Microstructural changes predict higher-order visual deficits after posterior cerebral artery stroke

Poster Presentation 33.411: Sunday, May 17, 2026, 8:30 am – 12:30 pm, Pavilion
Session: Functional Organization of Visual Pathways: Subcortical, clinical

Sara Ajina1 (), Yuheng Zeng1, Alexander Leff2, Matthew Lambon Ralph3, Randi Starrfelt4; 1Wellcome Centre for Human Neuroimaging, UCL Queen Square Institute of Neurology, University College London, 2Institute of Cognitive Neuroscience, UCL Queen Square Institute of Neurology, University College London, 3MRC Cognition and Brain Sciences Unit, University of Cambridge, 4Department of Psychology, University of Copenhagen

Posterior cerebral artery (PCA) stroke can cause diverse higher-order visual deficits affecting word, object, and face perception, yet the supporting white-matter changes remain poorly understood. Here, we examined microstructural changes across visual white-matter pathways and their relationship to domain-specific perceptual deficits in chronic PCA stroke. Forty-four chronic PCA stroke survivors and 21 healthy controls completed a comprehensive visual battery including face, object and word processing (“Back of the Brain” project). Domain-specific composite scores were derived using principal component analysis. All participants underwent 3T structural and diffusion MRI. Lesions were corrected using virtual brain grafting prior to multimodal-parcellation of 55 visual regions, 11 subregions. Whole-brain tractography between visual subregions used constrained spherical deconvolution. Tracts were weighted with SIFT2, normalized by intracranial volume. Ten major tracts were defined from control data. Mean Fractional Anisotropy was extracted and related to perceptual scores for left- and right-sided stroke using correlation and regression analyses; age and lesion volume as covariates. Word processing showed a modest overall association with left-hemisphere tract integrity (F=8.1,FDR_p=0.05), alongside a strong ‘stroke-hemisphere’ x ‘tract-hemisphere’ interaction: in left-sided stroke, word scores related primarily to left-hemisphere pathway integrity, whereas in right-sided stroke, performance related more strongly to right-hemisphere pathways (F=13.0,FDR_p=0.006). Face processing showed clear right-hemisphere dominance, associated with integrity in right ventral and dorsal stream tracts (F=23.6, FDR_p=0.0001). Object processing was selectively associated with integrity of right parahippocampal–ventral stream tract (F=10.4,FDR_p=0.02). Regression analyses showed that in right-sided stroke, integrity of right and left lateral temporal–STS pathways predicted face (R²=0.9, FDR_p=0.003) and word (R²=0.8, FDR_p=0.007) performance respectively, independent of lesion volume and age. Word deficits depend largely on tract integrity in the damaged hemisphere with a slight left-hemisphere bias, whereas face and object deficits show strong right-hemisphere dominance. Associations with contralesional tract integrity suggest interhemispheric plasticity and/or transneuronal degeneration may also influence visual outcomes.

Acknowledgements: This project was funded by the Independent Research Fund Denmark (Sapere Aude to R.S.; DFF - 4180-00201) and supported by a programme grant and intramural funding to M.A.L.R. from the Medical Research Council [MR/R023883/1; MC_UU_00005/18], and a Clinical Research Career Development Fellowship to S.A. from Wellcome [224655/Z/21/Z].