Structural brain disconnectivity mapping of post-stroke fatigue

Abstract

Stroke patients commonly suffer from post stroke fatigue (PSF). Despite a general consensus that brain perturbations constitute a precipitating event in the multifactorial etiology of PSF, the specific predictive value of conventional lesion characteristics such as size and localization remains unclear. The current study represents a novel approach to assess the neural correlates of PSF in chronic stroke patients. While previous research has focused primarily on lesion location or size, with mixed or inconclusive results, we targeted the extended structural network implicated by the lesion, and evaluated the added explanatory value of a structural disconnectivity approach with regards to the brain correlates of PSF. To this end, we estimated individual structural brain disconnectome maps in 84 S survivors in the chronic phase (≥3 months post stroke) using information about lesion location and normative white matter pathways obtained from 170 healthy individuals. PSF was measured by the Fatigue Severity Scale (FSS). Voxel wise analyses using non-parametric permutation-based inference were conducted on disconnectome maps to estimate regional effects of disconnectivity. Associations between PSF and global disconnectivity and clinical lesion characteristics were tested by linear models, and we estimated Bayes factor to quantify the evidence for the null and alternative hypotheses, respectively. The results revealed no significant associations between PSF and disconnectome measures or lesion characteristics, with moderate evidence in favor of the null hypothesis. These results suggest that symptoms of post-stroke fatigue among chronic stroke patients are not simply explained by lesion characteristics or the extent and distribution of structural brain disconnectome, and are discussed in light of methodological considerations.

Keywords: Brain mapping; Lesion; MRI; Post-stroke fatigue; Stroke; Structural disconnectome.

Fig. 1

Heatmap displaying lesion overlap across stroke patients by 70 slices (2 mm thickness) from z(voxel) = 1 to z = 70. Maximal overlap was 8, but for illustration purposes, the color scale saturates at 5.

Fig. 2

Distributions and group differences in FSS and PHQ for healthy controls (HC) and stroke patients. Red line denotes cut off value for clinically significant symptom load. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 3

Individual lesions (blue) and associated disconnectome maps (yellow–red). Probability for disconnection ranges from 10 (yellow) to 100 (red). Patient A: right cerebral white matter lesion, Patient B: brain stem lesion, Patient C: left and right cerebral cortex and white matter lesions. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)