ADNI Imaging Inventory and Protocol for MRI and Clinical Classification of Alzheimer Disease

Abstract

Combining magnetic resonance imaging (MRI) with clinical information requires reliable participant linkage and evaluation that controls diagnostic overlap and data leakage. This study aims to describe an available Alzheimer’s Disease Neuroimaging Initiative (ADNI) imaging inventory and propose a comparison of MRI–clinical fusion with simpler models. We describe the supplied T1-weighted MRI catalogue and three separate ADNI4 illustrative linkages. The proposed study uses demographic and APOE predictors, with cognitive scores evaluated separately. Participant-level nested cross-validation will assess discrimination, calibration and explanation stability. The catalogue contains 7,263 MRI records from 824 participants: 1,325 screening, 274 baseline and 5,664 follow-up records. Phase labels comprise 7,213 ADNI1 and 50 ADNIGO records. The three ADNI4 examples illustrate differences in clinical timing and missing information. The final eligible cohort, quality-control exclusions and model-performance results are not yet available. The available evidence supports an imaging-inventory description and proposed analytical protocol. Completed participant-level linkage and validation are required before concluding diagnostic performance or clinical usefulness.