Brain age and cognitive functioning in first-episode bipolar disorder

BACKGROUND

There is significant heterogeneity in cognitive function in patients with bipolar I disorder (BDI); however, there is a dearth of research into biological mechanisms that might underlie cognitive heterogeneity, especially at disease onset. To this end, this study investigated the association between accelerated or delayed age-related brain structural changes and cognition in early-stage BDI.

METHODS

First episode patients with BDI (n = 80) underwent cognitive assessment to yield demographically normed composite global and domain-specific scores in verbal memory, non-verbal memory, working memory, processing speed, attention, and executive functioning. Structural magnetic resonance imaging data were also collected from all participants and subjected to machine learning to compute the brain-predicted age difference (brainPAD), calculated by subtracting chronological age from age predicted by neuroimaging data (positive brainPAD values indicating age-related acceleration in brain structural changes and negative values indicating delay). Patients were divided into tertiles based on brainPAD values, and cognitive performance compared amongst tertiles with ANCOVA.

RESULTS

Patients in the lowest (delayed) tertile of brainPAD values (brainPAD range -17.9 to -6.5 years) had significantly lower global cognitive scores (p = 0.025) compared to patients in the age-congruent tertile (brainPAD range -5.3 to 2.4 yrs), and significantly lower verbal memory scores (p = 0.001) compared to the age-congruent and accelerated (brainPAD range 2.8 to 16.1 yrs) tertiles.

CONCLUSION

These results provide evidence linking cognitive dysfunction in the early stage of BDI to apparent delay in typical age-related brain changes. Further studies are required to assess how age-related brain changes and cognitive functioning evolve over time.

Predicting 'Brainage' in the Developmental Period using Structural MRI, Morphometric Similarity, and Machine Learning

Brain development is regularly studied using structural MRI. Recently, studies have used a combination of statistical learning and large-scale imaging databases of healthy-children to predict an individual's age from structural MRI. This data-driven, 'brainage' typically differs from the subjects chronological age, with this difference a potential measure of individual difference. Few studies have leveraged higher-order or connectomic representations of structural MRI data for this brainage approach. We leveraged morphometric similarity as a network-level approach to structural MRI to generate predictive models of age. We benchmarked these novel brain-age approaches using morphometric similarity against more typical, single feature (i.e. cortical thickness) approaches. We showed that these novel methods did not outperform cortical thickness or cortical volume measures. All models were significantly biased by age, but robust to motion confounds. The main results show that, whilst morphometric similarity mapping may be a novel way to leverage additional information from a T1-weighted structural MRI beyond individual features, in the context of a brain-age framework, morphometric similarity does not explain more variance than individual structural features. Morphometric similarity as a network-level approach to structural MRI may be poorly positioned to study individual differences in brain development in healthy individuals.

Towards the Interpretability of Deep Learning Models for Multi-modal Neuroimaging: Finding Structural Changes of the Ageing Brain

Brain-age (BA) estimates based on deep learning are increasingly used as neuroimaging biomarker for brain health; however, the underlying neural features have remained unclear. We combined ensembles of convolutional neural networks with Layer-wise Relevance Propagation (LRP) to detect which brain features contribute to BA. Trained on magnetic resonance imaging (MRI) data of a population-based study (n=2637, 18-82 years), our models estimated age accurately based on single and multiple modalities, regionally restricted and whole-brain images (mean absolute errors 3.37-3.86 years). We find that BA estimates capture aging at both small and large-scale changes, revealing gross enlargements of ventricles and subarachnoid spaces, as well as white matter lesions, and atrophies that appear throughout the brain. Divergence from expected aging reflected cardiovascular risk factors and accelerated aging was more pronounced in the frontal lobe. Applying LRP, our study demonstrates how superior deep learning models detect brain-aging in healthy and at-risk individuals throughout adulthood.

Advanced Brain-Age in Psychotic Psychopathology: Evidence for Transdiagnostic Neurodevelopmental Origins

Schizophrenia is characterized by abnormal brain structure such as global reductions in gray matter volume. Machine learning models trained to estimate the age of brains from structural neuroimaging data consistently show advanced brain-age to be associated with schizophrenia. Yet, it is unclear whether advanced brain-age is specific to schizophrenia compared to other psychotic disorders, and whether evidence that brain structure is "older" than chronological age actually reflects neurodevelopmental rather than atrophic processes. It is also unknown whether advanced brain-age is associated with genetic liability for psychosis carried by biological relatives of people with schizophrenia. We used the Brain-Age Regression Analysis and Computation Utility Software (BARACUS) prediction model and calculated the residualized brain-age gap of 332 adults (163 individuals with psychotic disorders: 105 schizophrenia, 17 schizoaffective disorder, 41 bipolar I disorder with psychotic features; 103 first-degree biological relatives; 66 controls). The model estimated advanced brain-ages for people with psychosis in comparison to controls and relatives, with no differences among psychotic disorders or between relatives and controls. Specifically, the model revealed an enlarged brain-age gap for schizophrenia and bipolar disorder with psychotic features. Advanced brain-age was associated with lower cognitive and general functioning in the full sample. Among relatives, cognitive performance and schizotypal symptoms were related to brain-age gap, suggesting that advanced brain-age is associated with the subtle expressions associated with psychosis. Exploratory longitudinal analyses suggested that brain aging was not accelerated in individuals with a psychotic disorder. In sum, we found that people with psychotic disorders, irrespective of specific diagnosis or illness severity, show indications of non-progressive, advanced brain-age. These findings support a transdiagnostic, neurodevelopmental formulation of structural brain abnormalities in psychotic psychopathology.

Brain-Age Prediction Using Shallow Machine Learning: Predictive Analytics Competition 2019

As we age, our brain structure changes and our cognitive capabilities decline. Although brain aging is universal, rates of brain aging differ markedly, which can be associated with pathological mechanism of psychiatric and neurological diseases. Predictive models have been applied to neuroimaging data to learn patterns associated with this variability and develop a neuroimaging biomarker of the brain condition. Aiming to stimulate the development of more accurate brain-age predictors, the Predictive Analytics Competition (PAC) 2019 provided a challenge that included a dataset of 2,640 participants. Here, we present our approach which placed between the top 10 of the challenge. We developed an ensemble of shallow machine learning methods (e.g., Support Vector Regression and Decision Tree-based regressors) that combined voxel-based and surface-based morphometric data. We used normalized brain volume maps (i.e., gray matter, white matter, or both) and features of cortical regions and anatomical structures, like cortical thickness, volume, and mean curvature. In order to fine-tune the hyperparameters of the machine learning methods, we combined the use of genetic algorithms and grid search. Our ensemble had a mean absolute error of 3.7597 years on the competition, showing the potential that shallow methods still have in predicting brain-age.