Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Longitudinal white matter changes in language tracts in the first year after stroke
Poster Session D, Thursday, October 1, 4:30 - 6:30 pm, Wangari Maathai
Viktoriia Pozdniakova1, Lily Walljasper2, Jillian L. Entrup2, Sarah M. Schneck2, Steffen Bollmann1, Lena Oestreich1, Michael de Riesthal2, Stephen M. Wilson1; 1The University of Queensland, 2Vanderbilt University Medical Center
Longitudinal changes in white matter (WM) language tracts following stroke-induced aphasia remain poorly understood, with existing studies reporting inconsistent findings [1,2,3]. Given that significant recovery of language function occurs over the first year after stroke [4], we used diffusion-weighted MRI to examine microstructural and volumetric changes in bilateral language-relevant WM tracts at 1 month, 3 months, and 12 months after stroke onset. This study included 55 individuals with post-stroke aphasia (25 female; age range 18–83, M = 57.0, SD = 14.5 years). Diffusion-weighted images were acquired (32 directions, b=1000 s/mm², plus one b=0 volume, 1.7×1.7×2.3mm voxels). Single-shell-optimised multi-tissue constrained spherical deconvolution (SS3T-CSD) was performed using MRtrix3 [5] via a QSIRecon pipeline [6]. Bilateral language-relevant tracts—the arcuate (AF), uncinate (UF), inferior longitudinal (ILF), inferior fronto-occipital (IFO), and middle longitudinal (MdLF) fasciculi—were segmented using TractSeg [7]. Fractional anisotropy (FA), axial diffusivity (AD), radial diffusivity (RD), and tract volume were extracted per tract and timepoint. For each diffusion metric, a linear mixed-effects model was fitted with timepoint, tract, hemisphere, and their interactions as fixed effects, participant as a random intercept, and age, sex, and lesion volume as covariates. AD and RD increased significantly over 12 months (AD: F(2,1283)=33.01, p<.001; RD: F(2,1293)=9.43, p<.001), with the majority of change occurring between 3 and 12 months rather than 1 and 3 months post-stroke (AD: β[1m→3m]=3.032×10-5, β[1m→12m]=1.02×10-4 mm²/s; RD: β[1m→3m]=2.94×10-5, β[1m→12m]=6.99×10-5 mm²/s). Both effects were strongly lateralized, with significant time×hemisphere interactions (AD: F(2,1264)=33.07, p<.001; RD: F(2,1264)=10.13, p<.001); hemisphere-specific follow-up models confirmed significant diffusivity increases in the left hemisphere (AD and RD both p<.001) but no change in the right (AD: p=.30; RD: p=.98), consistent with Wallerian degeneration driven by the ipsilateral stroke lesion. Within the left hemisphere, the trajectory of AD increase varied significantly across tracts (time×tract: p<.001), with ILF and UF showing the largest increases, while RD increases were more uniform across tracts. FA showed no significant change over time (p=.43) and no significant interactions, consistent with parallel axonal and myelin degeneration, as proportional increases in AD and RD leave FA unchanged. Tract volume did not change significantly over 12 months (p=.33), suggesting that despite progressive microstructural deterioration, bulk tissue loss had not yet occurred within this timeframe. These findings suggest progressive Wallerian degeneration in ipsilesional language tracts over the first year post-stroke, with microstructural changes preceding detectable volume loss. 1. Schlaug et al., Ann N Y Acad Sci 2009;1169:385–394 2. van Hees et al., Neurorehabil Neural Repair 2014;28:325–334 3. Soliman et al., Magn Reson Imaging 2023;95:19–26 4. Wilson et al., Brain 2023;146:1021–1039 5. Tournier et al., NeuroImage 2019;202:116137 6. Cieslak et al., Nat Methods 2021;18:775–778 7. Wasserthal et al., NeuroImage 2018;183:239–253
Topic Areas: Disorders: Acquired,