ABSTRACT NUMBER: ESOC2026A1101 DISSOCIATION BETWEEN PROCESSING SPEED AND NEUROIMAGING BURDEN ACROSS CEREBRAL SMALL VESSEL DISEASE ETIOLOGIES

Abstract Background and aims Cerebral small vessel disease (SVD) contributes to stroke and cognitive decline and is identified by neuroimaging markers. However, clinical and radiological presentation varies widely between SVD etiologies, complicating the associations between neuroimaging markers and cognitive function. Processing speed is often impaired in individuals with SVD, but its relationship with neuroimaging markers of SVD remains poorly understood. We investigated these relationships in a multi-etiological SVD cohort. Methods We included 90 participants recruited at Assistance Publique – Hôpitaux de Paris (AP-HP): 30 with sporadic SVD, 30 with cerebral autosomal dominant arteriopathy with subcortical infarcts and leukoencephalopathy (CADASIL) and 30 with cerebral amyloid angiopathy (CAA). All participants underwent multimodal neuroimaging at 3T and 7T and neuropsychological assessment. All analyzes were controlled for age, gender and education. Results Participants with sporadic SVD had a mean (SD) age of 62.4 (9.9) years, 57.7 (13.7) years in the CADASIL cohort, and 73.8 (8.4) years in the CAA cohort. Most neuroimaging markers showed significant differences across SVD etiologies (Figure 1). In contrast, processing speed performance did not differ between SVD etiologies. Only within the CAA group, higher PSMD (p-FDR = 0.04), higher free water fraction (p-FDR = 0.06), and lower diffusion tensor image analysis along the PVS (DTI-ALPS) index (p-FDR = 0.06) were associated with slower processing speed. Conclusions In a multi-etiological SVD cohort with comparable processing speed, our preliminary results suggest that neuroimaging-processing speed relationships are etiology-specific, highlighting that cognitive impairment in SVD arises from distinct pathologies rather than universal mechanisms. Exploring additional SVD markers may clarify processing speed-neuroimaging link. Conflict of interest Nothing to disclose Figure 1 - belongs to Results