Task-free functional MRI in cervical dystonia reveals multi-network changes that partially normalize with botulinum toxin

Abstract

Cervical dystonia is characterized by involuntary, abnormal movements and postures of the head and neck. Current views on its pathophysiology, such as faulty sensorimotor integration and impaired motor planning, are largely based on studies of focal hand dystonia. Using resting state fMRI, we explored whether cervical dystonia patients have altered functional brain connectivity compared to healthy controls, by investigating 10 resting state networks. Scans were repeated immediately before and some weeks after botulinum toxin injections to see whether connectivity abnormalities were restored. We here show that cervical dystonia patients have reduced connectivity in selected regions of the prefrontal cortex, premotor cortex and superior parietal lobule within a distributed network that comprises the premotor cortex, supplementary motor area, primary sensorimotor cortex, and secondary somatosensory cortex (sensorimotor network). With regard to a network originating from the occipital cortex (primary visual network), selected regions in the prefrontal and premotor cortex, superior parietal lobule, and middle temporal gyrus areas have reduced connectivity. In selected regions of the prefrontal, premotor, primary motor and early visual cortex increased connectivity was found within a network that comprises the prefrontal cortex including the anterior cingulate cortex and parietal cortex (executive control network). Botulinum toxin treatment resulted in a partial restoration of connectivity abnormalities in the sensorimotor and primary visual network. These findings demonstrate the involvement of multiple neural networks in cervical dystonia. The reduced connectivity within the sensorimotor and primary visual networks may provide the neural substrate to expect defective motor planning and disturbed spatial cognition. Increased connectivity within the executive control network suggests excessive attentional control and while this may be a primary trait, perhaps contributing to abnormal motor control, this may alternatively serve a compensatory function in order to reduce the consequences of the motor planning defect inflicted by the other network abnormalities.

Conflict of interest statement

Competing Interests: The authors have declared that no competing interests exist.

Figures

Figure 1. Between-group effects for the SMN, ECN and PVN.

Depicted here are the between-group effects for three RSNs. Between-group effects are corrected for family-wise errors (p≤0.013 for A. and B.; p≤0.05 (blue) and p≤0.013 (orange) for C.). A. shows frontal regions and precentral regions abnormally connected to the SMN, indicating decreased connectivity within the CD group. B. shows brain regions linked to the ECN, exhibiting increased connectivity for the CD group. C. The PVN shows CD-related decreased connectivity of several regions including PFC, PMC, SM1, and visual and temporal areas. *The right column (green) shows the original RSNs used in the dual regression approach, thresholded at Z = 2,0. These are PICA spatial maps of healthy subjects derived from Smith et al. Images are t-statistics overlaid on the MNI-152 standard brain. The left hemisphere of the brain corresponds to the right side in this image.

Figure 2. Treatment-related effects for the SMN and PVN.

Depicted here are the treatment-related effects for two RSNs, corrected for family-wise errors (p≤0.05). A. shows the ventral premotor cortex abnormally connected to the SMN, but with an increase of connectivity after BoNT treatment (in orange t = 1>t = 0, in blue t = 1> t = 2). In the sagittal plane, the effect for t = 1> t = 2 (blue) is projected on the left hemisphere for graphical purposes. B. shows areas in the visual cortex and primary motor cortex linked to the PVN, also exhibiting increased connectivity after BoNT treatment. *The right column (green) shows the original RSNs used in the dual regression approach, thresholded at Z = 2,0. These are PICA spatial maps of healthy subjects derived from Smith et al. Images are t-statistics overlaid on the MNI-152 standard brain. The left hemisphere of the brain corresponds to the right side in this image.