Biomarkers.

Area of Science:

• Neuroscience
• Medical Imaging
• Machine Learning

Background:

• Mild Cognitive Impairment (MCI) is a precursor to Alzheimer's disease.
• Gray matter decay in specific brain regions (Hippocampus, Entorhinal cortex, Cerebral cortex, Frontal lobe, Temporal lobe, Parietal lobe, Occipital lobe) is linked to cognitive decline.
• Accurate prediction of MCI conversion is crucial for timely intervention.

Purpose of the Study:

• To develop an improved prediction model for MCI conversion to Alzheimer's disease.
• To identify key brain regions and their interconnections involved in MCI progression.
• To leverage novel feature extraction techniques for enhanced diagnostic accuracy.

Main Methods:

• A novel auto-encoder based method was used for feature extraction from seven key brain regions.
• A graph representation was constructed, with nodes as brain regions and edge weights based on feature similarity.
• Graph features were flattened into a 1-D vector for classification using Support Vector Machine (SVM) with linear kernel and forward feature selection.

Main Results:

• The method was validated on 75 MCI-Converts and 112 MCI-Non Converts from the Alzheimer's Disease Neuroimaging Initiative (ADNI) dataset.
• The classification achieved a high F-score of 95.4%.
• This performance surpasses current state-of-the-art methods in predicting MCI conversion.

Conclusions:

• The proposed method offers region-specific insights into MCI conversion prediction.
• It enables network analysis of brain region interactions, revealing underlying mechanisms of MCI progression.
• The approach demonstrates high accuracy, generalizability, and potential for clinical application.