Data-driven causal model discovery and personalized prediction in Alzheimer’s disease


Professor Guang Lin and his collaborators developed a data-driven causal model to describe the biomarker dynamics for all eligible subjects in ADNI-1 database and make personalized predictions for patients who provide enough longitudinal biomarker data points. This work hopes to benefit Alzheimer’s disease early diagnosis and personalized predictions. This work was accepted to be published in Nature Digital Medicine.

The following figure presents a flowchart of the proposed data-driven modeling approach. Given the initialized ODE model, a causal model is obtained by fitting the ADNI dataset and DPS model through sparse learning; secondly, the ADNI dataset is used to calibrate the population parameters in the causal model and obtain the population model; thirdly, a sensitivity analysis is applied to analyze the sensitivity of each population parameters and determine the sensitive personalized parameters, and a simulate study is conducted to validate the population model. Then, the personalized model is obtained by calibrating the sensitive personalized parameters with the use of personalized data. A prediction is made by the personalized model in the end.

Department of Mathematics, Purdue University, 150 N. University Street, West Lafayette, IN 47907-2067

Phone: (765) 494-1901 - FAX: (765) 494-0548  Contact Us

© 2022 Purdue University | An equal access/equal opportunity university | Copyright Complaints

Trouble with this page? Disability-related accessibility issue? Please contact the College of Science.

Maintained by Science IT