Unveiling causal connections: Long-term particulate matter exposure and type 2 diabetes mellitus mortality in Southern China

Evidence of the potential causal links between long-term exposure to particulate matters (PM, i.e., PM1, PM2.5, and PM1-2.5) and T2DM mortality based on large cohorts is limited. In contrast, the existing evidence usually suffers from inherent bias with the traditional association assessment. A prospective cohort of 580,757 participants in the southern region of China were recruited during 2009 and 2015 and followed up through December 2020. PM exposure at each residential address was estimated by linking to the well-established high-resolution simulation dataset. Hazard ratios (HRs) were calculated using time-varying marginal structural Cox models, an established causal inference approach, after adjusting for potential confounders. During follow-up, a total of 717 subjects died from T2DM. For every 1 μg/m3 increase in PM2.5, the adjusted HRs and 95% confidence interval (CI) for T2DM mortality was 1.036 (1.019-1.053). Similarly, for every 1 μg/m3 increase in PM1 and PM1-2.5, the adjusted HRs and 95% CIs were 1.032 (1.003-1.062) and 1.085 (1.054-1.116), respectively. Additionally, we observed a generally more pronounced impact among individuals with lower levels of education or lower residential greenness which as measured by the Normalized Difference Vegetation Index (NDVI). We identified substantial interactions between NDVI and PM1 (P-interaction = 0.003), NDVI and PM2.5 (P-interaction = 0.019), as well as education levels and PM1 (P-interaction = 0.049). The study emphasizes the need to consider environmental and socio-economic factors in strategies to reduce T2DM mortality. We found that PM1, PM2.5, and PM1-2.5 heighten the peril of T2DM mortality, with education and green space exposure roles in modifying it.

Potential causal links of long-term air pollution with lung cancer incidence: From the perspectives of mortality and hospital admission in a large cohort study in southern China

Evidence on the potential causal links of long-term air pollution exposure with lung cancer incidence (reflected by mortality and hospital admission) was limited, especially based on large cohorts. We examined the relationship between lung cancer and long-term exposure to particulate matter (PM, including PM2.5 , PM10 and PM10-2.5 ) and nitrogen dioxide (NO2 ) among a large cohort of general Chinese adults using causal inference approaches. The study included 575 592 participants who were followed up for an average of 8.2 years. The yearly exposure of PM and NO2 was estimated through satellite-based random forest approaches and the ordinary kriging method, respectively. Marginal structural Cox models were used to examine hazard ratios (HRs) of mortality and hospital admission due to lung cancer following air pollution exposure, adjusting for potential confounders. The HRs of mortality due to lung cancer were 1.042 (95% confidence interval [CI]: 1.033-1.052), 1.032 (95% CI:1.024-1.041) and 1.052 (95% CI:1.041-1.063) for each 1 μg/m3 increase in PM2.5 , PM10 and NO2 , respectively. In addition, we observed statistically significant effects of PMs on hospital admission due to lung cancer. The HRs (95%CI) were 1.110 (1.027-1.201), 1.067 (1.020-1.115) and 1.079 (1.010-1.153) for every 1 μg/m3 increase in PM2.5 , PM10 , PM10-2.5 , respectively. Furthermore, we found larger effect estimates among the elderly and those who exercised more frequently. We provided the most comprehensive evidence of the potential causal links between two outcomes of lung cancer and long-term air pollution exposure. Relevant policies should be developed, with special attention to protecting the vulnerable groups of the population.

Causal discovery analysis: A promising tool in advancing precision medicine for eating disorders

OBJECTIVE

Precision medicine (i.e., individually tailored treatments) represents an optimal goal for treating complex psychiatric disorders, including eating disorders. Within the eating disorders field, most treatment development efforts have been limited in their ability to identify individual-level models of eating disorder psychopathology and to develop and apply an individually tailored treatment for a given individual's personalized model of psychopathology. In addition, research is still needed to identify causal relationships within a given individual's model of eating disorder psychopathology. Addressing this limitation of the current state of precision medicine-related research in the field will allow us to progress toward advancing research and practice for eating disorders treatment.

METHOD

We present a novel set of analytic tools, causal discovery analysis (CDA) methods, which can facilitate increasingly fine-grained, person-specific models of causal relations among cognitive, behavioral, and affective symptoms.

RESULTS

CDA can advance the identification of an individual's causal model that maintains that individuals' eating disorder psychopathology.

