Search

We present the new R package instrument to perform Bayesian estimation of person explanatory multidimensional item response theory. The package implements an exploratory multidimensional item response theory model and a higher-order multidimensional item response theory model, a type of confirmatory multidimensional item response theory. Explanation of person parameters is accomplished by fixed and random effect linear regression models. Estimation is carried out using Hamiltonian Monte Carlo in Stan. In this article, we provide a detailed description of the models; we use the instrument package to demonstrate fitting explanatory item response models with fixed and random effects (i.e., mixed modeling) of person parameters in R; and, we perform a simulation study to evaluate the performance of our implementation of the models. [ABSTRACT FROM AUTHOR]

This study aimed to investigate how intermittent hyperoxic exposure (three cycles of 21% O2 [10 min] and 30% O2 [15 min]) affects exercise performance in mice. Three hours after the acute exposure, there was an observed increase in mRNA levels of phosphofructokinase (Bayes factor [BF] ≥ 10), mitochondrial transcription factor‐A (BF ≥10), PPAR‐α (BF ≥3), and PPAR‐γ (BF ≥3) in the red gastrocnemius muscle (Gr). Four weeks of exercise training under intermittent (INT), but not continuous (HYP), hyperoxia significantly (BF ≥30) increased maximal exercise capacity compared to normoxic exercise‐trained (ET) group. INT group exhibited significantly higher activity levels of 3‐hydroxyacyl‐CoA‐dehydrogenase (HAD) in Gr (BF = 7.9) compared to ET group. Pyruvate dehydrogenase complex activity levels were significantly higher in INT group compared to ET group in white gastrocnemius, diaphragm, and left ventricle (BF ≥3). NT‐PGC1α protein levels in Gr (BF = 7.7) and HAD activity levels in Gr (BF = 6.9) and soleus muscles (BF = 3.3) showed a significant positive correlation with maximal work values. These findings suggest that exercise training under intermittent hyperoxia is a beneficial strategy for enhancing endurance performance by improving fatty acid and pyruvic acid utilization. [ABSTRACT FROM AUTHOR]

Controlling assistive robots can be challenging for some users, especially those lacking relevant experience. Augmented Reality (AR) User Interfaces (UIs) have the potential to facilitate this task. Although extensive research regarding legged manipulators exists, comparatively little is on their UIs. Most existing UIs leverage traditional control interfaces such as joysticks, Hand-Held (HH) controllers and 2D UIs. These interfaces not only risk being unintuitive, thus discouraging interaction with the robot partner, but also draw the operator's focus away from the task and towards the UI. This shift in attention raises additional safety concerns, particularly in potentially hazardous environments where legged manipulators are frequently deployed. Moreover, traditional interfaces limit the operators' availability to use their hands for other tasks. Towards overcoming these limitations, in this article, we provide a user study comparing an AR Head-Mounted Display (HMD) UI we developed for controlling a legged manipulator against off-the-shelf control methods for such robots. This user study involved 27 participants and 135 trials, from which we gathered over 405 completed questionnaires. These trials involved multiple navigation and manipulation tasks with varying difficulty levels using a Boston Dynamics's Spot, a 7 df Kinova robot arm and a Robotiq 2F-85 gripper that we integrated into a legged manipulator. We made the comparison between UIs across multiple dimensions relevant to a successful human–robot interaction. These dimensions include cognitive workload, technology acceptance, fluency, system usability, immersion and trust. Our study employed a factorial experimental design with participants undergoing five different conditions, generating longitudinal data. Due to potential unknown distributions and outliers in such data, using parametric methods for its analysis is questionable, and while non-parametric alternatives exist, they may lead to reduced statistical power. Therefore, to analyse the data that resulted from our experiment, we chose Bayesian data analysis as an effective alternative to address these limitations. Our results show that AR UIs can outpace HH-based control methods and reduce the cognitive requirements when designers include hands-free interactions and cognitive offloading principles into the UI. Furthermore, the use of the AR UI together with our cognitive offloading feature resulted in higher usability scores and significantly higher fluency and Technology Acceptance Model scores. Regarding immersion, our results revealed that the response values for the AR Immersion questionnaire associated with the AR UI are significantly higher than those associated with the HH UI, regardless of the main interaction method with the former, i.e., hand gestures or cognitive offloading. Derived from the participants' qualitative answers, we believe this is due to a combination of factors, of which the most important is the free use of the hands when using the HMD, as well as the ability to see the real environment without the need to divert their attention to the UI. Regarding trust, our findings did not display discernible differences in reported trust scores across UI options. However, during the manipulation phase of our user study, where participants were given the choice to select their preferred UI, they consistently reported higher levels of trust compared to the navigation category. Moreover, there was a drastic change in the percentage of participants that selected the AR UI for completing this manipulation stage after incorporating the cognitive offloading feature. Thus, trust seems to have mediated the use and non-use of the UIs in a dimension different from the ones considered in our study, i.e., delegation and reliance. Therefore, our AR HMD UI for the control of legged manipulators was found to improve human–robot interaction across several relevant dimensions, underscoring the critical role of UI design in the effective and trustworthy utilisation of robotic systems. [ABSTRACT FROM AUTHOR]

