A Bayesian multivariate factor analysis model for evaluating an intervention by using observational time series data on multiple outcomes.

A problem that is frequently encountered in many areas of scientific research is that of estimating the effect of a non-randomized binary intervention on an outcome of interest by using time series data on units that received the intervention ('treated') and units that did not ('controls'). One popular estimation method in this setting is based on the factor analysis (FA) model. The FA model is fitted to the preintervention outcome data on treated units and all the outcome data on control units, and the counterfactual treatment-free post-intervention outcomes of the former are predicted from the fitted model. Intervention effects are estimated as the observed outcomes minus these predicted counterfactual outcomes. We propose a model that extends the FA model for estimating intervention effects by jointly modelling the multiple outcomes to exploit shared variability, and assuming an auto-regressive structure on factors to account for temporal correlations in the outcome. Using simulation studies, we show that the method proposed can improve the precision of the intervention effect estimates and achieve better control of the type I error rate (compared with the FA model), especially when either the number of preintervention measurements or the number of control units is small. We apply our method to estimate the effect of stricter alcohol licensing policies on alcohol-related harms.

Estimating the prevalence of missing experiments in a neuroimaging meta-analysis.

Coordinate-based meta-analyses (CBMA) allow researchers to combine the results from multiple functional magnetic resonance imaging experiments with the goal of obtaining results that are more likely to generalize. However, the interpretation of CBMA findings can be impaired by the file drawer problem, a type of publication bias that refers to experiments that are carried out but are not published. Using foci per contrast count data from the BrainMap database, we propose a zero-truncated modeling approach that allows us to estimate the prevalence of nonsignificant experiments. We validate our method with simulations and real coordinate data generated from the Human Connectome Project. Application of our method to the data from BrainMap provides evidence for the existence of a file drawer effect, with the rate of missing experiments estimated as at least 6 per 100 reported. The R code that we used is available at https://osf.io/ayhfv/.

A Bayesian Approach for the Use of Athlete Performance Data Within Anti-doping.

The World Anti-doping Agency currently collates the results of all doping tests for athletes involved in elite sporting competition with the aim of improving the fight against doping. Existing anti-doping strategies involve either the direct detection of use of banned substances, or abnormal variation in metabolites or biological markers related to their use. As the aim of any doping regime is to enhance athlete competitive performance, it is interesting to consider whether performance data could be used within the fight against doping. In this regard, the identification of unexpected increases in athlete performance could be used as a trigger for their closer scrutiny via a targeted anti-doping testing programme. This study proposes a Bayesian framework for the development of an "athlete performance passport" and documents some initial findings and limitations of such an approach. The Bayesian model was retrospectively applied to the competitive results of 1,115 shot put athletes from 1975 to 2016 in order establish the interindividual variability of intraindividual performance in order to create individualized career performance trajectories for a large number of presumed clean athletes. Data from athletes convicted for doping violations (3.69% of the sample) was used to assess the predictive performance of the Bayesian framework with a probit model. Results demonstrate the ability to detect performance differences (~1 m) between doped and presumed clean athletes, and achieves good predictive performance of doping status (i.e., doped vs. non-doped) with a high area under the curve (AUC = 0.97). However, the model prediction of doping status was driven by the correct classification of presume non-doped athletes, misclassifying doped athletes as non-doped. This lack of sensitivity is likely due to the need to accommodate additional longitudinal covariates (e.g., aging and seasonality effects) potentially affecting performance into the framework. Further research is needed in order to increase the framework structure and improve its accuracy and sensitivity.

Spatial Bayesian latent factor regression modeling of coordinate-based meta-analysis data.

Now over 20 years old, functional MRI (fMRI) has a large and growing literature that is best synthesised with meta-analytic tools. As most authors do not share image data, only the peak activation coordinates (foci) reported in the article are available for Coordinate-Based Meta-Analysis (CBMA). Neuroimaging meta-analysis is used to (i) identify areas of consistent activation; and (ii) build a predictive model of task type or cognitive process for new studies (reverse inference). To simultaneously address these aims, we propose a Bayesian point process hierarchical model for CBMA. We model the foci from each study as a doubly stochastic Poisson process, where the study-specific log intensity function is characterized as a linear combination of a high-dimensional basis set. A sparse representation of the intensities is guaranteed through latent factor modeling of the basis coefficients. Within our framework, it is also possible to account for the effect of study-level covariates (meta-regression), significantly expanding the capabilities of the current neuroimaging meta-analysis methods available. We apply our methodology to synthetic data and neuroimaging meta-analysis datasets.

The coordinate-based meta-analysis of neuroimaging data.

Neuroimaging meta-analysis is an area of growing interest in statistics. The special characteristics of neuroimaging data render classical meta-analysis methods inapplicable and therefore new methods have been developed. We review existing methodologies, explaining the benefits and drawbacks of each. A demonstration on a real dataset of emotion studies is included. We discuss some still-open problems in the field to highlight the need for future research.

Bayesian latent factor regression for functional and longitudinal data.

In studies involving functional data, it is commonly of interest to model the impact of predictors on the distribution of the curves, allowing flexible effects on not only the mean curve but also the distribution about the mean. Characterizing the curve for each subject as a linear combination of a high-dimensional set of potential basis functions, we place a sparse latent factor regression model on the basis coefficients. We induce basis selection by choosing a shrinkage prior that allows many of the loadings to be close to zero. The number of latent factors is treated as unknown through a highly-efficient, adaptive-blocked Gibbs sampler. Predictors are included on the latent variables level, while allowing different predictors to impact different latent factors. This model induces a framework for functional response regression in which the distribution of the curves is allowed to change flexibly with predictors. The performance is assessed through simulation studies and the methods are applied to data on blood pressure trajectories during pregnancy.