Deep brain stimulation: Connectivity profile for bradykinesia alleviation

Subthalamic deep brain stimulation may alleviate bradykinesia in Parkinson patients. Research suggests that this stimulation effect may be mediated by brain networks like the corticocerebellar loop. This study investigated the connectivity between stimulation sites and cortical and subcortical structures to identify connections for effective stimulation.

We retrospectively investigated 21 patients with Parkinson disease with bilateral subthalamic deep brain stimulation. Stimulation effectiveness in reducing bradykinesia, tremor, and rigidity was evaluated for each electrode contact in brain hemispheres contralateral to the affected hemibody. Dysarthric side effects were also examined. Probabilistic tractography based on diffusion-weighted imaging was performed in individual patient-specific brains using electrode contacts as seeds. Connectivity profiles of contacts with effective and noneffective stimulation were compared.

Connectivity profiles of effective and noneffective contacts differed. Moreover, the connectivity profile for bradykinesia differed from that for rigidity, tremor, or dysarthria. Regarding bradykinesia, effective contacts were significantly more often connected with the ipsilateral superior cerebellar peduncle and the ipsilateral dentate nucleus, which correspond to the ipsilateral portion of the cerebellothalamocortical pathway. Rigidity was mitigated by stimulation of ascending brainstem and intralaminar thalamic connections. Tremor alleviation was related to connections with the internal capsule (anterior limb) and the pallidum. Dysarthric side effects were associated with connections to the supplementary motor area and the decussating cerebellothalamocortical pathway.

Whereas bradykinesia seems to be mitigated by stimulation of the ascending, ipsilateral cerebellothalamocortical pathway, stimulation of the descending corticopontocerebellar pathway may be ineffective. Rigidity, tremor, and dysarthric side effects seem to be influenced by different neural networks. ANN NEUROL 2019;85:852-864.