Explainable Machine Learning Models for Alzheimer's Diagnosis Using Routine and Low-Cost Clinical Data

Area of Science:

• Neuroscience
• Artificial Intelligence
• Medical Diagnostics

Background:

• Alzheimer's Disease (AD) presents a growing global health concern, characterized by progressive neurodegeneration leading to cognitive decline.
• Current diagnostic methods for AD face challenges including long waiting times for specialist evaluation and high costs.
• There is a critical need for rapid, accessible, and cost-effective diagnostic tools for Alzheimer's Disease.

Purpose of the Study:

• To develop and evaluate an explainable deep learning model for diagnosing Alzheimer's Disease (AD).
• To utilize routine, low-cost clinical data for AD diagnosis, including demographics, patient history, and automatically acquirable neuropsychological test results.
• To compare the diagnostic performance of the developed AI models against clinicians under similar data constraints.

Main Methods:

• Analysis of a large dataset from the National Alzheimer's Coordinating Center (167,364 observations, 1,024 features).
• Development of two classification models: a deep neural network and a Light Gradient Boosting Machine.
• Focus on explainable AI to ensure transparency in the diagnostic process.

Main Results:

• The deep neural network achieved 90% accuracy and a 0.96 ROC-AUC.
• The Light Gradient Boosting Machine model reached 90% accuracy with a 0.97 ROC-AUC.
• Both models demonstrated diagnostic performance comparable to, and in some aspects superior to, clinicians.

Conclusions:

• Explainable AI models can effectively diagnose Alzheimer's Disease using readily available clinical data.
• These AI tools offer a promising, cost-effective, and efficient alternative or supplement to traditional diagnostic pathways.
• The study highlights the potential of AI in addressing the diagnostic challenges associated with Alzheimer's Disease.