Predicting brain age with deep learning from raw imaging data results in a reliable and heritable biomarker
Introduction
The human brain changes across the adult lifespan. This process of brain ageing occurs in accord with a general decline in cognitive performance, cognitive ageing. Although the changes associated with brain ageing are not explicitly pathological, with increasing age comes increasing risk of neurodegenerative disease and dementia (Abbott, 2011). However, the wide range of onset ages for age-associated brain diseases indicates that the effects of ageing on the brain vary greatly between individuals. Thus, advancing our understanding of brain ageing and identifying biomarkers of the process are vital to help improve detection of early-stage neurodegeneration and predict age-related cognitive decline.
One promising approach to identifying individual differences in brain ageing derives from the research showing that neuroimaging data can be used to accurately predict chronological age in healthy individuals, using machine learning (Dosenbach et al., 2010, Franke et al., 2010). By ‘learning’ the correspondence between patterns in structural or functional neuroimaging data and an age ‘label’, machine-learning algorithms can formulate massively high-dimensional regression models, fitting large neuroimaging datasets as independent variables to predict chronological age as the dependent variable. The resulting brain-based age predictions are generally highly accurate, particularly when algorithms learn from large training datasets and are applied to novel or ‘left-out’ data (i.e., test datasets).
Neuroimaging-derived age predictions have been explored in the context of different brain diseases. By training models on healthy individuals, brain-based predictions of age can then be made in independent clinical samples. If ‘brain-predicted age’ is greater than an individual's chronological age, this is thought to reflect some aberrant accumulation of age-related changes to the brain. The degree of this ‘added’ brain ageing can be simply quantified by subtracting chronological age from brain-predicted age. This approach is being used more frequently and has demonstrated increased brain-predicted age in adults with mild cognitive impairment who progress to Alzheimer's (Franke and Gaser, 2012, Gaser et al., 2013), after traumatic brain injury (Cole et al., 2015), in schizophrenia (Koutsouleris et al., 2013, Schnack et al., 2016), HIV (Cole et al., 2017c), epilepsy (Pardoe et al., 2017), Down's syndrome (Cole et al., 2017a) and diabetes (Franke et al., 2013). At the same time, brain-predicted age has been used to demonstrate protective influences on brain ageing, including meditation (Luders et al., 2016) and increased levels of education and physical exercise (Steffener et al., 2016). Evidently, the extent to which one's brain resembles the typical structure or function appropriate for one's age can be affected by both positive and negative influences. By conceptualising brain ageing in this manner, highly-complex multivariate datasets and statistical procedures can be reformulated into an intuitively straightforward and widely-applicable biomarker. However, the practicality of using such a marker clinically, its reliability and relevance for normal variation in brain ageing need to be further demonstrated.
One hindrance to clinical applications for neuroimaging generally is the time needed for image ‘post-processing’ after acquisition (referred to as ‘pre-processing’ by neuroimagers), which can take hours or days, while clinical decisions often need to occur in minutes or less. Regardless of learning algorithm, previous brain-predicted age studies have required several pre-processing stages. Such steps are typically a sequence of data transformations that produce a representation of the original images that is sufficiently structured, compact and informative to support machine learning. These include the removal of non-brain tissue (i.e., skull stripping or brain extraction), affine or non-linear image registration, interpolation and smoothing. While pre-processing may reduce noise and permit voxelwise inter-individual statistical comparisons, there are numerous additional assumptions required for any pre-processing pipeline. These assumptions are often not met, particularly when analysing brain images containing gross pathology (Avants et al., 2008, Liu et al., 2015) and can even be an increased source of error. Recently, however, modelling methods that require little or no image pre-processing have become available, including so-called ‘deep learning’.
