Related Experiment Video
Updated: Aug 29, 2025

Author Spotlight: Advancing Alzheimer's Research – Exploring Early Detection and Multi-Omics Approaches
Published on: December 15, 2023
Data-driven causal model discovery and personalized prediction in Alzheimer's disease
Haoyang Zheng1, Jeffrey R Petrella2, P Murali Doraiswamy3
1School of Mechanical Engineering, Purdue University, West Lafayette, 47907, IN, USA.
This study introduces a novel data-driven method to build causal models for Alzheimer's disease (AD) biomarkers. The approach accurately predicts future cognitive status by analyzing biomarker trajectories without prior assumptions.
Area of Science:
- Computational biology
- Neuroscience
- Biostatistics
Background:
- Alzheimer's disease (AD) research generates vast biomarker data, necessitating robust modeling techniques.
- Existing mathematical models for AD biomarkers are often empirical or based on pre-existing hypotheses of disease pathophysiology.
- Deriving causal models and parameters purely from data, without hypothesis bias, remains a significant challenge in computational modeling.
Purpose of the Study:
- To develop and validate an innovative, data-driven approach for constructing and parameterizing causal models of Alzheimer's disease biomarker trajectories.
- To overcome limitations of hypothesis-driven and purely empirical modeling in computational causal modeling for AD.
- To enable personalized prediction of disease progression using learned causal relationships.
Main Methods:
- Developed an integrated computational approach combining causal model learning, population parameterization, and parameter sensitivity analysis.
- Applied the methodology to the Alzheimer's Disease Neuroimaging Initiative (ADNI) database, a large multicenter collection of AD biomarker data.
- Incorporated personalized prediction capabilities for individual subject modeling.
Main Results:
- Successfully revealed several data-driven causal models characterizing biomarker trajectories across different stages of Alzheimer's disease.
- Demonstrated the calibration of personalized models for individual subjects within the study cohort.
- Achieved accurate predictions of future cognitive status based on the developed personalized models.
Conclusions:
- The proposed data-driven modeling approach effectively builds and parameterizes causal models for Alzheimer's disease biomarkers.
- This methodology facilitates a deeper understanding of AD pathophysiology and biomarker dynamics.
- Personalized models derived from this approach show promise for accurate prediction of cognitive decline in Alzheimer's disease patients.
More Related Videos
09:38Generalized Psychophysiological Interaction PPI Analysis of Memory Related Connectivity in Individuals at Genetic Risk for Alzheimer's Disease
Published on: November 14, 2017
06:52Fabrication of Amyloid-β-Secreting Alginate Microbeads for Use in Modelling Alzheimer's Disease
Published on: July 6, 2019
Related Concept Videos
Alzheimer's Disease: Overview
The clinical diagnosis of AD hinges on the presence of memory and other cognitive impairments. Biomarkers, such as changes in Aβ...
Alzheimer's Disease: Treatment
Dementia
The progression of dementia is generally gradual....