Deep Learning for Dementia Disease Prediction Using the ADNI Dataset: A Comprehensive Review

Alzheimer's disease (AD) and dementia in general constitute one of the key neurological problems in the current twenty-first century and currently affect more than 55 million people worldwide. Accurate and timely diagnostic remains one of the biggest challenges, requiring automation and application of computational techniques. Over the last ten years, deep learning models emerged as the most important tool to diagnose dementia, using the rich repository of biomarkers obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI). The current literature review gives an overview of various deep learning methods employed for dementia prediction from the ADNI dataset in 2011-2024. Machine learning methods, CNNs, RNNs, LSTMs, transformers, GNNs, and GANs are mentioned. Various features such as input form, classification problem, architecture, and related performance metrics are reviewed. Such issues as data scarcity, heterogeneity of sites, class imbalance, model interpretability, and challenges in clinical implementation are also important. Emerging fields like federated learning, XAI, foundation models, and multimodal data integration are outlined. The review is intended as a systematic guide for navigating the fast-developing field of artificial intelligence-assisted dementia detection