Neurovascular Dysfunction in Retinal Diseases: Biological Mechanisms and Emerging AI-based diagnostic approaches

In the world, retinal diseases are one of the most important causes of permanent vision loss, and they are now known to be more than just vascular diseases; these are diseases of complex interaction between the neuronal, vascular, glial, and immune elements. The retinal neurovascular unit, which consists of retinal neurons, endothelial cells, pericytes, Müller glia, astrocytes and microglia, is a unit that is pivotal to maintaining the homeostasis of the retina via coordinated neurovascular coupling and blood-retinal barrier integrity. This highly integrated system is disrupted in the pathogenesis of major retinal diseases such as diabetic retinopathy , age-related macular degeneration, glaucoma and retinal vein occlusion. Recent data indicates that neurovascular dysfunction usually occurs before clinically apparent vascular lesions, leading to a paradigm change from a vascular to a model of early neuronal damage, inflammation, oxidative stress, and impaired cellular communication. Research for decades pointed to the retinal neurovascular unit as a target for early therapeutic interventions in diabetic retinopathy. The development of new imaging techniques for the retina, such as optical coherence tomography angiography (OCTA), adaptive optics and hyperspectral imaging, have made it possible to visualize micro-vascular and micro-neural changes in the retina like never before. At the same time, artificial intelligence (AI), especially deep learning, multimodal learning, and explainable AI (XAI), has revolutionized retinal diagnostics with automated detection of retinal diseases, their severity, and prediction of the disease prognosis and management. The recent advances in foundation models and multimodal AI systems that combine fundus photography, OCT, OCTA, and clinical metadata, signify a significant leap in precision ophthalmology. This review explores the biological mechanisms involved in retinal neurovascular dysfunction, reviews its contribution to the major retinal diseases, and critically analyzes recent advances, current limitations, and future prospects for the computational intelligence in clinical retinal care.