Exploring the Impact of Biological Biomarkers on Brain Aging Through Deep Learning Model
Poster No:
896
Submission Type:
Abstract Submission
Authors:
Lee Cheng-Hsiu1, Chen-Yuan Kuo1,2, Pei-Lin Lee3, Lee Han-Jui4,5, Kun-Hsien Chou6,1, Lin Ching-Po7,1,6
Institutions:
1Institute of Neuroscience, National Yang Ming Chiao Tung University, Taipei, Taiwan, 2Department of Neurology, Neurological Institute, Taipei Veterans General Hospital, Taipei, Taiwan, 3Center for Healthy Longevity and Aging Sciences, National Yang Ming Chiao Tung University, Taipei, Taiwan, 4Department of Radiology, Taipei Veterans General Hospital, Taipei, none, 5School of Medicine, National Yang Ming Chiao Tung University, Taipei, Taiwan, 6Brain research center, National Yang Ming Chiao Tung University, Taipei, Taiwan, 7Department of Education and Research, Taipei City Hospital, Taipei, Taiwan
First Author:
Co-Author(s):
Chen-Yuan Kuo
Institute of Neuroscience, National Yang Ming Chiao Tung University|Department of Neurology, Neurological Institute, Taipei Veterans General Hospital
Taipei, Taiwan|Taipei, Taiwan
Institute of Neuroscience, National Yang Ming Chiao Tung University|Department of Neurology, Neurological Institute, Taipei Veterans General Hospital
Taipei, Taiwan|Taipei, Taiwan
Pei-Lin Lee
Center for Healthy Longevity and Aging Sciences, National Yang Ming Chiao Tung University
Taipei, Taiwan
Center for Healthy Longevity and Aging Sciences, National Yang Ming Chiao Tung University
Taipei, Taiwan
Lee Han-Jui
Department of Radiology, Taipei Veterans General Hospital|School of Medicine, National Yang Ming Chiao Tung University
Taipei, none|Taipei, Taiwan
Department of Radiology, Taipei Veterans General Hospital|School of Medicine, National Yang Ming Chiao Tung University
Taipei, none|Taipei, Taiwan
Kun-Hsien Chou
Brain research center, National Yang Ming Chiao Tung University|Institute of Neuroscience, National Yang Ming Chiao Tung University
Taipei, Taiwan|Taipei, Taiwan
Brain research center, National Yang Ming Chiao Tung University|Institute of Neuroscience, National Yang Ming Chiao Tung University
Taipei, Taiwan|Taipei, Taiwan
Lin Ching-Po
Department of Education and Research, Taipei City Hospital|Institute of Neuroscience, National Yang Ming Chiao Tung University|Brain research center, National Yang Ming Chiao Tung University
Taipei, Taiwan|Taipei, Taiwan|Taipei, Taiwan
Department of Education and Research, Taipei City Hospital|Institute of Neuroscience, National Yang Ming Chiao Tung University|Brain research center, National Yang Ming Chiao Tung University
Taipei, Taiwan|Taipei, Taiwan|Taipei, Taiwan
Introduction:
As individuals age, the brain undergoes gradual structural and functional changes intricately linked to physiological and metabolic biomarkers. Previous studies have demonstrated that biomarkers associated with chronic diseases, such as diabetes and obesity, significantly influence brain aging [1][2]. The interplay between these biomarkers and the brain forms a complex biological network, reflecting overall health and contributing to neurodegenerative risks and functional decline. However, the intricate relationship between brain aging and these biomarkers remained poorly understood. This study employed deep learning to estimate brain biological age and elucidated the relationships between these biomarkers and brain aging.
Methods:
Clinical dataset
This study included a total of 458 healthy individuals (age range = 20-83, 280 M/178 F) recruited from Taipei Veterans General Hospital, Taiwan. All participants underwent brain MRI scans using 3T MRI. Clinical assessments and laboratory tests were performed, including age, sex, BMI, HbA1c, LDL, HDL, AST, ALT, current smoking status, diabetes prevalence, hemoglobin levels, triglycerides, eGFR, and total cholesterol levels.
Training dataset and Brain age model construction
We gathered 1,400 healthy participants from five sites in Taiwan as the training dataset. All T1-weighted images were preprocessed following standardized protocols (bias correction, brain extraction, registration, and min-max normalization). A 3D DenseNet121 architecture from Project MONAI [3] was utilized to construct the brain age prediction model. During the training process, the model was configured with a 16 batch size, 150 epochs, Adam optimizer, and loss function (MAE). To prevent overfitting, early stopping was implemented. The trained model was subsequently applied to the clinical dataset to estimate the brain age.
Statistical analysis
Brain age gap (BAG; brain age - chronological age) was calculated for each participant in the clinical dataset. Based on BAG values (±2 years), participants were categorized into three groups: delayed aging (DA), normal aging (NA), and accelerated aging (AA). Descriptive statistics and hypothesis testing were conducted to examine significant differences across groups. Furthermore, odds ratio (OR) analysis was performed to assess the impact of each biological indicator on BAG. Forest plots were used to illustrate the effects and confidence intervals of these indicators within each group, providing a comprehensive visualization of their associations with brain aging.
