Voice Handicap Index. [Color figure can be viewed at wileyonlinelibrary.com]