DISCUSSION

In the current article, we (1) introduce CDA methods as a set of promising analytic tools for developing precision medicine methods for eating disorders including the potential strengths and weaknesses of CDA, (2) provide recommendations for future studies utilizing this approach, and (3) outline the potential clinical implications of using CDA to generate personalized models of eating disorder psychopathology.

PUBLIC SIGNIFICANCE STATEMENT

CDA provides a novel statistical approach for identifying causal relationships among variables of interest for a given individual. Person-specific causal models may offer a promising approach to individualized treatment planning and inform future personalized treatment development efforts for eating disorders.

Modeling the impact of calorie-reduction interventions on population prevalence and inequalities in childhood obesity in the Southampton Women's Survey

BACKGROUND

In the United Kingdom, rates of childhood obesity are high and inequalities in obesity have widened in recent years. Children with obesity face heightened risks of living with obesity as adults and suffering from associated morbidities. Addressing population prevalence and inequalities in childhood obesity is a key priority for public health policymakers in the United Kingdom and elsewhere. Where randomized controlled trials are not possible, potential policy actions can be simulated using causal modeling techniques.

OBJECTIVES

Using data from the Southampton Women's Survey (SWS), a cohort with high quality dietary and lifestyle data, the potential impact of policy-relevant calorie-reduction interventions on population prevalence and inequalities of childhood obesity was investigated.

METHODS

Predicted probabilities of obesity (using UK90 cut-offs) at age 6-7 years were estimated from logistic marginal structural models adjusting for observed calorie consumption at age 3 years (using food diaries) and confounding. A series of policy-relevant intervention scenarios were modeled to simulate reductions in energy intake (differing in effectiveness, the targeting mechanisms, and level of uptake).

RESULTS

At age 6-7 years, 8.3% of children were living with obesity, after accounting for observed energy intake and confounding. A universal intervention to lower median energy intake to the estimated average requirement (a 13% decrease), with an uptake of 75%, reduced obesity prevalence by 1% but relative and absolute inequalities remained broadly unchanged.

CONCLUSIONS

Simulated interventions substantially reduced population prevalence of obesity, which may be useful in informing policymakers.

Potential mechanisms of the fatigue-reducing effect of cognitive-behavioral therapy in cancer survivors: Three randomized controlled trials

OBJECTIVE

Fatigue is a common symptom among cancer survivors that can be successfully treated with cognitive-behavioral therapy (CBT). Insights into the working mechanisms of CBT are currently limited. The aim of this study was to investigate whether improvements in targeted cognitive-behavioral variables and reduced depressive symptoms mediate the fatigue-reducing effect of CBT.

METHODS

We pooled data from three randomized controlled trials that tested the efficacy of CBT to reduce severe fatigue. In all three trials, fatigue severity (checklist individual strength) decreased significantly following CBT. Assessments were conducted pre-treatment and 6 months later. Classical mediation analysis testing a pre-specified model was conducted and its results compared to those of causal discovery, an explorative data-driven approach testing all possible causal associations and retaining the most likely model.

RESULTS

Data from 250 cancer survivors (n = 129 CBT, n = 121 waitlist) were analyzed. Classical mediation analysis suggests that increased self-efficacy and decreased fatigue catastrophizing, focusing on symptoms, perceived problems with activity and depressive symptoms mediate the reduction of fatigue brought by CBT. Conversely, causal discovery and post-hoc analyses indicate that fatigue acts as mediator, not outcome, of changes in cognitions, sleep disturbance and depressive symptoms.

CONCLUSIONS

Cognitions, sleep disturbance and depressive symptoms improve during CBT. When assessed pre- and post-treatment, fatigue acts as a mediator, not outcome, of these improvements. It seems likely that the working mechanism of CBT is not a one-way causal effect but a dynamic reciprocal process. Trials integrating intermittent assessments are needed to shed light on these mechanisms and inform optimization of CBT.

