Early Prediction of Alzheimer's Disease Using Null Longitudinal Model-Based Classifiers

Area of Science:

• Neuroimaging
• Biomarkers
• Computational Neuroscience

Background:

• Incipient Alzheimer's Disease (AD) diagnosis is challenging due to overlapping brain changes with normal aging and mild cognitive impairment (MCI).
• Understanding normative age-related brain atrophy and growth is crucial for differentiating pathological changes.
• Early detection of AD and MCI is vital for timely intervention and improved patient outcomes.

Purpose of the Study:

• To develop an age-based null model of brain structure changes using Magnetic Resonance Imaging (MRI) data.
• To differentiate normal aging from early pathological changes associated with MCI and AD.
• To utilize machine learning for early diagnosis and prediction of conversion to AD.

Main Methods:

• Linear Mixed Effects modeling of 166 MRI-based biomarkers over a 5-year follow-up in healthy controls.
• Development of an age-based null model characterizing normal brain atrophy and growth patterns.
• Application of residual-based Support Vector Machines (SVM) for classification and prediction of MCI/AD conversion.

Main Results:

• Significant reductions in cortical volumes and thicknesses, with notable gender differences and greater hippocampal atrophy.
• High classification accuracies for AD vs. Healthy Controls (HC) (94.11%) and MCI vs. HC (83.77%).
• SVM models predicted MCI to AD conversion 1.9 years earlier in females and 1.4 years earlier in males compared to standard methods.

Conclusions:

• The developed null model effectively characterizes normal age-related brain changes.
• Residual-based SVM analysis demonstrates potential for early and accurate diagnosis of MCI and AD.
• This approach offers a promising tool for identifying individuals at risk of AD progression, enabling earlier clinical intervention.