Fine motor task function of the non-paretic hand can be predicted from a constrained motor connectome.

Research Report
Population: Adult

Lara A Boyd, PT, PhD, Professor, University of British Columbia lara.boyd@ubc.ca

Sue Peters, PT, PhD Candidate, University of British Columbia s.peters@alumni.ubc.ca

Katie Wadden, MSc, PhD Candidate, University of British Columbia k.wadden@alumni.ubc.ca

Jason Neva, PhD, Postdoctoral Fellow, University of British Columbia jason.neva@ubc.ca

Keywords: Prediction, Multimodal Neural Imaging, Connectome, Motor Function

Purpose/hypothesis: The nonparetic hand may compensate for reduced motor function of the paretic arm after stroke. Animal models of stroke have shown that white matter damage distant from the lesion may impact function. Thus, structural reorganization of both hemispheres post-stroke may contribute to the resulting altered function of both paretic and nonparetic hands. Multimodal neuroimaging, combining structural and functional measures, may provide new insight into the underlying motor network in the brain. This study’s purpose was to examine whether motor function of the nonparetic hand could be predicted from a constrained motor connectome (CMC) derived from a multimodal neuroimaging approach.

Subjects: 26 (age 65.6±10.7SD years, post-stroke duration (PSD) 68.1±66.5SD months)

Methods: Participants underwent diffusion imaging (DTI) at a 3T magnetic resonance imaging (MRI) research centre. A bilateral sensorimotor network mask, created from a functional MRI connectivity analysis demonstrating a connected motor network in healthy participants during motor performance, was used to extract white matter information from a CMC. The clusters from this analysis represent an idealized motor network and were used as seed regions to construct a CMC for each participant. To quantify motor performance for this study, the Wolf Motor Function (WMF) and Fugl-Meyer (FM) upper extremity tests were assessed. For the WMF, movement time to complete 15 items with paretic and nonparetic arms was determined. A task rate was calculated as 60seconds/performance time (s) with a score of ‘0’ given if a participant was unable to perform the task. The WMF was also divided into fine (WMF-f) and gross (WMF-g) motor tasks with separate task rates calculated. Pearson’s correlations were computed for the WMF, WMF-f, and WMF-g, with diffusion measures (apparent diffusion coefficient (ADC), fractional anisotrophy (FA)). A regression followed for the nonparetic WMF-f with age, PSD, and ADC as predictors.

Results: On average, participants were moderately impaired (FM 48.0±16.8). The nonparetic WMF-f correlated with ADC (r=-0.41, p=0.04), and not with FA (r=0.27, p=0.18). Multiple hierarchical regression analysis revealed that lower ADC within the CMC predicted faster fine motor task rates of the nonparetic hand (r2=0.36, p=0.05) and accounted for an additional 12% of the variance after age/PSD.

Conclusions: A CMC can predict nonparetic fine motor hand performance after stroke. The diffusion metric ADC, characterized motor performance better than FA.

Clinical relevance: Applying this connectome to the post-stroke brain, may provide a biomarker for sensorimotor function, and lends support to the theory that both hemispheres are important for reorganization after stroke. Combining fMRI and DTI to create a CMC may provide a novel neurophysiological method to define the microstructural properties of a network of regions to predict motor performance after stroke. Interestingly, results show that dysfunction in the underlying CMC after stroke is a key variable in predicting motor performance of the nonparetic hand.

Citation:
Boyd, Lara A, PT, PhD; Peters, Sue , PT; Wadden, Katie P, MSc; Neva, Jason L, PhD. Fine motor task function of the non-paretic hand can be predicted from a constrained motor connectome.. Poster Presentation. IV STEP Conference, American Physical Therapy Association, Columbus, OH, July 17, 2016. Online. https://u.osu.edu/ivstep/poster/abstracts/078_boyd-et-al/

Return