Enhanced language models for predicting Alzheimer's disease pathology
ADLLM can predict composite outcomes by processing multimodal inputs, capturing longitudinal patterns, and providing interpretable attention-based variable importance. The model consistently performed well across cohorts and outcome tasks, emphasizing its potential as a scalable and clinically relevant tool for AD diagnosis and progression monitoring.
Authors
- Kevin Shen (ORCID: https://orcid.org/0000-0002-1773-038X)
- Jingshen Wang (ORCID: https://orcid.org/0000-0002-1432-3834)
- Yanping Li (ORCID: https://orcid.org/0000-0002-0412-2748)
- Joseph Giorgio
- Susan M. Landau
- William J. Jagust
- Marcell Borhi
- Qiuran Rita Lyu
Institutions
- Lawrence Berkeley National Laboratory (US)
- University of California, San Francisco (US)
- Nankai University (CN)
- Indiana University Bloomington (US)
- University of California, Berkeley (US)
Publication Details
- Journal
- Alzheimer s & Dementia
- Published
- 2025-12-01
- DOI
- https://doi.org/10.1002/alz70860_105841
- Primary Topic
- Machine Learning in Healthcare
- Type
- article
- Field-Weighted Citation Impact
- 0.00