Brain aging comprises many modes of structural and functional change with distinct genetic and biophysical associations

Smith, Stephen M.; Elliott, Lloyd T.; Alfaro-Almagro, Fidel; McCarthy, Paul; Nichols, Thomas E.; Douaud, Gwenaëlle; Miller, Karla L.

doi:10.7554/elife.52677

Cited by 147 publications

(106 citation statements)

References 57 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1. In line with existing findings about sex differences such as post-menopausal accelerated brain ageing [40], stratification by sex was also conducted in this study. The images were acquired at 1mm isotropic resolution using a 3D MPRAGE Sequence [18] and processed using the UK Biobank pipeline [41].…”

Section: Data and Preprocessingmentioning

confidence: 75%

“…Through utilising a CNN we are able to create a model that has freedom to investigate spatial signatures, rather than relying on spatial priors or target structures, in contrast to other methods proposed to explore brain age in the UK Biobank. In both [39] and [40] image-derived phenotypes are used to train a linear model to predict brain age, with the former using the IDPs to create a single estimate of brain age and in the latter they explore whether multiple modalities can produce distinct modes that relate to different processes involved in ageing and disease. These methods allow us to explore different aspects of ageing using multimodal data but they require prior choice of the IDPs whereas our method is able to consider the whole of the input image and therefore there is potential for our method to pick up on different types of changes compared to the IDP-based models, especially those with subtle changes.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Learning Patterns of the Ageing Brain in MRI using Deep Convolutional Networks

Dinsdale

Bluemke

Smith

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

Both normal ageing and neurodegenerative diseases cause morphological changes to the brain. Age-related brain changes are subtle, nonlinear, and spatially and temporally heterogenous, both within a subject and across a population. Machine learning models are particularly suited to capture these patterns and can produce a model that is sensitive to changes of interest, despite the large variety in healthy brain appearance. In this paper, the power of convolutional neural networks (CNNs) and the rich UK Biobank dataset, the largest database currently available, are harnessed to address the problem of predicting brain age. We developed a 3D CNN architecture to predict chronological age, using a training dataset of 12,802 T1-weighted MRI images and a further 6,885 images for testing. The proposed method shows competitive performance on age prediction, but, most importantly, the CNN prediction errors ΔBrain Age = AgePredicted - AgeTrue correlated significantly with many clinical measurements from the UK Biobank in the female and male groups. In addition, having used images from only one imaging modality in this experiment, we examined the relationship between ΔBrain Age and the image-derived phenotypes (IDPs) from all other imaging modalities in the UK Biobank, showing correlations consistent with known patterns of ageing. Furthermore, we show that the use of nonlinearly registered images to train CNNs can lead to the network being driven by artefacts of the registration process and missing subtle indicators of ageing, limiting the clinical relevance. Due to the longitudinal aspect of the UK Biobank study, in the future it will be possible to explore whether the ΔBrain Age from models such as this network were predictive of any health outcomes.

show abstract

Section: Data and Preprocessingmentioning

confidence: 75%

Section: Introductionmentioning

confidence: 99%

Learning Patterns of the Ageing Brain in MRI using Deep Convolutional Networks

Dinsdale

Bluemke

Smith

et al. 2020

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…With about 5000 training subjects, SFCN achieves 2.28 years MAE while 3D-ResNet with tensor regression achieves 2.58 years (Kolbeinsson et al, 2019). For the larger training set with more than 10,000 subjects, linear regression with multimodality IDPs (including fMRI and DTI features) achieves an MAE of 2.9 years (Smith et al, 2020(Smith et al, , 2019, whereas SFCN obtains the best MAE in UK Biobank with an MAE of 2.14 years.…”

Section: The Performance Of Sfcn In Uk Biobank Datamentioning

confidence: 99%

“…Our presented strategy achieves state-of-the-art results in brain age prediction. Table 3 shows a summary of previously reported brain age prediction MAE results (Kolbeinsson et al, 2019;Ning et al, 2018;Smith et al, 2020). To eliminate the effect of sample size differences (i.e., to make these comparisons as meaningful as possible), we trained SFCN with comparable training set sizes as the previous studies, and compared performance with those.…”

Section: The Performance Of Sfcn In Uk Biobank Datamentioning

confidence: 99%

Accurate brain age prediction with lightweight deep neural networks

Han

Gong

Beckmann

et al. 2019

Preprint

View full text Add to dashboard Cite

These authors contributed equally † Corresponding to Han Peng, han.peng@ndcn.ox.ac.uk Highlights • A lightweight deep learning model, Simple Fully Convolutional Network (SFCN), ispresented, achieving state-of-the-art brain age prediction performance in UK Biobank MRI brain imaging data. • Even with limited number of training subjects (e.g., 50), SFCN performs better than widely-used regression models. • A semi-multimodal ensemble strategy is proposed and achieved first place in the PAC 2019 brain age prediction challenge. • Linear regression can remove brain age predication bias (even on unlabelled data) while maintaining state-of-the-art performance. AbstractDeep learning has huge potential for accurate disease prediction with neuroimaging data, but the prediction performance is often limited by training-dataset size and compute memory requirements. To address this, we propose a deep convolutional neural network model, Simple Fully Convolutional Network (SFCN), for accurate prediction of brain age using T1weighted structural MRI data. Compared with other popular deep network architectures, SFCN has fewer parameters, so is more compatible with small dataset size and 3D volume data. The network architecture was combined with several techniques for boosting performance, including data augmentation, pre-training, model regularization, model ensemble and prediction bias correction. We compared our overall SFCN approach with several widely-used machine learning models. It achieved state-of-the-art performance in UK Biobank data (N = 14,503), with mean absolute error (MAE) = 2.14y in brain age prediction and 99.3% in sex classification. SFCN also won (both parts of) the 2019 Predictive Analysis Challenge for brain age prediction, involving 79 competing teams (N = 2,638, MAE = 2.90y). We describe here the details of our approach, and its optimisation and validation. Our approach can easily be generalised to other tasks using different image modalities, and is released on GitHub.

show abstract

“…Studies showed that individuals' chronological age can be predicted accurately from brain MRI scans . Brain age delta, the difference of a subject's predicted (brain) age and chronological age, is linked with a variety of biological factors within the healthy population (Smith et al, 2020b), and group differences can be found in disease populations (Cole et al, 2019;Kaufmann et al, 2019). Yet, accurate prediction of a subject's age in healthy population is still a challenging task.…”

Section: Introductionmentioning

confidence: 99%

Optimising a Simple Fully Convolutional Network (SFCN) for accurate brain age prediction in the PAC 2019 challenge

Gong

Beckmann

Vedaldi

et al. 2020

Preprint

View full text Add to dashboard Cite

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 places in both of the two objectives in the PAC 2019 brain age prediction challenge: Mean absolute error (MAE) = 2.90 years without bias removal, and MAE = 2.95 years with bias removal.

show abstract

Brain aging comprises many modes of structural and functional change with distinct genetic and biophysical associations

Cited by 147 publications

References 57 publications

Learning Patterns of the Ageing Brain in MRI using Deep Convolutional Networks

Learning Patterns of the Ageing Brain in MRI using Deep Convolutional Networks

Accurate brain age prediction with lightweight deep neural networks

Optimising a Simple Fully Convolutional Network (SFCN) for accurate brain age prediction in the PAC 2019 challenge

Contact Info

Product

Resources

About