Neuroimaging Biomarkers for Early Diagnosis of Dementia: From Established Methods to AI-Enhanced Precision

BackgroundDementia remains a major clinical and public health challenge, and the window for effective intervention depends heavily on timely and accurate diagnosis. Neuroimaging has become central to dementia research and care, offering non-invasive readouts that can support early detection, differential diagnosis, and disease monitoring. At the same time, translating promising imaging markers into routine clinical use still faces key barriers, including inconsistent links to disease mechanisms and cognitive outcomes, variability across scanners and sites, and limited reproducibility. There is therefore a clear need for imaging approaches that are both biologically grounded and analytically robust, with harmonized methods that can generalize across clinical and research settings.GoalThis Research Topic aims to advance imaging-based diagnostics for dementia with a strong emphasis on early detection and clinically translatable biomarkers. Our goal is to bring together work that connects imaging readouts to disease mechanisms and cognitive outcomes, while demonstrating analytical rigor, reproducibility, and harmonization across scanners and sites. We also seek contributions that support clinical adoption through concise, prospective, multisite validation, and that consider health-economic, regulatory, and equity dimensions that can broaden access, including in resource-constrained settings.ScopeWe invite submissions spanning established and emerging imaging modalities, including structural MRI morphometry, diffusion MRI and microstructural metrics, perfusion imaging, susceptibility- and myelin-sensitive techniques, PET tracers, and hybrid PET/MRI. We welcome quantitative pipelines and responsible uses of artificial intelligence, radiomics, and machine learning where they provide interpretable value beyond standard practice, are benchmarked against clinical standards, and are externally validated.Specific themes of interest include, but are not limited to:Imaging biomarkers for early detection, differential diagnosis, and monitoring in dementiaMechanistically grounded imaging readouts linked to cognitive outcomesReproducible workflows, scanner/site harmonization, and multisite validation approachesInterpretable AI, radiomics, and machine learning methods with external validation and clinical benchmarkingHealth-economic, regulatory, and equity considerations supporting clinical implementation and accessWe welcome contributions in the form of Original Research, Systematic Reviews (including narrative, scoping, meta-analytical, or umbrella review approaches), Brief Research Reports, Protocols, Case Reports, and Perspective articles. Topic editors: Karsten Mueller, Simon Fristed Eskildsen, Lucia Ballerini Source: https://www.frontiersin.org/research-topics/75977/neuroimaging-biomarkers-for-early-diagnosis-of-dementia-from-established-methods-to-ai-enhanced-precision