Alzheimer's Disease Early Diagnosis Using Manifold-Based Semi-Supervised Learning.

Area of Science:

• Medical Imaging and Diagnostics
• Machine Learning in Healthcare
• Neuroscience

Background:

• Alzheimer's disease (AD) is a leading cause of death, necessitating early prediction and prevention.
• Diagnosing AD involves complex, multivariate, and heterogeneous data from various medical tests.
• Manual analysis of medical imaging data for AD diagnosis is challenging and time-consuming.

Purpose of the Study:

• To propose a novel, efficient approach for early Alzheimer's disease diagnosis using brain MRI classification.
• To develop a semi-supervised learning framework for accurate prediction of AD in its early stages.
• To improve the accuracy and reduce the error rate in classifying mild Alzheimer's disease (MCI) and normal cognition (NC) using MRI data.

Main Methods:

• Voxel morphometry analysis to extract critical AD-related features from MRI and gray matter (GM) volumes.
• Principal Component Analysis (PCA) for dimension reduction of extracted features to enhance analysis speed and accuracy.
• A hybrid manifold learning framework with label propagation for classifying MCI/NC using limited labeled training data.

Main Results:

• The proposed method achieved a classification accuracy of 93.86% on the OASIS database of MRI brain images.
• The approach demonstrated a 3% lower error rate compared to the best existing methods for AD classification.
• Effective extraction and utilization of discriminative features between healthy and Alzheimer's-affected brains were achieved.

Conclusions:

• The developed manifold-based semi-supervised learning framework offers an efficient and accurate method for early Alzheimer's disease diagnosis.
• The hybrid approach effectively leverages MRI data for improved classification of mild Alzheimer's disease.
• This novel technique shows significant potential in advancing the early detection and management of Alzheimer's disease.