Prediction of Early Alzheimer Disease by Hippocampal Volume Changes under Machine Learning Algorithm

Area of Science:

• Neuroimaging and Machine Learning
• Computational Neuroscience
• Biomedical Data Analysis

Background:

• Alzheimer's disease (AD) diagnosis relies on identifying neurodegenerative changes.
• Hippocampal volume reduction is a key indicator in early-stage AD.
• Machine learning (ML) offers potential for automated prediction of AD.

Purpose of the Study:

• To evaluate the application value of different ML algorithms for early AD prediction.
• To assess prediction accuracy based on hippocampal volume changes from MRI scans.
• To compare the performance of Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF) models.

Main Methods:

• Utilized 84 cases from the American Alzheimer's Disease Neuroimaging Initiative (ADNI) database.
• Categorized participants into normal, early mild cognitive impairment (e-MCI), and late mild cognitive impairment (l-MCI) groups.
• Extracted hippocampal subregion volume features from MRI, trained and tested SVM, DT, and RF models for e-MCI prediction.

Main Results:

• Significant hippocampal volume reductions were observed in e-MCI and l-MCI groups compared to controls.
• Hippocampal subregion volumes correlated positively with cognitive scores (e.g., logical memory test-delay recall).
• The RF model achieved the highest predictive accuracy (98.33%) for the training set and outperformed DT for e-MCI and l-MCI prediction.

Conclusions:

• ML models, particularly RF, based on hippocampal volume changes can effectively predict early AD.
• These findings provide a valuable reference for the diagnosis and treatment of AD patients.
• Hippocampal subregion analysis using ML shows promise for early AD detection.