Bayesian stroke modeling details sex biases in the white matter substrates of aphasia

Abstract

Ischemic cerebrovascular events often lead to aphasia. Previous work provided hints that such strokes may affect women and men in distinct ways. Women tend to suffer strokes with more disabling language impairment, even if the lesion size is comparable to men. In 1,401 patients, we isolated data-led representations of anatomical lesion patterns and hand-tailored a Bayesian analytical solution to carefully model the degree of sex divergence in predicting language outcomes ~3 months after stroke. We located lesion-outcome effects in the left-dominant language network that highlight the ventral pathway as a core lesion focus across different tests of language performance. We provide newly detailed evidence for sex-specific brain-behavior associations in the domain-general networks associated with cortico-subcortical pathways, with unique contributions of the fornix in women and cingular fiber bundles in men. Our collective findings suggest diverging white matter substrates in how stroke causes language deficits in women and men. Clinically acknowledging such sex disparities has the potential to improve personalized treatment for stroke patients worldwide.

Bayesian modeling disentangles language versus executive control disruption in stroke

“…Next, we devised and deployed a principled Bayesian framework. A fully probabilistic framework is well suited to trace out how brain tissue impairments impact interindividual differences in key cognitive domains after acute ischemic stroke 24,25 and carefully quantify the extent of overlap of these brain-behavior correspondences. The combination of a comprehensive cognitive test battery with factor analysis in a Bayesian modeling framework should capture the impact of stroke-induced dysfunction in distributed brain regions for key faculties of human cognition.…”

“…Finally, all models took into account several covariates, including age, age 2 , sex, time since stroke onset, education years, pre-morbid cognitive performance and total lesion volume following previous research. 24,25 Details on the Bayesian model specification and posterior Bayesian estimates for all parameters and models are available in the SI Materials and Methods section.…”

“…Details about structural brain image assessment and resulting segmented lesion patterns were described elsewhere. 24,25 Neuropsychological and demographic data were obtained from the Korean-Vascular Cognitive Impairment Harmonization Standards-Neuropsychology Protocol. 31,46 Our analyses focused on established assessments of global cognitive function (MMSE) 32 , language (Korean short version of the BNT) 33 , executive speech function (Korean version of phonemic and semantic fluency tests) 31 , executive control functions (TMT, version A and B and DSCT) 34,35 , visuospatial functioning (RCF) 36 , as well as verbal learning and memory function (SVL) 32 , thus indexing a spectrum of language production and executive control functions in each patient.…”

“…To this end, we applied a previously established set of 10 distinct topographical lesion configurations, so-called "lesion atoms" (lesion patterns), in a data-driven fashion based on non-negative matrix factorization (NMF). 24,25,47 Note that the definition of lesion patterns was symmetrically identical in both hemispheres.…”

“…Moreover, only a few studies to date take a network perspective and consider the role of both hemispheres across various cognitive domains in patients with stroke. e.g., 15,24,25 While the few existing studies lend initial evidence for substantial lateralization differences between language and executive control, they do not address the functional interplay of domains.…”