AI and the World's Hard Problems · Medicine & drug discovery · Nature (computational biology) · 2026-10-20
EEG-based Alzheimer’s disease detection using two-branch multi-scale deep learning with integrated gradients-guided channel selection
Early diagnosis of Alzheimer's disease is critical, and EEG offers a non-invasive, low-cost alternative to neuroimaging. The proposed Fused-MSCA-2B, a two-branch multi-scale channel-attentive 1D CNN, processes broadband multichannel EEG directly, while Integrated Gradients ranks channel relevance to reduce the montage from 19 to 12 channels. Evaluated on two benchmark datasets, this IG-guided 12-channel configuration achieved 98.60% ± 1.18% recording-level accuracy.