Generalization of diffusion magnetic resonance imaging–based brain age prediction model through transfer learning

• Published in: NeuroImage (Orlando, Fla.) Vol. 217; p. 116831
• Main Authors: Chen, Chang-Le, Hsu, Yung-Chin, Yang, Li-Ying, Tung, Yu-Hung, Luo, Wen-Bin, Liu, Chih-Min, Hwang, Tzung-Jeng, Hwu, Hai-Gwo, Isaac Tseng, Wen-Yih
• Format: Journal Article
• Language: English
• Published: United States Elsevier Inc 15.08.2020; Elsevier Limited; Elsevier
• Subjects: Adolescent; Adult; Age; Aged; Aged, 80 and over; Aging; Brain - diagnostic imaging; Brain - growth & development; Brain age; Brain research; Datasets; Diffusion harmonization; Diffusion Magnetic Resonance Imaging; Diffusion MRI; Experiments; Feasibility Studies; Female; Hospitals; Humans; Image Processing, Computer-Assisted; Learning algorithms; Machine Learning; Magnetic resonance imaging; Male; Medical imaging; Mental disorders; Middle Aged; Neural network; Neurodegeneration; Neuroimaging; Prediction models; Predictive Value of Tests; Reproducibility of Results; Schizophrenia; Schizophrenia - diagnostic imaging; Schizophrenic Psychology; Statistical analysis; Transfer learning; Transfer, Psychology - physiology; White matter; Young Adult; Diffusion harmonization; Diffusion MRI; Neural network; Transfer learning; White matter; Brain age
• ISSN: 1053-8119; 1095-9572
• DOI: 10.1016/j.neuroimage.2020.116831

Brain age prediction models using diffusion magnetic resonance imaging (dMRI) and machine learning techniques enable individual assessment of brain aging status in healthy people and patients with brain disorders. However, dMRI data are notorious for high intersite variability, prohibiting direct application of a model to the datasets obtained from other sites. In this study, we generalized the dMRI-based brain age model to different dMRI datasets acquired under different imaging conditions. Specifically, we adopted a transfer learning approach to achieve domain adaptation. To evaluate the performance of transferred models, brain age prediction models were constructed using a large dMRI dataset as the source domain, and the models were transferred to three target domains with distinct acquisition scenarios. The experiments were performed to investigate (1) the tuning data size needed to achieve satisfactory performance for brain age prediction, (2) the feature types suitable for different dMRI acquisition scenarios, and (3) performance of the transfer learning approach compared with the statistical covariate approach. By tuning the models with relatively small data size and certain feature types, optimal transferred models were obtained with significantly improved prediction performance in all three target cohorts (p ​< ​0.001). The mean absolute error of the predicted age was reduced from 13.89 to 4.78 years in Cohort 1, 8.34 to 5.35 years in Cohort 2, and 8.74 to 5.64 years in Cohort 3. The test–retest reliability of the transferred model was verified using dMRI data acquired at two timepoints (intraclass correlation coefficient ​= ​0.950). Clinical sensitivity of the brain age prediction model was investigated by estimating the brain age in patients with schizophrenia. The prediction made by the transferred model was not significantly different from that made by the reference model. Both models predicted significant brain aging in patients with schizophrenia as compared with healthy controls (p ​< ​0.001); the predicted age difference of the transferred model was 4.63 and 0.26 years for patients and controls, respectively, and that of the reference model was 4.39 and −0.09 years, respectively. In conclusion, transfer learning approach is an efficient way to generalize the dMRI-based brain age prediction model. Appropriate transfer learning approach and suitable tuning data size should be chosen according to different dMRI acquisition scenarios.