The resurgence of interest in artificial neural networks for learning data representations, deep learning, offers a new way of approaching statistical modelling in neuroimaging, thanks to improvements in computing infrastructure. When sufficiently large volumes of data are available, no ‘hand-engineering’ (i.e., manually selecting a priori which features should be used as input) is needed as the deep learning algorithm is able to infer a compact representation of the data, starting only with raw images as input, which is optimally tailored for the particular predictive modelling task at hand. In this respect, deep learning offers several practical advantages for high-dimensional prediction tasks, that should enable the learning of both physiologically-relevant representations and latent relationships (Plis et al., 2014). Of particular interest to us is the potential for deep learning techniques, such as convolutional neural networks (CNN), to make predictions from raw, unprocessed neuroimaging data, thus obviating the reliance on time-consuming pre-processing and improving the clinical applicability of models of brain ageing.
Beyond improving clinical applicability, a biomarker of brain ageing needs to relate to naturally occurring variation, such as that caused by genetic factors. Many aspects of brain ageing and susceptibility to age-related brain disease are thought to be under genetic influence (Lee and Sachdev, 2014, Lu et al., 2004, Peters, 2006, Teter and Finch, 2004). Therefore, demonstrating a brain ageing biomarker is sensitive to genetic influences gives some external, genetic, validity to the measure. Furthermore, if a neuroimaging biomarker is heritable, this motivates further research into specific candidate genes, or sets of genes, that may affect this aspect of brain ageing. These candidate genes can then, in turn, provide biological targets for pharmacological interventions which aim to improve brain health in older adults.
Another important facet of any biomarker is reliability. If a biomarker is to be evaluated longitudinally, in clinical trials or research settings, to track change over time, establishing test-retest reliability is vital. Furthermore, as many neuroimaging studies are now international collaborative efforts, data collection often takes place across multiple scanning sites. Therefore, between-scanner reliability, which indicates that a method of obtaining a biomarker is generalizable to data acquired from other sites, is of increasing importance.
In this work, we sought to establish the credentials of CNN-predicted age as a potential biomarker of brain ageing in three different ways: 1) Demonstrate that CNNs can accurately predict age using structural neuroimaging data and compare predictions using pre-processed and ‘raw’ input data; 2) Establish the heritability of brain-predicted age using a sample of monozygotic and dizygotic twins; 3) Assess both the test-retest (i.e., within-scanner) and multi-centre (i.e., between-scanner) reliability of brain-predicted age.
Section snippets
Datasets
All neuroimaging data used in the study were T1-weighted MRI scans. Details of the participants in the specific samples and the respective acquisition parameters used are outlined below:
Convolutional neural networks accurately predict age using neuroimaging
Analysis showed that our CNN method could accurately predict the chronological age of healthy adults, using either processed volumetric maps or raw T1-MRI data (see Table 1). Prediction accuracy was similar for GPR. The lowest MAE achieved was using GM data and CNN analysis (MAE = 4.16 years), though other predictions were generally comparable. Using single tissues (i.e., GM or WM) did not appreciably alter the prediction accuracy compared to using all available input data for each subject
Discussion
Using 3D convolutional neural networks, we accurately estimated chronological age from raw T1-weighted MRI brain scans of healthy adults. The accuracy of CNN for age prediction was also high when using processed GM and WM voxelwise images, and was comparable with age estimations made using GPR. Brain-predicted age estimates were significantly heritable and showed high levels of within-scanner and between-scanner reliability. These findings support the idea that deep learning methods can
Conclusions
Deep learning models based on T1-MRI can accurately predict chronological age in healthy individuals. This can be achieved using raw MRI data, with a minimum of processing necessary to generate an accurate age prediction. These estimates of brain-predicted age are also considerably heritable, giving external, genetic, validity to the measure and motivating its use in genetic studies of brain ageing. Finally, our analysis showed the brain-predicted age is highly reliable and thus appropriate for
Acknowledgements
The TwinsUK study was funded by the Wellcome Trust, Medical Research Council, European Commision’s Seventh Framework Program (FP7/2007-2013, GA No 259749). The study also receives support from the National Institute for Health Research (NIHR), BioResource, Clinical Research Facility and Biomedical Research Centre based at Guy's and St Thomas' NHS Foundation Trust in partnership with King's College London. The STudy Of Reliability of MRI (STORM) was funded by the NIHR Biomedical Research Centre
References (68)
A fast diffeomorphic image registration algorithm
NeuroImage
(2007)- et al.
