SUPERVISED DEEP TREE IN ALZHEIMER'S DISEASE.

Area of Science:

• Neuroscience
• Artificial Intelligence
• Medical Diagnostics

Background:

• Alzheimer's disease (AD) is a progressive neurodegenerative disorder with pathological changes beginning decades before clinical symptoms.
• Early diagnosis of AD is crucial for intervention and treatment due to the irreversible nature of its pathology.
• Current diagnostic methods often struggle to capture the continuum of AD progression, hindering a full understanding of its mechanisms.

Purpose of the Study:

• To develop a novel method for integrating Alzheimer's disease progression and individual prediction.
• To model the continuum of AD pathology using a supervised deep tree model.
• To enable more accurate predictions for new subjects based on their disease progression stage.

Main Methods:

• Proposed a supervised deep tree model (SDTree) to represent AD progression.
• Utilized nonlinear reversed graph embedding to model progression as a tree in a latent space.
• Encoded the continuum of AD progression into the tree structure for analysis and prediction.

Main Results:

• The SDTree model successfully represented the continuum of AD progression.
• The model demonstrated capability in making predictions for new subjects.
• Achieved promising results on a classification task using the Alzheimer's Disease Neuroimaging Initiative dataset.

Conclusions:

• The developed SDTree model offers a new approach to understanding and predicting Alzheimer's disease progression.
• This method enhances the ability to describe the continuum of AD pathology.
• The findings support the potential of deep learning models for early AD diagnosis and personalized intervention.