PhD thesis defense. Advances for Alzheimer’s disease diagnostics through the integration of spectroscopy and chemometrics

This thesis investigates vibrational spectroscopy in combination with chemometrics as a liquid biopsy for Alzheimer’s disease (AD) diagnostics before symptoms. A cross-validated model from cerebrospinal fluid Raman spectra achieved >90% accuracy across 2 different cohorts, while a more tutorial-like study showed how improper validation could inflate model performance and proposed strategies against overfitting. Hyperspectral imaging further revealed AD-related signals concentrated in inner dried biofluid regions. Together, the work highlights Raman spectroscopy’s diagnostic potential, with rigorous validation and spatially informed measurement practices.