Characterizing Functional Connectivity Differences in Aging Adults using Machine Learning on Resting State fMRI Data.

Area of Science:

• Neuroscience
• Brain Imaging
• Machine Learning

Background:

• The brain's intrinsic connectivity networks (ICNs) are functionally linked regions active during rest.
• Resting state functional magnetic resonance imaging (rs-fMRI) offers a task-independent method to study ICNs.
• Understanding age-related changes in ICNs is crucial for cognitive neuroscience.

Purpose of the Study:

• To investigate age-related differences in four major ICNs using rs-fMRI.
• To apply machine learning techniques, specifically Support Vector Machines (SVM), for classification and prediction of age.
• To assess the efficacy of SVM in detecting age-related variations in brain network connectivity.

Main Methods:

• Utilized rs-fMRI data from participants.
• Applied a linear Support Vector Machine (SVM) classifier to differentiate between young and old subjects.
• Employed a linear Support Vector Regressor (SVR) to predict continuous age.

Main Results:

• The SVM classifier achieved 84% accuracy in distinguishing between young and old individuals (p < 1x10^-7).
• The SVR age predictor demonstrated significant performance in continuous age prediction (R^2 = 0.419, p < 1x10^-8).
• Significant age-related differences in intrinsic brain connectivity were detected by SVM methods.

Conclusions:

• rs-fMRI can identify age-related differences in intrinsic brain connectivity.
• SVM algorithms are effective tools for classifying age groups and predicting age based on rs-fMRI data.
• These findings highlight the potential of machine learning in neuroimaging for understanding brain aging.