Whether micromobility is hurting or boosting transit ridership remains a matter of debate. Previous studies on this topic mainly use either individual level data or system level data. This paper provides insights into this debate through analyses of the connection between bike-share use and transit use at both the individual-level and the system-level. The analysis uses data from an intercept survey of bike-share users and system-level data on bike-share trips from the Sacramento region's dockless electric bike-share system prior to the COVID-19 pandemic. Our individual-level analysis results suggest that people in the Sacramento region are more likely to replace their transit use with bike-share than to use bike-share as a first- or last-mile transit connector. Certain socio-demographic groups, however, are more likely to use bike share to connect to transit compared to others. Analysis of the system-level data shows that the number of bike-share trips that begin or end near transit stops is positively associated with transit boarding or alightings at those stops conditional on variables known to directly influence transit ridership. In this study, individual- and system-level analyses lead to different conclusions about the relationship between bike-share and transit, suggesting that reliance on system-level data alone may not provide an accurate assessment of the relationship between bike-share and transit use. A detailed understanding of the relationship using both sources of data can assist in better policy formulation that benefits both modes. [ABSTRACT FROM AUTHOR]

Abstract This study aimed to investigate how intermittent hyperoxic exposure (three cycles of 21% O2 [10 min] and 30% O2 [15 min]) affects exercise performance in mice. Three hours after the acute exposure, there was an observed increase in mRNA levels of phosphofructokinase (Bayes factor [BF] ≥ 10), mitochondrial transcription factor‐A (BF ≥10), PPAR‐α (BF ≥3), and PPAR‐γ (BF ≥3) in the red gastrocnemius muscle (Gr). Four weeks of exercise training under intermittent (INT), but not continuous (HYP), hyperoxia significantly (BF ≥30) increased maximal exercise capacity compared to normoxic exercise‐trained (ET) group. INT group exhibited significantly higher activity levels of 3‐hydroxyacyl‐CoA‐dehydrogenase (HAD) in Gr (BF = 7.9) compared to ET group. Pyruvate dehydrogenase complex activity levels were significantly higher in INT group compared to ET group in white gastrocnemius, diaphragm, and left ventricle (BF ≥3). NT‐PGC1α protein levels in Gr (BF = 7.7) and HAD activity levels in Gr (BF = 6.9) and soleus muscles (BF = 3.3) showed a significant positive correlation with maximal work values. These findings suggest that exercise training under intermittent hyperoxia is a beneficial strategy for enhancing endurance performance by improving fatty acid and pyruvic acid utilization.

