Age Prediction Using Resting-State Functional MRI

Area of Science:

• Neuroscience
• Radiology
• Gerontology

Background:

• Cognitive aging varies significantly, making brain health assessment crucial.
• Brain age, a marker of neural health, can differ from chronological age and is linked to mortality and depression.
• Functional brain imaging offers deeper insights into aging than structural methods alone.

Purpose of the Study:

• To develop a predictive model for brain age using resting-state functional MRI (rsfMRI).
• To identify neural network correlations associated with abnormal brain aging.
• To establish a robust reference model for assessing brain health in adults.

Main Methods:

• Utilized rsfMRI data from 176 healthy participants (aged 18-78).
• Employed the Least Absolute Shrinkage and Selection Operator (LASSO) to identify 39 predictive rsfMRI correlations.
• Developed a normal reference model by removing 68 outliers, achieving a low prediction error.

Main Results:

• The developed model achieved a leave-one-out mean absolute error of 2.48 years.
• Abnormal aging predictors were identified and linked to the Default Mode Network (DMN).
• The model demonstrated superior accuracy compared to existing published models.

Conclusions:

• The study provides an accurate model for predicting brain age and screening for abnormal aging.
• The Default Mode Network (DMN) plays a significant role in brain aging processes.
• This approach can identify brain aging issues before cognitive impairment is evident.