Causal Datasheet for Datasets: An Evaluation Guide for Real-World Data Analysis and Data Collection Design Using Bayesian Networks

Developing data-driven solutions that address real-world problems requires understanding of these problems' causes and how their interaction affects the outcome-often with only observational data. Causal Bayesian Networks (BN) have been proposed as a powerful method for discovering and representing the causal relationships from observational data as a Directed Acyclic Graph (DAG). BNs could be especially useful for research in global health in Lower and Middle Income Countries, where there is an increasing abundance of observational data that could be harnessed for policy making, program evaluation, and intervention design. However, BNs have not been widely adopted by global health professionals, and in real-world applications, confidence in the results of BNs generally remains inadequate. This is partially due to the inability to validate against some ground truth, as the true DAG is not available. This is especially problematic if a learned DAG conflicts with pre-existing domain doctrine. Here we conceptualize and demonstrate an idea of a "Causal Datasheet" that could approximate and document BN performance expectations for a given dataset, aiming to provide confidence and sample size requirements to practitioners. To generate results for such a Causal Datasheet, a tool was developed which can generate synthetic Bayesian networks and their associated synthetic datasets to mimic real-world datasets. The results given by well-known structure learning algorithms and a novel implementation of the OrderMCMC method using the Quotient Normalized Maximum Likelihood score were recorded. These results were used to populate the Causal Datasheet, and recommendations could be made dependent on whether expected performance met user-defined thresholds. We present our experience in the creation of Causal Datasheets to aid analysis decisions at different stages of the research process. First, one was deployed to help determine the appropriate sample size of a planned study of sexual and reproductive health in Madhya Pradesh, India. Second, a datasheet was created to estimate the performance of an existing maternal health survey we conducted in Uttar Pradesh, India. Third, we validated generated performance estimates and investigated current limitations on the well-known ALARM dataset. Our experience demonstrates the utility of the Causal Datasheet, which can help global health practitioners gain more confidence when applying BNs.

How haematopoiesis research became a fertile ground for regulatory network biology as pioneered by Eric Davidson

PURPOSE OF REVIEW

This historical perspective reviews how work of Eric H. Davidson was a catalyst and exemplar for explaining haematopoietic cell fate determination through gene regulation.

RECENT FINDINGS

Researchers studying blood and immune cells pioneered many of the early mechanistic investigations of mammalian gene regulatory processes. These efforts included the characterization of complex gene regulatory sequences exemplified by the globin and T-cell/B-cell receptor gene loci, as well as the identification of many key regulatory transcription factors through the fine mapping of chromosome translocation breakpoints in leukaemia patients. As the repertoire of known regulators expanded, assembly into gene regulatory network models became increasingly important, not only to account for the truism that regulatory genes do not function in isolation but also to devise new ways of extracting biologically meaningful insights from even more complex information. Here we explore how Eric H. Davidson's pioneering studies of gene regulatory network control in nonvertebrate model organisms have had an important and lasting impact on research into blood and immune cell development.

SUMMARY

The intellectual framework developed by Davidson continues to contribute to haematopoietic research, and his insistence on demonstrating logic and causality still challenges the frontier of research today.

The differential genetic signatures related to climatic landscapes for jaguars and pumas on a continental scale

Modern and paleoclimate changes may have altered species dynamics by shifting species' niche suitability over space and time. We analyze whether the current genetic structure and isolation of the two large American felids, jaguar (Panthera onca) and puma (Puma concolor), are mediated by changes in climatic suitability and connection routes over modern and paleoclimatic landscapes. We estimate species distribution under 5 climatic landscapes (modern, Holocene, last maximum glaciations [LMG], average suitability, and climatic instability) and correlate them with individuals' genetic isolation through causal modeling on a resemblance matrix. Both species exhibit genetic isolation patterns correlated with LMG climatic suitability, suggesting that these areas may have worked as "allele refuges." However, the jaguar showed higher vulnerability to climate changes, responding to modern climatic suitability and connection routes, whereas the puma showed a continuous and gradual transition of genetic variation. Despite differential responsiveness to climate change, both species are subjected to the climatic effects on genetic configuration, which may make them susceptible to future climatic changes, since these are progressing faster and with higher intensity than changes in the paleoclimate. Thus, the effects of climatic changes should be considered in the design of conservation strategies to ensure evolutionary and demographic processes mediated by gene flow for both species.