The Gulf Stream is a vital limb of the North Atlantic circulation that influences regional climate, sea level, and hurricane activity. Given the Gulf Stream's relevance to weather and climate, many studies have attempted to estimate trends in its volumetric transport from various data sets, but results have been inconclusive, and no consensus has emerged whether it is weakening with climate change. Here we use Bayesian analysis to jointly assimilate multiple observational data sets from the Florida Straits to quantify uncertainty and change in Gulf Stream volume transport since 1982. We find with virtual certainty (probability P > 99%) that Gulf Stream volume transport through the Florida Straits declined by 1.2 ± 1.0 Sv in the past 40 years (95% credible interval). This significant trend has emerged from the data set only over the past ten years, the first unequivocal evidence for a recent multidecadal decline in this climate‐relevant component of ocean circulation. Plain Language Summary: The Gulf Stream is a major ocean current located off the East Coast of the United States. It carries a tremendous amount of seawater and along with it heat, carbon, and other ocean constituents. Because of this, the Gulf Stream plays an important role in weather and climate, influencing phenomena as seemingly unrelated as sea level along coastal Florida and temperature and precipitation over continental Europe. Given how important this ocean current is to science and society, scientists have tried to determine whether the Gulf Stream has undergone significant changes under global warming, but so far, they have not reached a firm conclusion. Here we report our effort to synthesize available Gulf Stream observations from the Florida Straits near Miami, and to assess whether and how the Gulf Stream transport there has changed since 1982. We conclude with a high degree of confidence that Gulf Stream transport has indeed slowed by about 4% in the past 40 years, the first conclusive, unambiguous observational evidence that this ocean current has undergone significant change in the recent past. Future studies should try to identify the cause of this change. Key Points: Gulf Stream volume transport through Florida Straits declined by 1.2 ± 1.0 Sv during the past 40 years (95% credible interval)We find a weakening trend in the Gulf Stream by applying Bayesian methods to synthesize cable, in situ, and satellite data sets congruently [ABSTRACT FROM AUTHOR]

Because fish that enter the interior Delta have poorer survival than those emigrating via the Sacramento River, understanding the mechanisms that drive entrainment rates at side channel junctions is critically important for the management of imperiled juvenile salmon. Here, we implement a previously proposed process-based conceptual model to study entrainment rates based on three linked elements: the entrainment zone, critical streakline, and cross-sectional distribution of fish. The critical streakline is the location along a channel cross-section immediately upstream of a junction that forms the spatial divide between parcels of water that enter a side channel or remain in the main channel. The critical streakline therefore divides the main channel into entrainment zones within which fish would likely enter each channel. Combined with information about the cross-sectional distribution of fish upstream of a junction, this conceptual model provides a means to predict fish entrainment into each channel. To apply this conceptual model, we combined statistical models of the critical streakline, the cross-sectional distribution of acoustic tagged juvenile Chinook salmon, and their probability of entrainment into Georgiana Slough. We fit joint beta regression and logistic regression models to acoustic telemetry data gathered in 2011 and 2012 to estimate the cross-sectional distribution of fish upstream of the junction, and to estimate the probability of entrainment for fish on either side of the critical streakline. We show that entrainment rates can be predicted by understanding how the combination of critical streakline position and cross-sectional distribution of fish co-vary as a function of environmental covariates. By integrating over individual positions and entrainment fates to arrive at population-level entrain probability in relation to environmental covariates, our model offers managers a simple but powerful tool to evaluate how alternative actions affect migrating fish.

Abstract The Gulf Stream is a vital limb of the North Atlantic circulation that influences regional climate, sea level, and hurricane activity. Given the Gulf Stream's relevance to weather and climate, many studies have attempted to estimate trends in its volumetric transport from various data sets, but results have been inconclusive, and no consensus has emerged whether it is weakening with climate change. Here we use Bayesian analysis to jointly assimilate multiple observational data sets from the Florida Straits to quantify uncertainty and change in Gulf Stream volume transport since 1982. We find with virtual certainty (probability P > 99%) that Gulf Stream volume transport through the Florida Straits declined by 1.2 ± 1.0 Sv in the past 40 years (95% credible interval). This significant trend has emerged from the data set only over the past ten years, the first unequivocal evidence for a recent multidecadal decline in this climate‐relevant component of ocean circulation.