Symmetric diffeomorphic image registration with cross-correlation: evaluating automated labeling of elderly and neurodegenerative brain
Med. Image Anal.
(2008) - et al.
A comparison of methods for the automated calculation of volumes and atrophy rates in the hippocampus
NeuroImage
(2008) - et al.
The heritability of volumes of brain structures and its relationship to age: a review of twin and family studies
Ageing Res. Rev.
(2014) - et al.
Critical ages in the life course of the adult brain: nonlinear subcortical aging
Neurobiol. Aging
(2013) - et al.
Estimating the age of healthy subjects from T1-weighted MRI scans using kernel methods: exploring the influence of various parameters
NeuroImage
(2010) - et al.
The genetics of cognitive ability and cognitive ageing in healthy older people
Trends Cognitive Sci.
(2011) - et al.
A global optimisation method for robust affine registration of brain images
Med. Image Anal.
(2001) - et al.
Reliability in multi-site structural MRI studies: effects of gradient non-linearity correction on phantom and human data
NeuroImage
(2006) - et al.
Efficient multi-scale 3D CNN with fully connected CRF for accurate brain lesion segmentation
Med. Image Anal.
(2017)
Deep MRI brain extraction: a 3D convolutional neural network for skull stripping
NeuroImage
Evaluation of volume-based and surface-based brain image registration methods
NeuroImage
Neighbourhood approximation using randomized forests
Med. Image Anal.
Genetic and environmental influences on the size of specific brain regions in midlife: the VETSA MRI study
NeuroImage
Genetic influences on cognitive functions in the elderly: a selective review of twin studies
Brain Res. Rev.
Estimating brain age using high-resolution pattern recognition: younger brains in long-term meditation practitioners
NeuroImage
Prediction of individual subject's age across the human lifespan using diffusion tensor imaging: a machine learning approach
NeuroImage
Structural brain changes in medically refractory focal epilepsy resemble premature brain aging
Epilepsy Res.
Motion and morphometry in clinical and nonclinical populations
NeuroImage
Differences between chronological and brain age are related to education and self-reported physical activity
Neurobiol. Aging
Caliban's heritance and the genetics of neuronal aging
Trends Neurosci.
Cortical thickness or grey matter volume? The importance of selecting the phenotype for imaging genetics studies
NeuroImage
Dementia: a problem for our age
Nature
A new look at the statistical model identification
IEEE Trans. Automat. Control
Quantitative genetic modeling of variation in human brain morphology
Cereb. Cortex
Heritability of brain volumes in older adults: the older Australian twins study
Neurobiol. Aging
OpenMx: an open source extended structural equation modeling framework
Psychometrika
The Wilson Effect: the increase in heritability of IQ with age
Twin Res. Hum. Genet.
Big data deep learning: challenges and perspectives
IEEE Access
Brain-predicted age in Down Syndrome is associated with β-amyloid deposition and cognitive decline
Neurobiol. Aging
Prediction of brain age suggests accelerated atrophy after traumatic brain injury
Ann. Neurol.
Brain age predicts mortality
Mol. Psychiatry
Increased brain-predicted aging in treated HIV disease
Neurology
Deep learning: methods and applications
Found. Trends Signal Process.
Cited by (579)
Unveiling the muscle-brain axis: A bidirectional mendelian randomization study investigating the causal relationship between sarcopenia-related traits and brain aging
2024, Archives of Gerontology and GeriatricsExtensive T1-weighted MRI preprocessing improves generalizability of deep brain age prediction models
2024, Computers in Biology and MedicineApplication of AI in biological age prediction
2024, Current Opinion in Structural BiologyDeep Learning and Geriatric Mental Health
2024, American Journal of Geriatric Psychiatry