Eye movement analysis during working memory tasks correlates with cognitive, clinical, and anatomical metrics in multiple sclerosis

Background Cognitive dysfunction involving working memory is common in multiple sclerosis (MS) but is often under-detected in clinical assessments. Noninvasive digital biomarkers like oculomotor metrics may enhance the detection of neurofunctional changes. Objective Determine whether oculomotor metrics during a visuospatial N- back task reflect cognitive dysfunction and correspond with clinical measures in patients with MS. Methods In a cross-sectional study, 38 patients completed an N- back working memory task with eye tracking to record gaze duration and task accuracy. Clinical measures included Symbol Digit Modalities Test (SDMT), retinal nerve fiber layer (RNFL), macular volume, Expanded Disability Status Scale (EDSS), and Nine-Hole Peg Test (9HPT). Linear regression models examined relationships between task accuracy, gaze duration, and clinical measures. Results Gaze duration and task accuracy were inversely correlated (r = −0.57, p < 0.001). Shorter gaze duration and higher task accuracy, indicators of better cognitive processing efficiency, were associated with better SDMT scores, greater RNFL thickness and macular volume, and better EDSS and 9HPT completion times. Conclusions Oculomotor metrics recorded during a working memory task are associated with cognitive, retinal, and functional measures in pwMS, supporting their potential as noninvasive biomarkers to complement standard clinical assessments in detecting cognitive dysfunction.