Mapping the subcortical connectome in Parkinson's disease patients undergoing deep brain stimulation

1509 

Abstract Submission 

Alaa Taha¹, Jason Kai², Mohamad Abbass³, Brendan Santyr¹, Greydon Gilmore³, Bradley Karat¹, Arun Thurairajah⁴, Ali Khan¹, Jonathan Lau⁵

¹University of Western Ontario, London, Ontario, ²Imaging Research Laboratories, Robarts Research Institute, Western University, London, Ontario, ³Department of Clinical Neurological Sciences, Division of Neurosurgery, London, Ontario, ⁴Department of Clinical Neurological Sciences, Division of Neurosurgery, Western University, London, ON, ⁵Department of Clinical Neurological Sciences, Division of Neurosurgery, London, ON

Alaa Taha  
University of Western Ontario
London, Ontario

Jason Kai  
Imaging Research Laboratories, Robarts Research Institute, Western University
London, Ontario

Mohamad Abbass  
Department of Clinical Neurological Sciences, Division of Neurosurgery
London, Ontario

Brendan Santyr  
University of Western Ontario
London, Ontario

Greydon Gilmore  
Department of Clinical Neurological Sciences, Division of Neurosurgery
London, Ontario

Bradley Karat  
University of Western Ontario
London, Ontario

Arun Thurairajah  
Department of Clinical Neurological Sciences, Division of Neurosurgery, Western University
London, ON

Ali Khan  
University of Western Ontario
London, Ontario

Jonathan Lau  
Department of Clinical Neurological Sciences, Division of Neurosurgery
London, ON

Deep brain stimulation (DBS) involves surgical implantation of electrodes to modulate aberrant brain networks [1]. DBS is a focal therapeutic option, while also offering a unique window to understand underlying mechanisms of targeted neural circuits. The most common application of DBS is to manage Parkinson's disease (PD) [1], specifically targeting the subthalamic nucleus (STN), a lens-like basal ganglia (BG) structure spanning ~10 millimeters [2]. The STN serves as an integrative hub that spatially compresses information projected from the cortex, BG, and thalamus [2]. As a result, DBS electrodes placed in the STN can modulate large brain networks, although the exact connections to cortical and subcortical structures remain poorly understood. Most studies employing tractography seek to understand cortical connections (i.e. within the cortex or between the subcortex and cortex) [3]. However, recapitulating the short-range subcortical connections (e.g., within or between the BG and thalamus), important for motor control, cognition, and emotion [4,5] is crucial to understanding PD progression and DBS mechanisms.
In this work, we 1) recapitulate literature validated subcortical connectivity in-vivo via 7-Tesla (7T) MRI, comparing PD and healthy controls (HC) and 2) employ patient-specific subcortical connectivity to predict DBS outcomes.

A total of 60 participants were included in this study, 26 were diagnosed with PD and underwent STN-DBS (age: 60.19 ± 7.08, median = 62.5; 11 female; 12 ± 2 years since diagnosis) and the rest are HCs (age: 45.48 ±13.66, median = 47.5; 15 female). The following sequences were employed: 1) 3D T1w MP2RAGE, 2) 3D optimized T2w fast-spin echo (T2 SPACE), and 3) Whole-brain dMRI, 1.5 mm isotropic, with B-values 1000/2000 and 95 directions. Minimal preprocessing was performed on all MRI modalities. This included gradient non-linearity distortion correction and FreeSurfer. Deep brain region of interest (ROI) segmentations were generated from FreeSurfer combined with the BigBrain subcortical atlas [6] propagated back to subject space, all ROIs were combined via Labelmerge [7]. Mrtrix3 [8] was employed for probabilistic tractography by generating an average group response from HC to estimate the fibre orientation distribution maps. Twenty million streamlines were generated per participant with terminating criteria being a 2 voxel radius from ROI. Apparent fibre density (AFD) was investigated in literature defined [9] motor, limbic, and associative circuits across HC and PD.
Lead-DBS [10] was employed for DBS electrode localization and stimulation volume modeling. We leverage the motor section of the Unified Parkinson's Disease Rating Scale (UPDRS) to capture clinical outcomes 1 month post-DBS device turn on. Patient subcortical connectome was filtered based on tracts traversing DBS electrode stimulated volumes which yielded best clinical improvement. Using leave-one-out cross validation, we predicted UPDRS improvement (%) from filtered patient subcortical connectome.

All individual circuits (i.e., motor, limbic, and associative) were highly correlated (r > 0.8, p < 0.0001) between HC and PD, indicating that the same circuits between the two groups were recapitulated (Figure 1A). We identified significant differences in tract density across 8 subcortical connections between HC and PD (Figure 1B). All DBS electrodes were localized successfully and visualized in a common space (Figure 2). Connectivity strength between subcortical node tracts traversing DBS stimulated volumes predicted UPDRS improvement with an average error of 31%.

We employ a suite of open software for neuroimaging, tractography, DBS analysis, and ML to investigate patient specific subcortical structures and their connections including in the context of DBS electrode implantation.

Deep Brain Stimulation ²

Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s)

Connectivity (eg. functional, effective, structural) ¹

Subcortical Structures

White Matter Anatomy, Fiber Pathways and Connectivity

Basal Ganglia

Computational Neuroscience

HIGH FIELD MR

Machine Learning

Movement Disorder

Neurological

Sub-Cortical

Tractography

WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC

Other - Deep Brain Stimulation (DBS)