Brain age prediction across the human lifespan using multimodal MRI data

Area of Science:

• Neuroimaging
• Biomarkers
• Aging Research

Background:

• Biological age estimation using brain information may identify neurological syndromes.
• Multimodal brain magnetic resonance imaging (MRI) features offer potential biomarkers for aging.

Purpose of the Study:

• To investigate how multimodal brain imaging data improve age prediction.
• To explore the predictive power of various structural and functional MRI features.
• To analyze the contribution of different brain hemispheres to age prediction.

Main Methods:

• Utilized partial least squares regression (PLSR) on multimodal brain MRI data.
• Analyzed a longevity dataset spanning ages 6 to 85 years.
• Evaluated ten different MRI features for their age prediction accuracy.

Main Results:

• Most multimodal MRI features, excluding cortical thickness, enhanced brain age prediction.
• The left hemisphere demonstrated a more significant contribution to age prediction than the right.
• A nonlinear relationship was observed between predicted age and the volume of MRI data.

Conclusions:

• Multimodal brain MRI data, particularly from the left hemisphere, are valuable for predicting chronological age.
• This approach offers a novel perspective on brain aging and potential neurological disorder biomarkers.
• Findings contribute to understanding the complex relationship between brain structure, function, and the aging process.