In a recent episode of Evidence Based Retina from Eyes On Eyecare, Duke’s Sharon Fekrat, MD, FACS, FASRS joined Rishi P. Singh, MD, FASRS to discuss how multimodal retinal imaging and artificial intelligence are reshaping the early detection of neurocognitive disease. Dr. Fekrat leads the iMIND Study Group, which has enrolled more than 2,000 participants across a wide range of neurologic conditions. Using OCT, OCT angiography, and ultra‑widefield imaging, the team has identified consistent patterns such as ganglion cell inner plexiform layer thinning and reduced retinal microvascular density that may signal neurodegenerative change long before symptoms appear. Early AI models trained on these multimodal images have already shown promise in distinguishing symptomatic Alzheimer’s disease and mild cognitive impairment from normal cognition.
While these findings highlight the potential of retinal imaging as a noninvasive biomarker for brain health, Dr. Fekrat emphasized that broader patient diversity and standardized imaging protocols are essential before this approach can be used in clinical care. As the team continues refining AI models and exploring whole‑fundus imaging, future applications could include population‑level screening, primary care triage, and earlier identification of patients for clinical trials. This work positions Duke at the forefront of efforts to use the eye as a window into neurocognitive disease.