The interpretation of deep brain stimulation through diffusion-weighted magnetic resonance imaging

Thesis

Advances in neuroimaging, particularly magnetic resonance imaging (MRI), have been fundamental to achieving improvements in clinical outcomes from deep brain stimulation (DBS). These improvements are essentially an increase in the reliability of the procedure as a result of improved accuracy in targeting a neuroanatomic structure of interest. However, many DBS procedures still result in highly variable outcomes between patients, raising the possibility that targeting remains suboptimal in these cases. These procedures may benefit from advanced forms of MRI that can extract unutilised anatomic features for a more individualised approach. Diffusion MRI tractography is a prime candidate for this but has to date mostly been applied to low outcome-variance procedures or with normative data which belies any project of individualisation. This thesis examines three cases of high outcome-variance DBS procedures and interprets them with patient-specific structural connectomes. DBS of the pedunculopontine nucleus in Parkinson’s patients with severe gait freezing was interpretable by stimulation-associated connectivity to primary motor cortex (M1), cerebellum, and caudal primary somatosensory cortex. This suggests connectivity maps of the ventrolateral pontine tegmentum could be useful as an adjunct to targeting. DBS of the ventrocaudal nucleus of thalamus (Vc) in patients with chronic neuropathic pain was interpretable through a homuncular framework superimposed on an individualised connectivity-based parcellation (CBP) of the thalamus aimed at segmenting the Vc. This suggests that CBP may have utility in targeting electrodes to Vc instead of relying on atlas coordinates. Lastly, DBS of the internal pallidum (GPi) for patients with cervical dystonia was interpretable through stimulation associated connectivity to the putamen, specifically the region of putamen characterised by dense input from M1. This was shown to be uncoupled from electrode coordinates, and therefore points to GPi connectivity maps as a promising tool with which to personalise the procedure. Retrospective data indicates diffusion MRI has promise to improve outcomes in DBS and underscores merit in the future pursuit of prospective work in this field.

Authors: Raghu, A

• Type of award: DPhil
• Level of award: Doctoral
• Awarding institution: University of Oxford
• Language: English