Machine learning analysis of neuroimaging data can accurately predict chronological age in healthy people. Deviations from healthy brain ageing have been associated with cognitive impairment and disease. Here we sought to further establish the credentials of 'brain-predicted age' as a biomarker of individual differences in the brain ageing process, using a predictive modelling approach based on deep learning, and specifically convolutional neural networks (CNN), and applied to both pre-processed and raw T1-weighted MRI data. Firstly, we aimed to demonstrate the accuracy of CNN brain-predicted age using a large dataset of healthy adults (N = 2001). Next, we sought to establish the heritability of brain-predicted age using a sample of monozygotic and dizygotic female twins (N = 62). Thirdly, we examined the test-retest and multi-centre reliability of brain-predicted age using two samples (within-scanner N = 20; between-scanner N = 11). CNN brain-predicted ages were generated and compared to a Gaussian Process Regression (GPR) approach, on all datasets. Input data were grey matter (GM) or white matter (WM) volumetric maps generated by Statistical Parametric Mapping (SPM) or raw data. CNN accurately predicted chronological age using GM (correlation between brain-predicted age and chronological age r = 0.96, mean absolute error [MAE] = 4.16 years) and raw (r = 0.94, MAE = 4.65 years) data. This was comparable to GPR brain-predicted age using GM data (r = 0.95, MAE = 4.66 years). Brain-predicted age was a heritable phenotype for all models and input data (h2 ≥ 0.5). Brain-predicted age showed high test-retest reliability (intraclass correlation coefficient [ICC] = 0.90-0.99). Multi-centre reliability was more variable within high ICCs for GM (0.83-0.96) and poor-moderate levels for WM and raw data (0.51-0.77). Brain-predicted age represents an accurate, highly reliable and genetically-influenced phenotype, that has potential to be used as a biomarker of brain ageing. Moreover, age predictions can be accurately generated on raw T1-MRI data, substantially reducing computation time for novel data, bringing the process closer to giving real-time information on brain health in clinical settings.
Objective: To establish whether HIV disease is associated with abnormal levels of age-related brain atrophy, by estimating apparent brain age using neuroimaging and exploring whether these estimates related to HIV status, age, cognitive performance, and HIV-related clinical parameters. Methods: A large sample of virologically suppressed HIV-positive adults (n = 162, age 45–82 years) and highly comparable HIV-negative controls (n = 105) were recruited as part of the Comorbidity in Relation to AIDS (COBRA) collaboration. Using T1-weighted MRI scans, a machine-learning model of healthy brain aging was defined in an independent cohort (n = 2,001, aged 18–90 years). Neuroimaging data from HIV-positive and HIV-negative individuals were then used to estimate brain-predicted age; then brain-predicted age difference (brain-PAD = brain-predicted brain age − chronological age) scores were calculated. Neuropsychological and clinical assessments were also carried out. Results: HIV-positive individuals had greater brain-PAD score (mean ± SD 2.15 ± 7.79 years) compared to HIV-negative individuals (−0.87 ± 8.40 years; b = 3.48, p < 0.01). Increased brain-PAD score was associated with decreased performance in multiple cognitive domains (information processing speed, executive function, memory) and general cognitive performance across all participants. Brain-PAD score was not associated with age, duration of HIV infection, or other HIV-related measures. Conclusion: Increased apparent brain aging, predicted using neuroimaging, was observed in HIV-positive adults, despite effective viral suppression. Furthermore, the magnitude of increased apparent brain aging related to cognitive deficits. However, predicted brain age difference did not correlate with chronological age or duration of HIV infection, suggesting that HIV disease may accentuate rather than accelerate brain aging.
The segmentation and volumetric quantification of white matter (WM) lesions play an important role in monitoring and studying neurological diseases such as multiple sclerosis (MS) or cerebrovascular disease. This is often interactively done using 2D magnetic resonance images. Recent developments in acquisition techniques allow for 3D imaging with much thinner sections, but the large number of images per subject makes manual lesion outlining infeasible. This warrants the need for a reliable automated approach. Here we aimed to improve k nearest neighbor (kNN) classification of WM lesions by optimizing intensity normalization and using spatial tissue type priors (TTPs). The kNN-TTP method used kNN classification with 3.0 T 3DFLAIR and 3DT1 intensities as well as MNI-normalized spatial coordinates as features. Additionally, TTPs were computed by nonlinear registration of data from healthy controls. Intensity features were normalized using variance scaling, robust range normalization or histogram matching. The algorithm was then trained and evaluated using a leave-one-out experiment among 20 patients with MS against a reference segmentation that was created completely manually. The performance of each normalization method was evaluated both with and without TTPs in the feature set. Volumetric agreement was evaluated using intra-class coefficient (ICC), and voxelwise spatial agreement was evaluated using Dice similarity index (SI). Finally, the robustness of the method across different scanners and patient populations was evaluated using an independent sample of elderly subjects with hypertension. The intensity normalization method had a large influence on the segmentation performance, with average SI values ranging from 0.66 to 0.72 when no TTPs were used. Independent of the normalization method, the inclusion of TTPs as features increased performance particularly by reducing the lesion detection error. Best performance was achieved using variance scaled intensity features and including TTPs in the feature set: this yielded ICC = 0.93 and average SI = 0.75 ± 0.08. Validation of the method in an independent sample of elderly subjects with hypertension, yielded even higher ICC = 0.96 and SI = 0.84 ± 0.14. Adding TTPs increases the performance of kNN based MS lesion segmentation methods. Best performance was achieved using variance scaling for intensity normalization and including TTPs in the feature set, showing excellent agreement with the reference segmentations across a wide range of lesion severity, irrespective of the scanner used or the pathological substrate of the lesions.
American children consume up to 27% of calories from high-fat and high-sugar snacks. Both sugar and fat consumption have been implicated as a cause of hepatic steatosis and obesity but the effect of meal pattern is largely understudied. We hypothesized that a high meal frequency, compared to consuming large meals, is detrimental in the accumulation of intrahepatic and abdominal fat. To test this hypothesis, we randomized 36 lean, healthy men to a 40% hypercaloric diet for 6 weeks or a eucaloric control diet and measured intrahepatic triglyceride content (IHTG) using proton magnetic resonance spectroscopy (1H-MRS), abdominal fat using magnetic resonance imaging (MRI), and insulin sensitivity using a hyperinsulinemic euglycemic clamp with a glucose isotope tracer before and after the diet intervention. The caloric surplus consisted of fat and sugar (high-fat-high-sugar; HFHS) or sugar only (high-sugar; HS) and was consumed together with, or between, the three main meals, thereby increasing meal size or meal frequency. All hypercaloric diets similarly increased body mass index (BMI). Increasing meal frequency significantly increased IHTG (HFHS mean relative increase of 45%; P = 0.016 and HS mean relative increase of 110%; P = 0.047), whereas increasing meal size did not (2-way analysis of variance [ANOVA] size versus frequency P = 0.03). Abdominal fat increased in the HFHS-frequency group (+63.3 ± 42.8 mL; P = 0.004) and tended to increase in the HS-frequency group (+46.5 ± 50.7 mL; P = 0.08). Hepatic insulin sensitivity tended to decrease in the HFHS-frequency group while peripheral insulin sensitivity was not affected. Conclusion: A hypercaloric diet with high meal frequency increased IHTG and abdominal fat independent of caloric content and body weight gain, whereas increasing meal size did not. This study suggests that snacking, a common feature in the Western diet, independently contributes to hepatic steatosis and obesity. (Trial registration: www.clinicaltrials.gov; nr.NCT01297738.) (Hepatology 2014;60:545–553)