Dataset Information

Deep transfer learning of structural magnetic resonance imaging fused with blood parameters improves brain age prediction.

ABSTRACT: Machine learning has been applied to neuroimaging data for estimating brain age and capturing early cognitive impairment in neurodegenerative diseases. Blood parameters like neurofilament light chain are associated with aging. In order to improve brain age predictive accuracy, we constructed a model based on both brain structural magnetic resonance imaging (sMRI) and blood parameters. Healthy subjects (n = 93; 37 males; aged 50-85 years) were recruited. A deep learning network was firstly pretrained on a large set of MRI scans (n = 1,481; 659 males; aged 50-85 years) downloaded from multiple open-source datasets, to provide weights on our recruited dataset. Evaluating the network on the recruited dataset resulted in mean absolute error (MAE) of 4.91 years and a high correlation (r = .67, p <.001) against chronological age. The sMRI data were then combined with five blood biochemical indicators including GLU, TG, TC, ApoA1 and ApoB, and 9 dementia-associated biomarkers including ApoE genotype, HCY, NFL, TREM2, Aβ40, Aβ42, T-tau, TIMP1, and VLDLR to construct a bilinear fusion model, which achieved a more accurate prediction of brain age (MAE, 3.96 years; r = .76, p <.001). Notably, the fusion model achieved better improvement in the group of older subjects (70-85 years). Extracted attention maps of the network showed that amygdala, pallidum, and olfactory were effective for age estimation. Mediation analysis further showed that brain structural features and blood parameters provided independent and significant impact. The constructed age prediction model may have promising potential in evaluation of brain health based on MRI and blood parameters.

SUBMITTER: Ren B

PROVIDER: S-EPMC8886664 | biostudies-literature | 2022 Apr

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Deep transfer learning of structural magnetic resonance imaging fused with blood parameters improves brain age prediction.

Ren Bingyu B Wu Yingtong Y Huang Liumei L Zhang Zhiguo Z Huang Bingsheng B Zhang Huajie H Ma Jinting J Li Bing B Liu Xukun X Wu Guangyao G Zhang Jian J Shen Liming L Liu Qiong Q Ni Jiazuan J

Human brain mapping 20211216 5

Machine learning has been applied to neuroimaging data for estimating brain age and capturing early cognitive impairment in neurodegenerative diseases. Blood parameters like neurofilament light chain are associated with aging. In order to improve brain age predictive accuracy, we constructed a model based on both brain structural magnetic resonance imaging (sMRI) and blood parameters. Healthy subjects (n = 93; 37 males; aged 50-85 years) were recruited. A deep learning network was firstly pretra ...[more]

PMID: 34913545

Similar Datasets

Project description:ObjectiveBrain tissue changes dynamically during aging. The purpose of this study was to use synthetic magnetic resonance imaging (syMRI) to evaluate the changes in relaxation values in different brain regions during brain aging and to construct a brain age prediction model.Materials and methodsQuantitative MRI was performed on 1,000 healthy people (≥ 18 years old) from September 2020 to October 2021. T1, T2 and proton density (PD) values were simultaneously measured in 17 regions of interest (the cerebellar hemispheric cortex, pons, amygdala, hippocampal head, hippocampal tail, temporal lobe, occipital lobe, frontal lobe, caudate nucleus, lentiform nucleus, dorsal thalamus, centrum semiovale, parietal lobe, precentral gyrus, postcentral gyrus, substantia nigra, and red nucleus). The relationship between the relaxation values and age was investigated. In addition, we analyzed the relationship between brain tissue values and sex. Finally, the participants were divided into two age groups: < 60 years old and ≥ 60 years old. Logistic regression analysis was carried out on the two groups of data. According to the weight of related factors, a brain age prediction model was established and verified.ResultsWe obtained the specific reference value range of different brain regions of individuals in different age groups and found that there were differences in relaxation values in brain tissue between different sexes in the same age group. Moreover, the relaxation values of most brain regions in males were slightly higher than those in females. In the study of age and brain relaxation, it was found that brain relaxation values were correlated with age. The T1 values of the centrum semiovale increased with age, the PD values of the centrum semiovale increased with age, while the T2 values of the caudate nucleus and lentiform nucleus decreased with age. Seven brain age prediction models were constructed with high sensitivity and specificity, among which the combined T1, T2 and PD values showed the best prediction efficiency. In the training set, the area under the curve (AUC), specificity and sensitivity were 0.959 [95% confidence interval (CI): 0.945-0.974], 91.51% and 89.36%, respectively. In the test cohort, the above indicators were 0.916 (95% CI: 0.882-0.951), 89.24% and 80.33%, respectively.ConclusionOur study provides specific reference ranges of T1, T2, and PD values in different brain regions from healthy adults of different ages. In addition, there are differences in brain relaxation values in some brain regions between different sexes, which help to provide new ideas for brain diseases that differ according to sex. The brain age model based on synthetic MRI is helpful to determine brain age.

Dataset Information

Deep transfer learning of structural magnetic resonance imaging fused with blood parameters improves brain age prediction.

Publications

Deep transfer learning of structural magnetic resonance imaging fused with blood parameters improves brain age prediction.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets