Multimodal Machine Learning for Alzheimer’s Disease Classification using ADNI Biomarker Fusion

Despite ongoing advances, accurate diagnosis of Alzheimer’s disease (AD) remains challenging due to its multifactorial nature, comorbidities, and clinical heterogeneity. Accordingly, approaches that combine multimodal data may improve AD classification. To investigate this, we evaluated classification performance using a preprocessed ADNI-3 dataset comprising a shared set of clinical/cognitive features along with four imaging modality-based cohorts: trimodal (MRI + amyloid PET + tau PET), MRI + amyloid PET, MRI + tau PET, and MRI-only. We trained a range of supervised machine learning (ML) and deep learning (DL) classifiers using stratified five-fold cross-validation and evaluated performance using accuracy, balanced accuracy, and macro-F1. Among all ML model-cohort combinations, the highest-performing configurations were gradient-boosted tree models, with CatBoost on MRI+tau PET (Macro-F1: 0.932 ± 0.043) and LightGBM on MRI+amyloid PET (Macro-F1: 0.924 ± 0.039). Among all DL configurations, the highest performance was achieved by AutoencoderFusion on MRI+amyloid PET (Macro-F1: 0.899 ± 0.042) and on MRI-only (Macro-F1: 0.896 ± 0.032), respectively. Our results indicate strong performance when combining clinical/cognitive and imaging data for AD classification in ADNI-3. While clinical/cognitive data provided the strongest classification signal, adding PET yielded modest improvements among ML configurations.