Researchers have developed a self-supervised framework called Delta2Gamma for detecting Alzheimer's disease using electroencephalography (EEG) recordings. Unlike traditional imaging-based diagnosis, EEG is portable and inexpensive but has noisy and varying signals with limited clinical labels. Delta2Gamma decomposes each recording into five neural rhythms (delta, theta, alpha, beta, gamma) and learns their representations from unlabeled data by contrasting augmented views of each signal. The framework achieves 92.4% accuracy in separating Alzheimer's disease from cognitively normal controls on a specific dataset. --- Why it matters: This matters to researchers in AI because it provides a low-cost and scalable solution for dementia screening, which is crucial for early detection and treatment. Delta2Gamma's self-supervised approach also demonstrates the potential of contrastive learning for medical imaging analysis. Source: https://arxiv.org/abs/2608.17231