Multiaxial fatigue failure is the most common problem in engineering structures. It is worth noting that short crack growth accounts for most of the fatigue life. Hence, it is necessary to study the short crack growth models for multiaxial fatigue life assessment. The primary focus of this study is to develop a hierarchical Bayesian linear regression method to estimate parameters in multiaxial fatigue crack growth model. The Bayesian method is used to estimate the intercept and slope of the regression equation for each loading path and ensemble test datasets. The method of this work was demonstrated on three multiaxial fatigue crack growth datasets. The main results obtained in this paper were that the parameters of the multiaxial fatigue crack growth model changed significantly with different loading paths, and the parameters of the model depended on the multiaxial loading path. Highlights: An equivalent strain‐based intensity factor ΔKESA was proposed.Hierarchical Bayesian linear regression is developed for parameter estimation.Hierarchical Bayesian model considers the differences of each loading paths.The parameters of the fatigue crack growth model depend on the loading path. [ABSTRACT FROM AUTHOR]

Investigating the relationship between task‐related hemodynamic responses and cortical excitability is challenging because it requires simultaneous measurement of hemodynamic responses while applying noninvasive brain stimulation. Moreover, cortical excitability and task‐related hemodynamic responses are both associated with inter‐/intra‐subject variability. To reliably assess such a relationship, we applied hierarchical Bayesian modeling. This study involved 16 healthy subjects who underwent simultaneous Paired Associative Stimulation (PAS10, PAS25, Sham) while monitoring brain activity using functional Near‐Infrared Spectroscopy (fNIRS), targeting the primary motor cortex (M1). Cortical excitability was measured by Motor Evoked Potentials (MEPs), and the motor task‐related hemodynamic responses were measured using fNIRS 3D reconstructions. We constructed three models to investigate: (1) PAS effects on the M1 excitability, (2) PAS effects on fNIRS hemodynamic responses to a finger tapping task, and (3) the correlation between PAS effects on M1 excitability and PAS effects on task‐related hemodynamic responses. Significant increase in cortical excitability was found following PAS25, whereas a small reduction of the cortical excitability was shown after PAS10 and a subtle increase occurred after sham. Both HbO and HbR absolute amplitudes increased after PAS25 and decreased after PAS10. The probability of the positive correlation between modulation of cortical excitability and hemodynamic activity was 0.77 for HbO and 0.79 for HbR. We demonstrated that PAS stimulation modulates task‐related cortical hemodynamic responses in addition to M1 excitability. Moreover, the positive correlation between PAS modulations of excitability and hemodynamics brought insight into understanding the fundamental properties of cortical function and cortical excitability. [ABSTRACT FROM AUTHOR]

Abstract Background Just Another Gibbs Sampling (JAGS) is a convenient tool to draw posterior samples using Markov Chain Monte Carlo for Bayesian modeling. However, the built-in function dinterval() for censored data misspecifies the default computation of deviance function, which limits likelihood-based Bayesian model comparison. Results To establish an automatic approach to specifying the correct deviance function in JAGS, we propose a simple and generic alternative modeling strategy for the analysis of censored outcomes. The two illustrative examples demonstrate that the alternative strategy not only properly draws posterior samples in JAGS, but also automatically delivers the correct deviance for model assessment. In the survival data application, our proposed method provides the correct value of mean deviance based on the exact likelihood function. In the drug safety data application, the deviance information criterion and penalized expected deviance for seven Bayesian models of censored data are simultaneously computed by our proposed approach and compared to examine the model performance. Conclusions We propose an effective strategy to model censored data in the Bayesian modeling framework in JAGS with the correct deviance specification, which can simplify the calculation of popular Kullback–Leibler based measures for model selection. The proposed approach applies to a broad spectrum of censored data types, such as survival data, and facilitates different censored Bayesian model structures.

