The Role of Medication Data to Enhance the Prediction of Alzheimer's Progression Using Machine Learning

Area of Science:

• Neuroscience
• Computational Biology
• Pharmacology

Background:

• Alzheimer's disease (AD) management requires early detection of progression.
• Existing studies often overlook longitudinal data and the influence of dementia medications.
• Longitudinal patient data is essential for understanding chronic disease trajectories.

Purpose of the Study:

• To develop a machine learning architecture for early AD progression detection.
• To investigate the impact of dementia drugs on AD progression using multimodal data.
• To compare the efficacy of various machine learning algorithms for predicting AD progression.

Main Methods:

• Utilized a multimodal dataset including Alzheimer's disease (AD) drug information and cognitive scores.
• Applied five machine learning techniques: Support Vector Machine, Random Forest, Logistic Regression, Decision Tree, and K-Nearest Neighbor.
• Conducted extensive experiments on a dataset of 1036 subjects from the Alzheimer's Disease Neuroimaging Initiative (ADNI).

Main Results:

• Fusing drug and cognitive score data improved the cross-validation performance of most machine learning algorithms.
• The study demonstrated the significant role of patient-medication interactions in AD progression.
• Early detection of AD progression was achieved with improved accuracy through multimodal data integration.

Conclusions:

• Machine learning models integrating longitudinal drug and cognitive data can effectively predict Alzheimer's disease progression.
• Dementia medications play a crucial role in influencing the behavioral trajectory of AD.
• Future research should incorporate medication data for more accurate AD progression modeling.