Estimation of Controlled Direct Effects in Longitudinal Mediation Analyses with Latent Variables in Randomized Studies

In a randomized study with longitudinal data on a mediator and outcome, estimating the direct effect of treatment on the outcome at a particular time requires adjusting for confounding of the association between the outcome and all preceding instances of the mediator. When the confounders are themselves affected by treatment, standard regression adjustment is prone to severe bias. In contrast, G-estimation requires less stringent assumptions than path analysis using SEM to unbiasedly estimate the direct effect even in linear settings. In this article, we propose a G-estimation method to estimate the controlled direct effect of treatment on the outcome, by adapting existing G-estimation methods for time-varying treatments without mediators. The proposed method can accommodate continuous and noncontinuous mediators, and requires no models for the confounders. Unbiased estimation only requires correctly specifying a mean model for either the mediator or the outcome. The method is further extended to settings where the mediator or outcome, or both, are latent, and generalizes existing methods for single measurement occasions of the mediator and outcome to longitudinal data on the mediator and outcome. The methods are utilized to assess the effects of an intervention on physical activity that is possibly mediated by motivation to exercise in a randomized study.

'Turning the tide' on hyperglycemia in pregnancy: insights from multiscale dynamic simulation modeling

INTRODUCTION

Hyperglycemia in pregnancy (HIP, including gestational diabetes and pre-existing type 1 and type 2 diabetes) is increasing, with associated risks to the health of women and their babies. Strategies to manage and prevent this condition are contested. Dynamic simulation models (DSM) can test policy and program scenarios before implementation in the real world. This paper reports the development and use of an advanced DSM exploring the impact of maternal weight status interventions on incidence of HIP.

METHODS

A consortium of experts collaboratively developed a hybrid DSM of HIP, comprising system dynamics, agent-based and discrete event model components. The structure and parameterization drew on a range of evidence and data sources. Scenarios comparing population-level and targeted prevention interventions were simulated from 2018 to identify the intervention combination that would deliver the greatest impact.

RESULTS

Population interventions promoting weight loss in early adulthood were found to be effective, reducing the population incidence of HIP by 17.3% by 2030 (baseline ('business as usual' scenario)=16.1%, 95% CI 15.8 to 16.4; population intervention=13.3%, 95% CI 13.0 to 13.6), more than targeted prepregnancy (5.2% reduction; incidence=15.3%, 95% CI 15.0 to 15.6) and interpregnancy (4.2% reduction; incidence=15.5%, 95% CI 15.2 to 15.8) interventions. Combining targeted interventions for high-risk groups with population interventions promoting healthy weight was most effective in reducing HIP incidence (28.8% reduction by 2030; incidence=11.5, 95% CI 11.2 to 11.8). Scenarios exploring the effect of childhood weight status on entry to adulthood demonstrated significant impact in the selected outcome measure for glycemic regulation, insulin sensitivity in the short term and HIP in the long term.

DISCUSSION

Population-level weight reduction interventions will be necessary to 'turn the tide' on HIP. Weight reduction interventions targeting high-risk individuals, while beneficial for those individuals, did not significantly impact forecasted HIP incidence rates. The importance of maintaining interventions promoting healthy weight in childhood was demonstrated.

A Bayesian Approach to the Analysis of Local Average Treatment Effect for Missing and Non-normal Data in Causal Modeling: A Tutorial With the ALMOND Package in R

One practical challenge in observational studies and quasi-experimental designs is selection bias. The issue of selection bias becomes more concerning when data are non-normal and contain missing values. Recently, a Bayesian robust two-stage causal modeling with instrumental variables was developed and has the advantages of addressing selection bias and handle non-normal data and missing data simultaneously in one model. The method provides reliable parameter and standard error estimates when missing data and outliers exist. The modeling technique can be widely applied to empirical studies particularly in social, psychological and behavioral areas where any of the three issues (e.g., selection bias, data with outliers and missing data) is commonly seen. To implement this method, we developed an R package named ALMOND (Analysis of LATE (Local Average Treatment Effect) for Missing Or/and Nonnormal Data). Package users have the flexibility to directly apply the Bayesian robust two-stage causal models or write their own Bayesian models from scratch within the package. To facilitate the application of the Bayesian robust two-stage causal modeling technique, we provide a tutorial for the ALMOND package in this article, and illustrate the application with two examples from empirical research.

