Brain morphology varies across the ageing trajectory and the prediction of a person's age using brain features can aid the detection of abnormalities in the ageing process.Existing studies on such "brain age prediction" vary widely in terms of their methods and type of data, so at present the most accurate and generalisable methodological approach is unclear. Therefore, we used the UK Biobank data set (N = 10,824, age range 47-73) to compare the performance of the machine learning models support vector regression, relevance vector regression and Gaussian process regression on whole-brain region-based or voxel-based structural magnetic resonance imaging data with or without dimensionality reduction through principal component analysis. Performance was assessed in the validation set through cross-validation as well as an independent test set. The models achieved mean absolute errors between 3.7 and 4.7 years, with those trained on voxel-level data with principal component analysis performing best. Overall, we observed little difference in performance between models trained on the same data type, indicating that the type of input data had greater impact on performance than model choice. All code is provided online in the hope that this will aid future research.

Normative modelling is an emerging method for quantifying how individuals deviate from the healthy populational pattern. Several machine learning models have been implemented to develop normative models to investigate brain disorders, including regression, support vector machines and Gaussian process models. With the advance of deep learning technology, the use of deep neural networks has also been proposed. In this study, we assessed normative models based on deep autoencoders using structural neuroimaging data from patients with Alzheimer’s disease (n = 206) and mild cognitive impairment (n = 354). We first trained the autoencoder on an independent dataset (UK Biobank dataset) with 11,034 healthy controls. Then, we estimated how each patient deviated from this norm and established which brain regions were associated to this deviation. Finally, we compared the performance of our normative model against traditional classifiers. As expected, we found that patients exhibited deviations according to the severity of their clinical condition. The model identified medial temporal regions, including the hippocampus, and the ventricular system as critical regions for the calculation of the deviation score. Overall, the normative model had comparable cross-cohort generalizability to traditional classifiers. To promote open science, we are making all scripts and the trained models available to the wider research community.

For most neuroimaging questions the range of possible analytic choices makes it unclear how to evaluate conclusions from any single analytic method. One possible way to address this issue is to evaluate all possible analyses using a multiverse approach, however, this can be computationally challenging and sequential analyses on the same data can compromise predictive power. Here, we establish how active learning on a low-dimensional space capturing the inter-relationships between pipelines can efficiently approximate the full spectrum of analyses. This approach balances the benefits of a multiverse analysis without incurring the cost on computational and predictive power. We illustrate this approach with two functional MRI datasets (predicting brain age and autism diagnosis) demonstrating how a multiverse of analyses can be efficiently navigated and mapped out using active learning. Furthermore, our presented approach not only identifies the subset of analysis techniques that are best able to predict age or classify individuals with autism spectrum disorder and healthy controls, but it also allows the relationships between analyses to be quantified.

Deep neural networks have brought remarkable breakthroughs in medical image analysis. However, due to their data-hungry nature, the modest dataset sizes in medical imaging projects might be hindering their full potential. Generating synthetic data provides a promising alternative, allowing to complement training datasets and conducting medical image research at a larger scale. Diffusion models recently have caught the attention of the computer vision community by producing photorealistic synthetic images. In this study, we explore using Latent Diffusion Models to generate synthetic images from high-resolution 3D brain images. We used T1w MRI images from the UK Biobank dataset (N=31,740) to train our models to learn about the probabilistic distribution of brain images, conditioned on covariables, such as age, sex, and brain structure volumes. We found that our models created realistic data, and we could use the conditioning variables to control the data generation effectively. Besides that, we created a synthetic dataset with 100,000 brain images and made it openly available to the scientific community.

