Unveiling Alzheimer's disease through brain age estimation using multi-kernel regression network and magnetic...

Area of Science:

• Neuroimaging
• Artificial Intelligence
• Biomarker Discovery

Background:

• Structural MRI reveals age-related brain changes.
• Brain-Predicted Age Difference (Brain-PAD) can indicate neurocognitive issues.
• Accurate brain age prediction aids in identifying brain abnormalities.

Purpose of the Study:

• To develop an accurate brain age prediction model using MRI.
• To utilize Brain-PAD for identifying potential neurocognitive impairments.
• To assess the utility of Brain-PAD as a biomarker for brain health.

Main Methods:

• Implemented a ResNet-50 deep network and multi-kernel extreme learning machine (MKELM) regression.
• Utilized MRI scans, segmenting into gray matter (GM) and white matter (WM) maps.
• Extracted features from GM and WM datasets for brain age prediction.

Main Results:

• Achieved low mean absolute error (MAE) and root mean square error (RMSE) on healthy controls.
• Brain-PAD values were significantly higher in Alzheimer's disease (AD) subjects compared to healthy controls (HC) and mild cognitive impairment (MCI) subjects.
• Identified specific brain region variations (parahippocampus, corpus callosum) in AD subjects.

Conclusions:

• Brain-PAD is a significant biomarker for assessing brain health.
• The developed framework accurately estimates brain age using MRI.
• Findings contribute to understanding brain structure and cognitive health, aiding early neurocognitive disease detection.