Window to the Mind: A Multi-Stage Computational Pipeline for Early Alzheimer's Intervention

Overview

Alzheimer's disease steals the memories of 55+ million worldwide, yet no effective diagnosis or treatment exists. The retina is composed of neural tissue, thus mirroring neurological damage and offering a non-invasive window into neural health. I trained a deep-learning model on 638 retinal images to detect Alzheimer's, then used gradient imaging to identify anatomical regions driving its decisions. The model achieved 91.8%±0.279% accuracy and independently identified the optic disc as its primary signal, confirming it learned real pathology. Alongside this, I discovered RetinAD-1, a novel prodrug candidate that overcomes the barriers preventing promising inhibitors from clinical use. RetinAD-1 achieved a binding score of −8.76 kcal/mol, passed all drug-likeness criteria, and demonstrated stable binding geometry through molecular dynamics. Together, this pipeline could detect and treat Alzheimer's years before symptoms appear, giving families the chance to act before it's too late.

Awards (2)

• Gold Medal
• Selected for CWSF 2026

Competition history

• CWSF 2026 Disease & Illness

Related projects

Closest projects by meaning, across every fair and year in the corpus.