Vincent van Gogh was one of the most influential artists of the Western world, having shaped the post-impressionist art movement by shifting its boundaries forward into abstract expressionism. His distinctive style, which was not valued by the art-buying public during his lifetime, is nowadays one of the most sought after. However, despite the great deal of attention from academic and artistic circles, one important question remains open: was van Gogh's original style a visual manifestation distinct from his troubled mind, or was it in fact a by-product of an impairment that resulted from the psychiatric illness that marred his entire life? In this paper, we use a previously published multi-scale model of brain function to piece together a number of disparate observations about van Gogh's life and art. In particular, we first quantitatively analyze the brushwork of his large production of self-portraits using the image autocorrelation and demonstrate a strong association between the contrasts in the paintings, the occurrence of psychiatric symptoms, and his simultaneous use of absinthe-a strong liquor known to affect gamma aminobutyric acid (GABA) alpha receptors. Secondly, we propose that van Gogh suffered from a defective function of parvalbumin interneurons, which seems likely given his family history of schizophrenia and his addiction to substances associated with GABA action. This could explain the need for the artist to increasingly amplify the contrasts in his brushwork as his disease progressed, as well as his tendency to merge esthetic and personal experiences into a new form of abstraction.

For most neuroimaging questions the huge range of possible analytic choices leads to the possibility that conclusions from any single analytic approach may be misleading. Examples of possible choices include the motion regression approach used and smoothing and threshold factors applied during the processing pipeline. Although it is possible to perform a multiverse analysis that evaluates all possible analytic choices, this can be computationally challenging and repeated sequential analyses on the same data can compromise inferential and predictive power. Here, we establish how active learning on a low-dimensional space that captures the inter-relationships between analysis approaches can be used to efficiently approximate the whole multiverse of analyses. This approach balances the benefits of a multiverse analysis without the accompanying cost to statistical power, computational power and the integrity of inferences. We illustrate this approach with a functional MRI dataset of functional connectivity across adolescence, demonstrating how a multiverse of graph theoretic and simple pre-processing steps can be efficiently navigated using active learning. Our study shows how this approach can identify the subset of analysis techniques (i.e., pipelines) which are best able to predict participants' ages, as well as allowing the performance of different approaches to be quantified.

The field of infant research is not immune from the reproducibility crisis in cognitive science and psychology. In their recent methodological article, Byers‐Heinlein et al. (2021) invited infant researchers to commit to produce robust findings by reporting reliability metrics for their variables of interest, improving data quality and quantity, and moving towards more sophisticated paradigms and analyses. We present a novel artificial intelligence‐enriched individualized approach that, in our view, is particularly promising to shed new light on infant and child development and promote good research practice in the field; neuroadaptive Bayesian optimization (NBO). NBO is a transformative method where the collected brain or behavioural data are processed in real time and used to identify the stimuli that maximize the individual's response. Applying NBO to infant research goes in the direction proposed by Byers‐Heinlein et al. (2021) and further, the method requires careful a priori choices that effectively correspond to preregistering the experimental design and analytic pipeline. In this commentary, we examine how the NBO approach embeds the six proposed solutions for more reliable infant research, encouraging transparency of the planned analyses and robustness of findings.

Normative modelling is an emerging method for quantifying how individuals deviate from the healthy populational pattern. Several machine learning models have been implemented to develop normative models to investigate brain disorders, including regression, support vector machines and Gaussian process models. With the advance of deep learning technology, the use of deep neural networks has also been proposed. In this study, we assessed normative models based on deep autoencoders using structural neuroimaging data from patients with Alzheimer's disease (n=206) and mild cognitive impairment (n=354). We first trained the autoencoder on an independent dataset (UK Biobank dataset) with 11,034 healthy controls. Then, we estimated how each patient deviated from this norm and established which brain regions were associated to this deviation. Finally, we compared the performance of our normative model against traditional classifiers. As expected, we found that patients exhibited deviations according to the severity of their clinical condition. The model identified medial temporal regions, including the hippocampus, and the ventricular system as critical regions for the calculation of the deviation score. Overall, the normative model had comparable cross-cohort generalizability to traditional classifiers. In order to promote open science, we are making all scripts and the trained models available to the wider research community.