AI Turns Simple EEG Scans Into Accurate Dementia Detectors
Neuroscience News • • health
Key Points:
- Researchers at Florida Atlantic University developed a deep learning model using EEG data to distinguish Alzheimer's disease (AD) from frontotemporal dementia (FTD) with 84% accuracy by analyzing both temporal and spectral brain activity patterns.
- The model identified slow delta waves in frontal and central brain regions as key biomarkers for both diseases, with AD showing widespread brain disruption and FTD exhibiting more localized changes in frontal and temporal lobes.
- This two-stage system first detects dementia presence, then differentiates AD from FTD, and also estimates disease severity with relative errors below 35% for AD and 15.5% for FTD, offering clinicians faster and more affordable diagnostic insights than MRI or PET scans.
- The approach enhances interpretability through Grad-C
