Abstract
Brain age prediction from brain MRI scans not only helps improve brain ageing modelling generally, but also provides benchmarks for predictive analysis methods. Brain-age delta, which is the difference between a subject's predicted age and true age, has become a meaningful biomarker for the health of the brain. Here, we report the details of our brain age prediction models and results in the Predictive Analysis Challenge 2019. The aim of the challenge was to use T1-weighted brain MRIs to predict a subject's age in multicentre datasets. We apply a lightweight deep convolutional neural network architecture, Simple Fully Convolutional Neural Network (SFCN), and combined several techniques including data augmentation, transfer learning, model ensemble, and bias correction for brain age prediction. The model achieved first place in both of the two objectives in the PAC 2019 brain age prediction challenge: Mean absolute error (MAE) = 2.90 years without bias removal (Second Place = 3.09 yrs; Third Place = 3.33 yrs), and MAE = 2.95 years with bias removal, leading by a large margin (Second Place = 3.80 yrs; Third Place = 3.92 yrs).
Highlights
Predictive analysis with data-driven machine learning algorithms brings huge promise in neuroimaging and neuroscience research
For models pretrained in UK Biobank and finetuned in Predictive Analysis Challenge (PAC) 2019, the number of output age bins is set to 40 to reduce coding effort
In PAC 2019, we find that ADAM, it overfits more than stochastic gradient descent optimizer (SGD), performs slightly better than SGD in the validation set
Summary
Predictive analysis with data-driven machine learning algorithms brings huge promise in neuroimaging and neuroscience research. Predictive analysis can help disease diagnosis, such as Alzheimer’s [1], Autism [2], ADHD [3] and schizophrenia [4], and helps in formulating new hypotheses [5] and identifying new biomarkers [6]. The predictive analysis paradigm brings new challenges. A fair way to compare predictive analysis models is needed. It is common practise to build models in a training set, and apply the models to a test set [7, 8]. Data is usually scarce for many diseases so that training a large deep learning model in such modest datasets is still hard [11]
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
