Cognitive biomarker prioritization in Alzheimer's Disease using brain morphometric data.

摘要: Cognitive assessments represent the most common clinical routine for diagnosis of Alzheimer’s Disease (AD). Given a large number cognitive assessment tools and time-limited office visits, it is important to determine proper set tests different subjects. Most current studies create guidelines test selection targeted population, but they are not customized each individual subject. In this manuscript, we develop machine learning paradigm enabling personalized prioritization. We adapt newly developed learning-to-rank approach $${\mathtt {PLTR}}$$ implement our paradigm. This method learns latent scoring function that pushes effective onto top prioritization list. also extend better separate less ones. Our empirical study on ADNI data shows proposed outperforms state-of-the-art baselines identifying prioritizing individual-specific biomarkers. conduct experiments in cross validation level-out settings. two settings, significantly best with improvement as much 22.1% 19.7%, respectively, features. The achieves superior performance biomarkers prioritized have great potentials facilitate diagnosis, disease subtyping, ultimately precision medicine AD.