Related Experiment Video
Updated: Sep 22, 2025
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
15.8K
Estimation of Human Cerebral Atrophy Based on Systemic Metabolic Status Using Machine Learning
Kaoru Sakatani1, Katsunori Oyama2, Lizhen Hu1
1Department of Human and Engineered Environmental Studies, Graduate School of Frontier Sciences, The University of Tokyo, Kashiwa, Japan.
Frontiers in Neurology
|May 19, 2022
Summary
This study shows that deep neural network models can estimate cerebral atrophy using basic blood test data. Even without age information, blood tests alone can predict brain changes, suggesting potential for early dementia screening.
Area of Science:
- Neuroscience
- Biomarkers
- Artificial Intelligence
Background:
- Systemic metabolic disorders are linked to cognitive function decline.
- Deep neural network (DNN) models can estimate cognitive function using basic blood tests without dementia-specific biomarkers.
- This study investigates the utility of DNN models and basic blood data for estimating cerebral atrophy.
Purpose of the Study:
- To assess if basic blood data can be used to estimate cerebral atrophy using a deep neural network (DNN) model.
- To evaluate the performance of a DNN model with and without subject age as input for estimating cerebral atrophy.
- To explore the correlation between systemic metabolic markers in blood and cerebral atrophy.
Main Methods:
- Utilized data from 1,310 subjects from the Brain Doc Bank.
- Assessed cerebral atrophy using an MRI-based index (GM-BHQ).
- Developed two DNN models to estimate GM-BHQ: one using age and blood data, another using only blood data.
Main Results:
- Age showed a strong negative correlation with cerebral atrophy (GM-BHQ, r = -0.71).
- Blood markers like BUN, ALP, and glucose correlated with age and GM-BHQ.
- The DNN model estimated GM-BHQ with significant positive correlations to ground truth (r = 0.70 with age, r = 0.58 without age).
Conclusions:
- Aging is a primary factor influencing cerebral atrophy.
- Basic blood tests, analyzed by DNN models, can estimate cerebral atrophy, potentially serving as a novel screening tool for dementia.
- Blood data reflecting systemic metabolic status may aid in personalized care strategies for cognitive health.