This study included a total of 458 healthy individuals (age range = 20-83, 280 M/178 F) recruited from Taipei Veterans General Hospital, Taiwan. All participants underwent brain MRI scans using 3T MRI. Clinical assessments and laboratory tests were performed, including age, sex, BMI, HbA1c, LDL, HDL, AST, ALT, current smoking status, diabetes prevalence, hemoglobin levels, triglycerides, eGFR, and total cholesterol levels.
Training dataset and Brain age model construction
We gathered 1,400 healthy participants from five sites in Taiwan as the training dataset. All T1-weighted images were preprocessed following standardized protocols (bias correction, brain extraction, registration, and min-max normalization). A 3D DenseNet121 architecture from Project MONAI [3] was utilized to construct the brain age prediction model. During the training process, the model was configured with a 16 batch size, 150 epochs, Adam optimizer, and loss function (MAE). To prevent overfitting, early stopping was implemented. The trained model was subsequently applied to the clinical dataset to estimate the brain age.
Statistical analysis
Brain age gap (BAG; brain age - chronological age) was calculated for each participant in the clinical dataset. Based on BAG values (±2 years), participants were categorized into three groups: delayed aging (DA), normal aging (NA), and accelerated aging (AA). Descriptive statistics and hypothesis testing were conducted to examine significant differences across groups. Furthermore, odds ratio (OR) analysis was performed to assess the impact of each biological indicator on BAG. Forest plots were used to illustrate the effects and confidence intervals of these indicators within each group, providing a comprehensive visualization of their associations with brain aging.
Results:
The risk factor profiles across the groups (DA, NA, AA) revealed significant differences. Although individuals in the AA group were significantly younger (mean age: 52.94 years, p=0.009), they exhibited a larger BAG (4.41±1.84 years) than other groups. Additionally, the AA group exhibited higher BMI (25.54 ± 4.00, p=0.040), triglyceride levels (152.75 ± 112.26 mg/dL, p=0.012), and hemoglobin levels (14.67 ± 1.47, p = 0.040). Men were significantly overrepresented in the AA group (83.1%, p=0.001) (Figure 1).
Significant risk factors for the AA group included male sex (OR=3.56, CI: 1.75–7.24), obesity (OR=2.25, CI: 1.07–3.61), and diabetes (OR=2.19, 95% CI: 1.02–4.71). In contrast, the DA group was characterized by a lower prevalence of male sex (OR = 0.54, CI: 0.36–0.81) and diabetes (OR = 0.61, CI: 0.38–0.98) (Figure 2).
Significant risk factors for the AA group included male sex (OR=3.56, CI: 1.75–7.24), obesity (OR=2.25, CI: 1.07–3.61), and diabetes (OR=2.19, 95% CI: 1.02–4.71). In contrast, the DA group was characterized by a lower prevalence of male sex (OR = 0.54, CI: 0.36–0.81) and diabetes (OR = 0.61, CI: 0.38–0.98) (Figure 2).
·Figure 1
·Figure 2
Conclusions:
This study utilized a deep learning-based brain age model to investigate accelerated brain aging and its relationship with physiological biomarkers. Our findings revealed that the AA group exhibited higher BMI, triglycerides, hemoglobin, and a greater proportion of men, all linked to elevated brain aging risk. Odds ratio analysis further identified obesity, smoking, and diabetes as significant risk factors, with significant associations observed in the AA group. Our study highlighted the importance of combining biological indicators with brain age predictions to identify high-risk populations.
Lifespan Development:
Aging 1
Modeling and Analysis Methods:
Classification and Predictive Modeling 2
Novel Imaging Acquisition Methods:
Anatomical MRI
Keywords:
Aging
Cognition
Data analysis
Machine Learning
MRI
STRUCTURAL MRI
1|2Indicates the priority used for review
Abstract Information
By submitting your proposal, you grant permission for the Organization for Human Brain Mapping (OHBM) to distribute your work in any format, including video, audio print and electronic text through OHBM OnDemand, social media channels, the OHBM website, or other electronic publications and media.
The Open Science Special Interest Group (OSSIG) is introducing a reproducibility challenge for OHBM 2025. This new initiative aims to enhance the reproducibility of scientific results and foster collaborations between labs. Teams will consist of a “source” party and a “reproducing” party, and will be evaluated on the success of their replication, the openness of the source work, and additional deliverables. Click here for more information. Propose your OHBM abstract(s) as source work for future OHBM meetings by selecting one of the following options:
Please indicate below if your study was a "resting state" or "task-activation” study.
Healthy subjects only or patients (note that patient studies may also involve healthy subjects):
Was this research conducted in the United States?
Were any human subjects research approved by the relevant Institutional Review Board or ethics panel? NOTE: Any human subjects studies without IRB approval will be automatically rejected.
Were any animal research approved by the relevant IACUC or other animal research panel? NOTE: Any animal studies without IACUC approval will be automatically rejected.
Please indicate which methods were used in your research:
For human MRI, what field strength scanner do you use?
Provide references using APA citation style.
2.Sellbom, K., & Gunstad, J. (2012). Cognitive decline and obesity: role of inflammation and vascular disease. Obesity Reviews.
3.Huang, Gao, et al. (2017). Densely connected convolutional networks. Proceedings of the IEEE conference on computer vision and pattern recognition.