Alzheimer’s Disease Detection Using Machine and Deep Learning: Empirical Evidence, Predictive Models, and Research Gaps, A Systematic Review

Abstract

Alzheimer’s disease is an irreversible neurodegenerative disorder that affects a growing proportion of the global population and represents a significant challenge for healthcare systems due to its gradual progression and the absence of a definitive cure, a context in which early detection through Machine Learning emerges as a promising strategy to support timely clinical interventions. This research aims to determine how Machine Learning models impact the detection of Alzheimer’s disease through the systematic analysis of the available empirical evidence. A systematic review was conducted under the PRISMA 2020 framework, considering studies published between 2019 and 2025 in IEEE Xplore, Web of Science, Scopus, EBSCOhost, and ScienceDirect; after applying inclusion criteria and quality assessment, 73 papers were selected. Support Vector Machine (18.91%), Random Forest (16.92%), and Decision Tree (13.93%) predominate as hegemonic models, while Python accounts for 45% of the developments; the functional category groups 70.3% of the theoretical definitions, and Neuroimaging Classification emerges as the only specialized thematic cluster. The field shows a progressive transition toward multimodal deep architectures, although critical gaps persist in clinical validation, interpretability, and ontological standardization, which guide future research.