Applying causal models to explore the mechanism of action of simvastatin in progressive multiple sclerosis

Understanding the mode of action of drugs is a challenge with conventional methods in clinical trials. Here, we aimed to explore whether simvastatin effects on brain atrophy and disability in secondary progressive multiple sclerosis (SPMS) are mediated by reducing cholesterol or are independent of cholesterol. We applied structural equation models to the MS-STAT trial in which 140 patients with SPMS were randomized to receive placebo or simvastatin. At baseline, after 1 and 2 years, patients underwent brain magnetic resonance imaging; their cognitive and physical disability were assessed on the block design test and Expanded Disability Status Scale (EDSS), and serum total cholesterol levels were measured. We calculated the percentage brain volume change (brain atrophy). We compared two models to select the most likely one: a cholesterol-dependent model with a cholesterol-independent model. The cholesterol-independent model was the most likely option. When we deconstructed the total treatment effect into indirect effects, which were mediated by brain atrophy, and direct effects, simvastatin had a direct effect (independent of serum cholesterol) on both the EDSS, which explained 69% of the overall treatment effect on EDSS, and brain atrophy, which, in turn, was responsible for 31% of the total treatment effect on EDSS [β = -0.037; 95% credible interval (CI) = -0.075, -0.010]. This suggests that simvastatin's beneficial effects in MS are independent of its effect on lowering peripheral cholesterol levels, implicating a role for upstream intermediate metabolites of the cholesterol synthesis pathway. Importantly, it demonstrates that computational models can elucidate the causal architecture underlying treatment effects in clinical trials of progressive MS.

Causality on longitudinal data: Stable specification search in constrained structural equation modeling

A typical problem in causal modeling is the instability of model structure learning, i.e., small changes in finite data can result in completely different optimal models. The present work introduces a novel causal modeling algorithm for longitudinal data, that is robust for finite samples based on recent advances in stability selection using subsampling and selection algorithms. Our approach uses exploratory search but allows incorporation of prior knowledge, e.g., the absence of a particular causal relationship between two specific variables. We represent causal relationships using structural equation models. Models are scored along two objectives: the model fit and the model complexity. Since both objectives are often conflicting, we apply a multi-objective evolutionary algorithm to search for Pareto optimal models. To handle the instability of small finite data samples, we repeatedly subsample the data and select those substructures (from the optimal models) that are both stable and parsimonious. These substructures can be visualized through a causal graph. Our more exploratory approach achieves at least comparable performance as, but often a significant improvement over state-of-the-art alternative approaches on a simulated data set with a known ground truth. We also present the results of our method on three real-world longitudinal data sets on chronic fatigue syndrome, Alzheimer disease, and chronic kidney disease. The findings obtained with our approach are generally in line with results from more hypothesis-driven analyses in earlier studies and suggest some novel relationships that deserve further research.

Inferring Causalities in Landscape Genetics: An Extension of Wright's Causal Modeling to Distance Matrices

Identifying landscape features that affect functional connectivity among populations is a major challenge in fundamental and applied sciences. Landscape genetics combines landscape and genetic data to address this issue, with the main objective of disentangling direct and indirect relationships among an intricate set of variables. Causal modeling has strong potential to address the complex nature of landscape genetic data sets. However, this statistical approach was not initially developed to address the pairwise distance matrices commonly used in landscape genetics. Here, we aimed to extend the applicability of two causal modeling methods-that is, maximum-likelihood path analysis and the directional separation test-by developing statistical approaches aimed at handling distance matrices and improving functional connectivity inference. Using simulations, we showed that these approaches greatly improved the robustness of the absolute (using a frequentist approach) and relative (using an information-theoretic approach) fits of the tested models. We used an empirical data set combining genetic information on a freshwater fish species (Gobio occitaniae) and detailed landscape descriptors to demonstrate the usefulness of causal modeling to identify functional connectivity in wild populations. Specifically, we demonstrated how direct and indirect relationships involving altitude, temperature, and oxygen concentration influenced within- and between-population genetic diversity of G. occitaniae.

Systems Genetics Approach Identifies Gene Pathways and Adamts2 as Drivers of Isoproterenol-Induced Cardiac Hypertrophy and Cardiomyopathy in Mice

We previously reported a genetic analysis of heart failure traits in a population of inbred mouse strains treated with isoproterenol to mimic catecholamine-driven cardiac hypertrophy. Here, we apply a co-expression network algorithm, wMICA, to perform a systems-level analysis of left ventricular transcriptomes from these mice. We describe the features of the overall network but focus on a module identified in treated hearts that is strongly related to cardiac hypertrophy and pathological remodeling. Using the causal modeling algorithm NEO, we identified the gene Adamts2 as a putative regulator of this module and validated the predictive value of NEO using small interfering RNA-mediated knockdown in neonatal rat ventricular myocytes. Adamts2 silencing regulated the expression of the genes residing within the module and impaired isoproterenol-induced cellular hypertrophy. Our results provide a view of higher order interactions in heart failure with potential for diagnostic and therapeutic insights.

