Related Experiment Video
Updated: Mar 22, 2026
Whole-brain Segmentation and Change-point Analysis of Anatomical Brain MRI—Application in Premanifest Huntington's Disease
Published on: June 9, 2018
Quantitative Machine Learning Analysis of Brain MRI Morphology throughout Aging
121000 W Ten Mile Rd., Department of Math and Computer Science, Lawrence Technological University, Southfield, Michigan 48075, United States.
Background:
While cognition is clearly affected by aging, it is unclear whether the process of brain aging is driven solely by accumulation of environmental damage, or involves biological pathways.
Methods:
We applied quantitative image analysis to profile the alteration of brain tissues during aging. A dataset of 463 brain MRI images taken from a cohort of 416 subjects was analyzed using a large set of low-level numerical image content descriptors computed from the entire brain MRI images. The correlation between the numerical image content descriptors and the age was computed, and the alterations of the brain tissues during aging were quantified and profiled using machine learning.
Results:
The comprehensive set of global image content descriptors provides high Pearson correlation of ~0.9822 with the chronological age, indicating that the machine learning analysis of global features is sensitive to the age of the subjects. Profiling of the predicted age shows several periods of mild changes, separated by shorter periods of more rapid alterations. The periods with the most rapid changes were around the age of 55, and around the age of 65.
Conclusion:
The results show that the process of brain aging of is not linear, and exhibit short periods of rapid aging separated by periods of milder change. These results are in agreement with patterns observed in cognitive decline, mental health status, and general human aging, suggesting that brain aging might not be driven solely by accumulation of environmental damage. Code and data used in the experiments are publicly available.
