Connectivity-Based Predictions of Hand Motor Outcome for Patients at the Subacute Stage After Stroke

Abstract

Background: Connectivity-based predictions of hand motor outcome have been proposed to be useful in stroke patients. We intended to assess the prognostic value of different imaging methods on short-term (3 months) and long-term (6 months) motor outcome after stroke.

Methods: We measured resting state functional connectivity (rsFC), diffusion weighted imaging (DWI) and grip strength in 19 stroke patients within the first days (5-9 days) after stroke. Outcome measurements for short-term (3 months) and long-term (6 months) motor function was assessed by the Motricity Index (MI) of the upper limb and the box and block test (BB). Patients were predominantly mildly affected since signed consent was necessary at inclusion. We performed a multiple stepwise regression analysis to compare the predictive value of rsFC, DWI and clinical measurements.

Results: Patients showed relevant improvement in both motor outcome tests. As expected grip strength at inclusion was a predictor for short- and long-term motor outcome as assessed by MI. Diffusion-based tract volume (DTV) of the tracts between ipsilesional primary motor cortex and contralesional anterior cerebellar hemisphere showed a strong trend (p = 0.05) for a predictive power for long-term motor outcome as measured by MI. DTV of the interhemispheric tracts between both primary motor cortices was predictive for both short- and long-term motor outcome in BB. rsFC was not associated with motor outcome.

Conclusions: Grip strength is a good predictor of hand motor outcome concerning strength-related measurements (MI) for mildly affected subacute patients. Therefore additional connectivity measurements seem to be redundant in this group. Using more complex movement recruiting bilateral motor areas as an outcome parameter, DTV and in particular interhemispheric pathways might enhance predictive value of hand motor outcome.

Keywords: diffusion weighted imaging; motor outcome; prediction; recovery; resting state fMRI; stroke.

Figure 1

Left: Inclusion/exclusion flowchart. Right: Lesion maps of each participant with lesions indicated in red.

Figure 2

Visualization of tracts and regions of interest (ROIs) used for diffusion-based tract volume (DTV) and rsFC analysis. (A) Pathways included in the probabilistic DWI-tracking demonstrated with probability tracking of all participants: left: M1^il to M1^cl; middle: M1^il to Pons^il; right: M1^il to Cb^cl. (B) Selected ROIs to calculate rsFC. Abbreviations: M1, primary motor cortex; dPMC, dorsal premotor cortex; SMA, supplementary motor area; Cb, cerebellum; il, ipsilesional; cl, contralesional; af, affected hand; na, non-affected hand.

Figure 3

Left graph: Plots of all motricity index (MI) scores for the affected hand during acute, 3 and 6 months. The bold line indicates the median. The staggered line indicates the maxima. Right graph: Plots of the box and block (BB) test for the affected hand, 3 and 6 months after stroke. The bold line indicates the median.

Figure 4

Linear regression of significant predictive measures with the box and block test (BB) after 3 and 6 months for the M1^il–M1^cl DTV. The tract course is demonstrated on a diffusion weighted image overlaid on a T1-weighted image both measured from a healthy control.