Artificial intelligence and machine learning-based models in evaluating Alzheimer’s disease and its association with diabetes mellitus

discrimination (AUC = 0.88).Among ML classifiers, random forest achieved the highest accuracy (88%) and AUC (0.90), outperforming traditional regression.MRI-derived ONSD and ventricular indices are significantly associated with papilledema and enable accurate non-invasive prediction of raised ICP.Machine learning-based classification further enhances diagnostic performance, supporting integration of quantitative MRI biomarkers with AI for neuro-ophthalmological assessment.However, prospective multicentre validation of the study is needed.