Predicting Alzheimer's Years Before Symptoms Through Geometric Brain Trajectories

This article introduces an innovative geometric approach to predict Alzheimer’s disease several years before clinical symptoms appear. Using brain MRI scans, we analyze hippocampo‑cortical trajectories through geometric invariants such as mean curvature (κ̄) and continuity (C). The comparison between normal spirals and broken spirals reveals robust predictive markers, detectable 5 to 7 years before the first signs. Beyond Alzheimer’s, this method opens perspectives for other neurodegenerative diseases and demonstrates the power of compact digital approaches compared to traditional data‑heavy methods.