Application of the Marginal Structural Model to Account for Suboptimal Adherence in a Randomized Controlled Trial

Background

There is considerable interest in adjusting for suboptimal adherence in randomized controlled trials. A per-protocol analysis, for example removes individuals who fail to achieve a minimal level of adherence. One can also reassign non-adherers to the control group, censor them at the point of non-adherence, or cross them over to the control. However, there are biases inherent in each of these methods. Here, we describe an application of causal modeling to address this issue.

Methods

The marginal structural model with inverse-probability weighting was implemented using a weighted generalized estimating equation model. Two ancillary models were developed to derive the weights. First, stepwise linear regression was used to model the observed percent weight loss, while stepwise logistic regression model was applied to model early discontinuation from the intervention. From these, participant- and time-specific weights were calculated.

Discussion

This model is complicated and requires careful attention to detail. Which variables to force into the ancillary models, how to construct interaction terms, and how to address time-dependent covariates must be considered. Nevertheless, it can be used to great effect to predict intervention effects at full adherence. Moreover, by contrasting these results against intention-to-treat results, insights can be gained into the intrinsic physiologic effect of the intervention.

Trial registration

ClinicalTrials.gov Identifier NCT00427193.

Mediation Analysis with Survival Outcomes: Accelerated Failure Time vs. Proportional Hazards Models

Objective: Survival time is an important type of outcome variable in treatment research. Currently, limited guidance is available regarding performing mediation analyses with survival outcomes, which generally do not have normally distributed errors, and contain unobserved (censored) events. We present considerations for choosing an approach, using a comparison of semi-parametric proportional hazards (PH) and fully parametric accelerated failure time (AFT) approaches for illustration.

Method: We compare PH and AFT models and procedures in their integration into mediation models and review their ability to produce coefficients that estimate causal effects. Using simulation studies modeling Weibull-distributed survival times, we compare statistical properties of mediation analyses incorporating PH and AFT approaches (employing SAS procedures PHREG and LIFEREG, respectively) under varied data conditions, some including censoring. A simulated data set illustrates the findings.

Results: AFT models integrate more easily than PH models into mediation models. Furthermore, mediation analyses incorporating LIFEREG produce coefficients that can estimate causal effects, and demonstrate superior statistical properties. Censoring introduces bias in the coefficient estimate representing the treatment effect on outcome—underestimation in LIFEREG, and overestimation in PHREG. With LIFEREG, this bias can be addressed using an alternative estimate obtained from combining other coefficients, whereas this is not possible with PHREG.

Conclusions: When Weibull assumptions are not violated, there are compelling advantages to using LIFEREG over PHREG for mediation analyses involving survival-time outcomes. Irrespective of the procedures used, the interpretation of coefficients, effects of censoring on coefficient estimates, and statistical properties should be taken into account when reporting results.

Larger neural responses produce BOLD signals that begin earlier in time

Functional MRI analyses commonly rely on the assumption that the temporal dynamics of hemodynamic response functions (HRFs) are independent of the amplitude of the neural signals that give rise to them. The validity of this assumption is particularly important for techniques that use fMRI to resolve sub-second timing distinctions between responses, in order to make inferences about the ordering of neural processes. Whether or not the detailed shape of the HRF is independent of neural response amplitude remains an open question, however. We performed experiments in which we measured responses in primary visual cortex (V1) to large, contrast-reversing checkerboards at a range of contrast levels, which should produce varying amounts of neural activity. Ten subjects (ages 22-52) were studied in each of two experiments using 3 Tesla scanners. We used rapid, 250 ms, temporal sampling (repetition time, or TR) and both short and long inter-stimulus interval (ISI) stimulus presentations. We tested for a systematic relationship between the onset of the HRF and its amplitude across conditions, and found a strong negative correlation between the two measures when stimuli were separated in time (long- and medium-ISI experiments, but not the short-ISI experiment). Thus, stimuli that produce larger neural responses, as indexed by HRF amplitude, also produced HRFs with shorter onsets. The relationship between amplitude and latency was strongest in voxels with lowest mean-normalized variance (i.e., parenchymal voxels). The onset differences observed in the longer-ISI experiments are likely attributable to mechanisms of neurovascular coupling, since they are substantially larger than reported differences in the onset of action potentials in V1 as a function of response amplitude.