Robust scientific inference is crucial to ensure evidence-based decision making. Accordingly, the selection of appropriate statistical tools and experimental designs is integral to achieve accuracy from data analytical processes. Environmental monitoring of water quality has become increasingly common and widespread as a result of technological advances, leading to an abundance of datasets. We conducted a scoping review of the water quality literature and found that correlation and linear regression are by far the most used statistical tools. However, the accuracy of inferences drawn from ordinary least squares (OLS) techniques depends on a set of assumptions, most prominently: (a) independence among observations, (b) normally distributed errors, (c) equal variances of errors, and (d) balanced designs. Environmental data, however, are often faced with temporal and spatial dependencies, and unbalanced designs, thus making OLS techniques not suitable to provide valid statistical inferences. Generalized least squares (GLS), linear mixed-effect models (LMMs), and generalized linear mixed-effect models (GLMMs), as well as Bayesian data analyses, have been developed to better tackle these problems. Recent progress in the development of statistical software has made these approaches more accessible and user-friendly. We provide a high-level summary and practical guidance for those statistical techniques. HIGHLIGHTS Correlation and linear regression are commonly used to assess water quality data.; Environmental data, however, are often characterized by temporal and spatial dependency structures in the data thus making ordinary least squares techniques inappropriate.; Generalized least squares, linear mixed, and generalized linear mixed-effect models, as well as Bayesian techniques, may be more suitable for such data.;

Probability theory is the “logic of science” (Jaynes, 2003) andBayesian data analysis (BDA) is the glue that brings that logicto data. BDA is a general, flexible alternative to standard statis-tical approaches (e.g., NHST) that provides the scientist withclarity and ease to address their personal scientific questions.Doing BDA in a probabilistic programming language (PPL) af-fords several additional advantages: a compositional approachto writing models, separation of model specification from al-gorithmic implementation (a la lm() in R), and continuity fromarticulating data analytic models to Bayesian cognitive mod-els. Furthermore, specifying one’s model and data analysisin a PPL allows you to search for “optimal experiments” forfree. This tutorial will walk the participant through the basicsof BDA to state-of-the-art applications, using an interactive on-line web-book and tools for integrating BDA into their existingworkflow.

This study was designed to (1) investigate the effects of acute exercise under intermittent hypoxia on muscle mRNA and protein levels, and (2) clarify the mechanisms by which exercise under intermittent hypoxia improves endurance capacity. Experiment‐1: Male mice were subjected to either acute endurance exercise, exercise under hypoxia (14% O2), exercise under intermittent hypoxia (Int, three cycles of room air [10 min] and 14% O2 [15 min]). At 3 h after exercise under intermittent hypoxia, sirtuin‐6 mRNA levels and nuclear prolyl hydroxylases‐2 protein levels were significantly upregulated in white gastrocnemius muscle in the Int group. Experiment‐2: Mice were assigned to sedentary control (Sed), normoxic exercise‐trained (ET), hypoxic exercise‐trained (HYP) or exercise‐trained under intermittent hypoxia (INT) groups. Exercise capacity was significantly greater in the INT group than in the ET and HYP group. Activity levels of citrate synthase were significantly greater in the INT group than in the HYP group in soleus (SOL) and red gastrocnemius muscles. In SOL, nuclear N‐terminal PGC1α levels were considerably increased by the INT training (95% confidence interval [CI]: 1.09–1.79). The INT significantly increased pyruvate dehydrogenase complex activity levels in left ventricle (LV). Monocarboxylate transporter‐4 protein levels were significantly increased after the INT training in LV. Capillary‐to‐fiber ratio values were significantly increased in SOL and were substantially increased in LV (CI: 1.10–1.22) after the INT training. These results suggest that exercise training under intermittent hypoxia represents a beneficial strategy for increasing endurance performance via improving metabolic properties and capillary profiles in several hind‐leg muscles and the heart. [ABSTRACT FROM AUTHOR]

