Aging-related volume changes in the brain and cerebrospinal fluid using artificial intelligence-automated...

Area of Science:

• Neuroimaging
• Artificial Intelligence
• Gerontology

Background:

• Aging leads to changes in brain structure and fluid volumes.
• Accurate segmentation of brain and cerebrospinal fluid (CSF) is crucial for understanding these changes.
• Artificial intelligence (AI) offers potential for automated and reliable volumetric analysis.

Purpose of the Study:

• To assess the reliability of an AI application for segmenting brain and CSF volumes.
• To evaluate age-related changes in brain and CSF volumes in healthy adults.

Main Methods:

• AI-based automatic segmentation of 21 brain and 5 CSF subregions from 3D T1-weighted MRI in 133 healthy volunteers (age 21-92).
• Comparison of automatically segmented total ventricle and subarachnoid space volumes with manual segmentations using intra-class correlation and Bland-Altman analysis.

Main Results:

• High reliability for AI segmentation: intra-class correlations of 0.986 (ventricles) and 0.882 (subarachnoid spaces).
• CSF volume increased significantly with age (approx. 30 mL/decade), from 18.7% in 20s to 33.7% in those >80.
• Ventricle volume showed minimal change (<2%) until age 60, then gradually increased.

Conclusions:

• The AI auto-segmentation application reliably quantifies CSF volumes.
• Intracranial CSF volume increases linearly with aging due to brain volume reduction.
• Ventricle volume remains stable until 60, then increases, potentially aiding in understanding adult hydrocephalus pathogenesis.