Validation of a causal model of agitation among institutionalized residents with dementia in Taiwan

The aim of this study was to test a causal model of the predictors of agitation among 405 nursing home residents in Taiwan with varying degrees of cognitive impairment. Chart review and behavioral observations were used to assess residents' physical and psychosocial condition. The final version of the model had a good fit. Cognitive function and depression had direct effects on agitation, and pain and functional ability had indirect effects on agitation via depression. Additionally, cognitive function and pain influenced functional ability directly, which in turn influenced depression and ultimately influenced agitation. The results suggest that effective management of agitation in demented residents requires identifying the needs underlying the behavior rather than directly treating the behavior itself.

Genetics of adverse reactions to haloperidol in a mouse diallel: a drug-placebo experiment and Bayesian causal analysis

Haloperidol is an efficacious antipsychotic drug that has serious, unpredictable motor side effects that limit its utility and cause noncompliance in many patients. Using a drug-placebo diallel of the eight founder strains of the Collaborative Cross and their F1 hybrids, we characterized aggregate effects of genetics, sex, parent of origin, and their combinations on haloperidol response. Treating matched pairs of both sexes with drug or placebo, we measured changes in the following: open field activity, inclined screen rigidity, orofacial movements, prepulse inhibition of the acoustic startle response, plasma and brain drug level measurements, and body weight. To understand the genetic architecture of haloperidol response we introduce new statistical methodology linking heritable variation with causal effect of drug treatment. Our new estimators, "difference of models" and "multiple-impute matched pairs", are motivated by the Neyman-Rubin potential outcomes framework and extend our existing Bayesian hierarchical model for the diallel (Lenarcic et al. 2012). Drug-induced rigidity after chronic treatment was affected by mainly additive genetics and parent-of-origin effects (accounting for 28% and 14.8% of the variance), with NZO/HILtJ and 129S1/SvlmJ contributions tending to increase this side effect. Locomotor activity after acute treatment, by contrast, was more affected by strain-specific inbreeding (12.8%). In addition to drug response phenotypes, we examined diallel effects on behavior before treatment and found not only effects of additive genetics (10.2-53.2%) but also strong effects of epistasis (10.64-25.2%). In particular: prepulse inhibition showed additivity and epistasis in about equal proportions (26.1% and 23.7%); there was evidence of nonreciprocal epistasis in pretreatment activity and rigidity; and we estimated a range of effects on body weight that replicate those found in our previous work. Our results provide the first quantitative description of the genetic architecture of haloperidol response in mice and indicate that additive, dominance-like inbreeding and parent-of-origin effects contribute strongly to treatment effect heterogeneity for this drug.

Improving causal inferences in risk analysis

Recent headlines and scientific articles projecting significant human health benefits from changes in exposures too often depend on unvalidated subjective expert judgments and modeling assumptions, especially about the causal interpretation of statistical associations. Some of these assessments are demonstrably biased toward false positives and inflated effects estimates. More objective, data-driven methods of causal analysis are available to risk analysts. These can help to reduce bias and increase the credibility and realism of health effects risk assessments and causal claims. For example, quasi-experimental designs and analysis allow alternative (noncausal) explanations for associations to be tested, and refuted if appropriate. Panel data studies examine empirical relations between changes in hypothesized causes and effects. Intervention and change-point analyses identify effects (e.g., significant changes in health effects time series) and estimate their sizes. Granger causality tests, conditional independence tests, and counterfactual causality models test whether a hypothesized cause helps to predict its presumed effects, and quantify exposure-specific contributions to response rates in differently exposed groups, even in the presence of confounders. Causal graph models let causal mechanistic hypotheses be tested and refined using biomarker data. These methods can potentially revolutionize the study of exposure-induced health effects, helping to overcome pervasive false-positive biases and move the health risk assessment scientific community toward more accurate assessments of the impacts of exposures and interventions on public health.

