Exploratory analysis of the associations of the brain age gap with cognitive function and amyloid-β accumulation: participants selection based on metabolic and physiological blood markers.
Authors
Affiliations (8)
Affiliations (8)
- Department of Radiological Technology, Graduate School of Health Science, Juntendo University, Tokyo, Japan.
- Department of Radiological Technology, Graduate School of Health Science, Juntendo University, Tokyo, Japan; Department of Radiology, Tokyo Medical University, Tokyo, Japan; Center for Evolutionary Cognitive Sciences, Graduate School of Art and Sciences, The University of Tokyo, Tokyo, Japan.
- Department of Radiological Technology, Kokura Memorial Hospital, Fukuoka, Japan.
- Department of Radiology, Tokyo Medical University, Tokyo, Japan; Jikei University School of Medicine, Division of Regenerative Medicine, Japan; National Institute of Advanced Industrial Science and Technology (AIST), Tsukuba, Japan.
- Department of Radiology, LSI Sapporo Clinic, Hokkaido, Japan.
- Department of Radiological Technology, Faculty of Health Sciences, Juntendo University, Tokyo, Japan.
- Department of Radiology, Juntendo University Graduate School of Medicine, Tokyo, Japan.
- Department of Radiological Technology, Graduate School of Health Science, Juntendo University, Tokyo, Japan. Electronic address: [email protected].
Abstract
The brain age gap (BAG) is defined as the difference between brain age estimated from MRI using artificial intelligence and chronological age, and has been proposed as a biomarker reflecting aging and neurodegeneration. However, the association between BAG and dementia-related biomarkers has yielded inconsistent findings in previous studies. Conventional training datasets have primarily been defined based on medical history and MRI findings, which may have included participants with underlying metabolic and physiological abnormalities, potentially contributing to these inconsistent results. In this exploratory study, we examined whether incorporating metabolic and physiological blood test parameters into participant selection would influence the association between BAG and dementia-related biomarkers. Using MRI and blood test data from 680 participants, we developed two machine learning-based brain age models. Model 1 was based on conventional selection criteria, whereas Model 2 additionally incorporated blood test parameters. BAG was estimated in 38 participants who underwent MRI and amyloid PET imaging. Differences between amyloid-positive and amyloid-negative groups were assessed using the Wilcoxon rank-sum test, and associations with standardized uptake value ratio, Centiloid scale, and Mini-Mental State Examination were evaluated using Spearman's rank correlation analysis. Model 2 showed higher BAG in the amyloid positive group and exhibited stronger associations with amyloid-β accumulation and cognitive function. These findings suggest that incorporating metabolic and physiological blood test parameters into participant selection may have emphasized the association between BAG and dementia-related biomarkers. However, because the independent test dataset was small, these results should be interpreted as exploratory, and validation in datasets is required.