Individual identification for different age groups using functional connectivity strength.

Area of Science:

• Neuroscience
• Brain Imaging
• Aging Research

Background:

• Individual differences in brain functional networks are pronounced with age.
• Aging is a risk factor for neurodegenerative diseases like Alzheimer's.
• Understanding age-related network changes is crucial.

Purpose of the Study:

• To explore discrepant functional network patterns in the aging population.
• To classify different age groups using novel atlases and functional connectivity strength (FCS).
• To evaluate the impact of global signal regression and feature selection methods.

Main Methods:

• Utilized two novel atlases for feature extraction.
• Employed whole and intra-hemispheric functional connectivity strength (FCS) for classification.
• Applied Support Vector Machine-Recursive Feature Elimination (SVM-RFE) for feature selection and Support Vector Machine (SVM) for classification.
• Assessed classifier robustness using Receiver Operating Characteristic (ROC) curves.

Main Results:

• Functional connectivity strength (FCS) effectively distinguished between different age groups.
• SVM-RFE enhanced classification accuracy and identified discriminative features.
• Classifiers showed similar performance regardless of the atlas used.

Conclusions:

• Successfully differentiated young, middle-aged, and elderly groups based on FCS.
• The SVM-RFE and SVM classifier approach demonstrated robustness, independence from specific atlases, and insensitivity to global signal effects.
• These findings suggest potential applications in future disease diagnosis using brain network analysis.