Multimodal neuroimaging to characterize symptom-specific networks in movement disorders

Abstract

Movement disorders, such as Parkinson's disease, essential tremor, and dystonia, are characterized by their predominant motor symptoms, yet diseases causing abnormal movement also encompass several other symptoms, including non-motor symptoms. Here we review recent advances from studies of brain lesions, neuroimaging, and neuromodulation that provide converging evidence on symptom-specific brain networks in movement disorders. Although movement disorders have traditionally been conceptualized as disorders of the basal ganglia, cumulative data from brain lesions causing parkinsonism, tremor and dystonia have now demonstrated that this view is incomplete. Several recent studies have shown that lesions causing a given movement disorder occur in heterogeneous brain locations, but disrupt common brain networks, which appear to be specific to each motor phenotype. In addition, findings from structural and functional neuroimaging in movement disorders have demonstrated that brain abnormalities extend far beyond the brain networks associated with the motor symptoms. In fact, neuroimaging findings in each movement disorder are strongly influenced by the constellation of patients' symptoms that also seem to map to specific networks rather than individual anatomical structures or single neurotransmitters. Finally, observations from deep brain stimulation have demonstrated that clinical changes, including both symptom improvement and side effects, are dependent on the modulation of large-scale networks instead of purely local effects of the neuromodulation. Combined, this multimodal evidence suggests that symptoms in movement disorders arise from distinct brain networks, encouraging multimodal imaging studies to better characterize the underlying symptom-specific mechanisms and individually tailor treatment approaches.

Fig. 1. Insights from lesion network mapping.

Lesions causing cervical dystonia are found in heterogeneous locations (example lesion locations shown in red, (A)) but are connected to a common network characterized by connectivity to the cerebellum and somatosensory cortex (B). Similarly, heterogeneous lesion locations causing other movement disorders (C) and non-motor symptoms (D) map to common networks, characterized by connectivity to distinct brain regions (here referred to as ‘network hubs’). Lesions causing Holmes tremor mapped to six regions (Table 1) of which the red nucleus, vermis and lateral cerebellum hubs are depicted here; parkinsonism localized to the claustrum; hemichorea-hemiballismus to the posterolateral putamen. Lesions causing depression mapped to the dorsolateral prefrontal cortex; loss of memory to the presubiculum and retrosplenial cortex; disordered agency to the precuneus. Note: Lesions and lesion networks have here been recreated with BioRender.com, functional connectivity profiles are shown with arbitrary units for illustrative purposes only, with warm colors representing positive functional connectivity from lesion locations and cool colors representing negative functional connectivity from lesion locations in (B)–(D). Original publications with exhaustive lesion location information: Cervical dystonia; Holmes tremor; Parkinsonism; Hemichorea; Depression; Memory loss; Disordered agency. AL alien limb syndrome.

Fig. 2. Insights from structural and molecular neuroimaging.

Meta-analyses and systematic reviews of neuroimaging studies have highlighted that structural abnormalities (A) are commonly implicated within the basal ganglia, and motor, frontal and occipital cortex in Parkinson’s disease^,; within thalamic, cerebellar, and cortical motor regions in dystonia^,; and within the cerebellum and cerebellar peduncles in essential tremor. Molecular imaging (B) implicates multiple neurotransmitters associated with symptoms of Parkinson’s disease, including an association between the dopaminergic (modeled in orange) system and bradykinesia/rigidity and apathy, the serotonergic (red) system with tremor and several non-motor symptoms^,, and the cholinergic (green) system with cognition and gait. Neurotransmitter systems are here modeled for visualization purposes only.

Fig. 3. Insights from neuromodulation.

Symptom-specific networks in Parkinson’s disease derived by associating connectome-based structural and functional connectivity profiles from stimulation sites (STN-DBS) to clinical outcomes (A)^,,,. Tracts and structures were obtained from the DBS tractography atlas (Middleborough et al.) and visualized on the BigBrain ultrahigh resolution template. Symptom specific networks in cervical dystonia with GPi-DBS, identified using [¹⁸F]FDG PET imaging (B). Warm colors represent increased metabolism and cool colors represent decreased metabolism. STN subthalamic nucleus, M1 primary motor cortex, Vim ventral intermedial nucleus of the thalamus, PSA posterior subthalamic area, SMA Supplementary motor area, NP Neuropsychiatric, brady bradykinesia, rigid rigidity, S1 primary somatosensory cortex, PMC pre-motor cortex, OC occipital cortex, ACC anterior cingulate cortex, MFC Middle frontal cortex.

Fig. 4. Non-exhaustive levels of evidence towards symptom-specific networks.

Evidence from group-level neuroimaging studies is built upon by validation in independent datasets, and perturbation of suspected symptom networks with neuromodulation. Evidence gathered across these levels can establish symptom-specific networks in movement disorders. Prior research studies provided as examples of each level of evidence: Vaillancourt et al.; Prodoehl et al.; Nagano-Saito et al.; Abe et al.; Möller et al.; Matthews et al.; Niethammer et al.; Joutsa et al.; Bédard et al.; Filip et al.; Chu et al.; Park et al..