Classification of Alzheimer's Disease Stages: An Approach Using PCA-Based Algorithm

Area of Science:

• Neuroscience
• Medical Imaging
• Machine Learning

Background:

• Early diagnosis of Alzheimer's disease (AD) is crucial for effective patient management and treatment.
• Principal Component Analysis (PCA) is a recognized algorithm for feature extraction and dimensionality reduction, notably in face recognition.
• The hippocampus is a brain region significantly impacted by Alzheimer's disease.

Purpose of the Study:

• To propose and evaluate a novel classification approach for Alzheimer's disease (AD) stages.
• To adapt a principal component analysis (PCA)-based algorithm for the classification of AD using neuroimaging data.
• To effectively classify different stages of AD, including mild cognitive impairment and advanced stages.

Main Methods:

• Feature extraction was performed on 100 images from 10 children using PCA, involving covariance matrix construction and eigenvalue/eigenvector calculation.
• Functional magnetic resonance imaging (fMRI) and magnetic resonance imaging (MRI) data were sourced from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database.
• Clusters of voxels from the hippocampus region were selected for mild cognitive impairment, AD stage 1, stage 2, and stage 3, utilizing eigenvectors derived from fMRI data.

Main Results:

• The PCA-based algorithm demonstrated effectiveness in classifying voxels corresponding to different stages of Alzheimer's disease.
• Eigenvectors associated with maximum eigenvalues of fMRI data proved instrumental in the classification process.
• The study successfully applied a face-recognition-inspired algorithm to neuroimaging data for AD staging.

Conclusions:

• The proposed PCA-based algorithm offers a viable method for the classification of Alzheimer's disease stages using fMRI data.
• The approach highlights the potential of leveraging established algorithms from other domains, like face recognition, for neurodegenerative disease research.
• Accurate classification of AD stages is essential for timely intervention and personalized treatment strategies.