Symptom-specific brain networks identified in rsfMRI data of Parkinson's disease patients

160 

Abstract Submission 

Jurgen Germann¹, Gavin Elias¹, Andrew Yang¹, Alexandre Boutet¹, Andres Lozano¹

¹University Health Network, University of Toronto, Toronto, Ontario

Jurgen Germann  
University Health Network, University of Toronto
Toronto, Ontario

Gavin Elias  
University Health Network, University of Toronto
Toronto, Ontario

Andrew Yang  
University Health Network, University of Toronto
Toronto, Ontario

Alexandre Boutet  
University Health Network, University of Toronto
Toronto, Ontario

Andres Lozano  
University Health Network, University of Toronto
Toronto, Ontario

Parkinson's disease (PD) is a neurodegenerative disorder; patients with PD exhibit motor symptoms such as tremor rigidity, bradykinesia, and axial impairment. (Poewe et al., 2017) Treatments for PD include pharmacological and surgical therapies such as deep brain stimulation (DBS). DBS is an invasive neuromodulation technique that can deliver immense therapeutic benefits in PD through the modulation of local and distal brain networks. (Lozano & Lipsman, 2013) Treatment response in DBS depends on the precise location of stimulation, and previous work has demonstrated that symptom-specific clinical improvement is associated with the stimulation of different diencephalic brain regions. (Boutet et al., 2021) The goal of this work was to identify symptom-specific brain networks using patient-specific pre-operative rsfMRI data. The presence of such networks would open an avenue for therapies to be tailored to individual patients. Furthermore, the work may identify symptom-specific target regions that would allow for the use of non-invasive neuromodulation techniques such as transcranial magnetic stimulation (TMS) or focused ultrasound (FUS).

Following ethics approval (University Health Network Research Ethics Board #15–9777), rsfMRI scans were prospectively acquired in 133 PD patients (47 female; average age: 62.3 years (Stdev 10.6); average disease duration: 10.5 years (Stdev 5.1)) as part of their pre-operative planning MRI session (field strength: 1.5-3T, TR: 1880-2200 ms, TE: 30-35 ms, flip angle: 50-85°, slice thickness: 2.5-4.5 mm, matrix: 64-88×64-88 voxels). Pre-operative baseline (Med-OFF) motor item scores of the Unified Parkinson's Disease Rating Scale (UPDRS-III) were also collected for each patient, and symptom-specific scores calculated. The rsfMRI data were processed using the BRANT toolbox (http://brant.brainnetome.org/). (Xu et al., 2018) Following preprocessing (removal of first 10 volumes; motion correction; normalization; resampling; denoising for nuisance variables; filtering with a temporal bandpass filter [0.01–0.08 Hz]; smoothing with a Gaussian kernel [6mm fwhm]), the functional connectivity of the motor network of each individual was assessed using the motor region of the thalamus as the seed region. The Pearson correlations between the time courses of the ROI and all other voxels in the brain were calculated and Fisher z-transformed. Additional maps were calculated using the supplementary motor area and the motor/premotor cortex as seeds. Using R (version 4.0.2) and RMINC, symptom-specific patterns of connectivity were calculated using a linear regression between individual rsfMRI connectivity and symptom scores.

Independent of seed region, each symptom was associated with distinct functional connectivity across various motor regions (Figure 1). Each symptom network involves distinct regions of the motor/premotor region (Figure 1). These are in close proximity to the 'hand knob,' a standard reference target for TMS, and could readily be targeted using non-invasive neuromodulation such as TMS or FUS toameliorate specific symptoms. Furthermore, each symptom-specific connectivity network showed unique connectivity in the diencephalic region that is targeted in DBS (Figure 2). The pattern of peak locations is highly similar to the optimal symptom-specific DBS stimulation locations derived from sweet-spot mapping previously reported. (Boutet et al., 2021)

This work demonstrates that individual preoperative rsfMRI shows distinct patterns associated with individual symptom severity. The symptom-specific brain networks identified may be used for individual treatment planning and provide potential brain targets for non-invasive neuromodulation techniques such as FUS or TMS. The availability of non-invasive techniques could allow more PD patients to benefit from neuromodulation therapies to alleviate motor symptoms.

Deep Brain Stimulation ²

Non-invasive Magnetic/TMS

Sonic/Ultrasound

Neurodegenerative/ Late Life (eg. Parkinson’s, Alzheimer’s) ¹

fMRI Connectivity and Network Modeling

Basal Ganglia

FUNCTIONAL MRI

Movement Disorder

Transcranial Magnetic Stimulation (TMS)

ULTRASOUND