Quantitative multimodal microstructural imaging associations with chronic stroke motor impairment

BackgroundCorticospinal tract (CST) integrity, assessed using diffusion weighted MR imaging, has frequently been associated with motor impairment following stroke. However, which specific biological features contribute to this relationship remains unknown. In the current paper we used quantitative multimodal imaging to shed light on the biophysical properties underlying this relationship in the CST and explore additional whole brain relationships with motor impairment. MethodsTwenty-seven chronic stroke survivors underwent MRI scanning, including multi-shell diffusion weighted imaging and quantitative Multi-Parameter Mapping, from which we derived ten metrics, sensitive to different microstructure properties. We tested the relationship between these MRI metrics and stroke survivors impairment scores on the Upper-Extremity Fugl Meyer (UE-FM), using multimodal and unimodal statistical inference, in both the CST and across the whole brain. ResultsUsing an ROI approach to replicate and extend on previous findings, a joint multimodal analysis revealed that greater CST microstructure asymmetry was associated with worse motor impairment. The strongest relationships were found between impairment and several modalities sensitive to free water and myelination. This result was further confirmed using a voxel-wise approach. Beyond the CST, modalities sensitive to free water, myelination and the Neurite Density Index, were also found to be related to motor impairment, particularly in the superior longitudinal fasciculus, corpus callosum and thalamic radiations, with magnetisation transfer saturation showing the strongest relationship in these areas. ConclusionModalities that are sensitive to myelin and free water exhibited the most significant correlation with motor impairment. This suggests that these specific biological features may account for the previously identified relationships between diffusion metrics and motor impairments. Myelin metrics in particular were also highly correlated with motor impairment in other brain regions. This study illustrates the advantages of using multimodal data to identify more specific biological factors that contribute to motor impairment in stroke patients.