Selection on QTL and complex traits in complex environments

Understanding genetic variation for complex traits in heterogeneous environments is a fundamental problem in biology. In this issue of Molecular Ecology, Fournier-Level et al. (2013) analyse quantitative trait loci (QTL)influencing ecologically important phenotypes in mapping populations of Arabidopsis thaliana grown in four habitats across its native European range. They used causal modelling to quantify the selective consequences of life history and morphological traits and QTL on components of fitness. They found phenology QTL colocalizing with known flowering time genes as well as novel loci. Most QTL influenced fitness via life history and size traits, rather than QTL having direct effects on fitness.Comparison of phenotypes among environments found no evidence for genetic trade-offs for phenology or growth traits, but genetic trade-offs for fitness resulted because flowering time had opposite fitness effects in different environments. These changes in QTL effects and selective consequences may maintain genetic variation among populations.

Preventing Alzheimer's disease-related gray matter atrophy by B-vitamin treatment

Is it possible to prevent atrophy of key brain regions related to cognitive decline and Alzheimer's disease (AD)? One approach is to modify nongenetic risk factors, for instance by lowering elevated plasma homocysteine using B vitamins. In an initial, randomized controlled study on elderly subjects with increased dementia risk (mild cognitive impairment according to 2004 Petersen criteria), we showed that high-dose B-vitamin treatment (folic acid 0.8 mg, vitamin B6 20 mg, vitamin B12 0.5 mg) slowed shrinkage of the whole brain volume over 2 y. Here, we go further by demonstrating that B-vitamin treatment reduces, by as much as seven fold, the cerebral atrophy in those gray matter (GM) regions specifically vulnerable to the AD process, including the medial temporal lobe. In the placebo group, higher homocysteine levels at baseline are associated with faster GM atrophy, but this deleterious effect is largely prevented by B-vitamin treatment. We additionally show that the beneficial effect of B vitamins is confined to participants with high homocysteine (above the median, 11 µmol/L) and that, in these participants, a causal Bayesian network analysis indicates the following chain of events: B vitamins lower homocysteine, which directly leads to a decrease in GM atrophy, thereby slowing cognitive decline. Our results show that B-vitamin supplementation can slow the atrophy of specific brain regions that are a key component of the AD process and that are associated with cognitive decline. Further B-vitamin supplementation trials focusing on elderly subjets with high homocysteine levels are warranted to see if progression to dementia can be prevented.

A tutorial on propensity score estimation for multiple treatments using generalized boosted models

The use of propensity scores to control for pretreatment imbalances on observed variables in non-randomized or observational studies examining the causal effects of treatments or interventions has become widespread over the past decade. For settings with two conditions of interest such as a treatment and a control, inverse probability of treatment weighted estimation with propensity scores estimated via boosted models has been shown in simulation studies to yield causal effect estimates with desirable properties. There are tools (e.g., the twang package in R) and guidance for implementing this method with two treatments. However, there is not such guidance for analyses of three or more treatments. The goals of this paper are twofold: (1) to provide step-by-step guidance for researchers who want to implement propensity score weighting for multiple treatments and (2) to propose the use of generalized boosted models (GBM) for estimation of the necessary propensity score weights. We define the causal quantities that may be of interest to studies of multiple treatments and derive weighted estimators of those quantities. We present a detailed plan for using GBM to estimate propensity scores and using those scores to estimate weights and causal effects. We also provide tools for assessing balance and overlap of pretreatment variables among treatment groups in the context of multiple treatments. A case study examining the effects of three treatment programs for adolescent substance abuse demonstrates the methods.

Lipid adjustment in the analysis of environmental contaminants and human health risks

The literature on exposure to lipophilic agents such as polychlorinated biphenyls (PCBs) is conflicting, posing challenges for the interpretation of potential human health risks. Laboratory variation in quantifying PCBs may account for some of the conflicting study results. For example, for quantification purposes, blood is often used as a proxy for adipose tissue, which makes it necessary to model serum lipids when assessing health risks of PCBs. Using a simulation study, we evaluated four statistical models (unadjusted, standardized, adjusted, and two-stage) for the analysis of PCB exposure, serum lipids, and health outcome risk (breast cancer). We applied eight candidate true causal scenarios, depicted by directed acyclic graphs, to illustrate the ramifications of misspecification of underlying assumptions when interpreting results. Statistical models that deviated from underlying causal assumptions generated biased results. Lipid standardization, or the division of serum concentrations by serum lipids, was observed to be highly prone to bias. We conclude that investigators must consider biology, biologic medium (e.g., nonfasting blood samples), laboratory measurement, and other underlying modeling assumptions when devising a statistical plan for assessing health outcomes in relation to environmental exposures.