Background: Just Another Gibbs Sampling (JAGS) is a convenient tool to draw posterior samples using Markov Chain Monte Carlo for Bayesian modeling. However, the built-in function dinterval() for censored data misspecifies the default computation of deviance function, which limits likelihood-based Bayesian model comparison. Results: To establish an automatic approach to specifying the correct deviance function in JAGS, we propose a simple and generic alternative modeling strategy for the analysis of censored outcomes. The two illustrative examples demonstrate that the alternative strategy not only properly draws posterior samples in JAGS, but also automatically delivers the correct deviance for model assessment. In the survival data application, our proposed method provides the correct value of mean deviance based on the exact likelihood function. In the drug safety data application, the deviance information criterion and penalized expected deviance for seven Bayesian models of censored data are simultaneously computed by our proposed approach and compared to examine the model performance. Conclusions: We propose an effective strategy to model censored data in the Bayesian modeling framework in JAGS with the correct deviance specification, which can simplify the calculation of popular Kullback–Leibler based measures for model selection. The proposed approach applies to a broad spectrum of censored data types, such as survival data, and facilitates different censored Bayesian model structures. [ABSTRACT FROM AUTHOR]

We introduce the new package dmbc that implements a Bayesian algorithm for clustering a set of binary dissimilarity matrices within a model-based framework. Specifically, we consider the case when S matrices are available, each describing the dissimilarities among the same n objects, possibly expressed by S subjects (judges), or measured under different experimental conditions, or with reference to different characteristics of the objects themselves. In particular, we focus on binary dissimilarities, taking values 0 or 1 depending on whether or not two objects are deemed as dissimilar. We are interested in analyzing such data using multidimensional scaling (MDS). Differently from standard MDS algorithms, our goal is to cluster the dissimilarity matrices and, simultaneously, to extract an MDS configuration specific for each cluster. To this end, we develop a fully Bayesian three-way MDS approach, where the elements of each dissimilarity matrix are modeled as a mixture of Bernoulli random vectors. The parameter estimates and the MDS configurations are derived using a hybrid Metropolis-Gibbs Markov Chain Monte Carlo algorithm. We also propose a BIC-like criterion for jointly selecting the optimal number of clusters and latent space dimensions. We illustrate our approach referring both to synthetic data and to a publicly available data set taken from the literature. For the sake of efficiency, the core computations in the package are implemented in C/C++. The package also allows the simulation of multiple chains through the support of the parallel package.

Abstract This study was designed to (1) investigate the effects of acute exercise under intermittent hypoxia on muscle mRNA and protein levels, and (2) clarify the mechanisms by which exercise under intermittent hypoxia improves endurance capacity. Experiment‐1: Male mice were subjected to either acute endurance exercise, exercise under hypoxia (14% O2), exercise under intermittent hypoxia (Int, three cycles of room air [10 min] and 14% O2 [15 min]). At 3 h after exercise under intermittent hypoxia, sirtuin‐6 mRNA levels and nuclear prolyl hydroxylases‐2 protein levels were significantly upregulated in white gastrocnemius muscle in the Int group. Experiment‐2: Mice were assigned to sedentary control (Sed), normoxic exercise‐trained (ET), hypoxic exercise‐trained (HYP) or exercise‐trained under intermittent hypoxia (INT) groups. Exercise capacity was significantly greater in the INT group than in the ET and HYP group. Activity levels of citrate synthase were significantly greater in the INT group than in the HYP group in soleus (SOL) and red gastrocnemius muscles. In SOL, nuclear N‐terminal PGC1α levels were considerably increased by the INT training (95% confidence interval [CI]: 1.09–1.79). The INT significantly increased pyruvate dehydrogenase complex activity levels in left ventricle (LV). Monocarboxylate transporter‐4 protein levels were significantly increased after the INT training in LV. Capillary‐to‐fiber ratio values were significantly increased in SOL and were substantially increased in LV (CI: 1.10–1.22) after the INT training. These results suggest that exercise training under intermittent hypoxia represents a beneficial strategy for increasing endurance performance via improving metabolic properties and capillary profiles in several hind‐leg muscles and the heart.

Second language (L2) learners are known to have difficulty producing inflection in obligatory contexts reliably. According to the Prosodic Transfer Hypothesis (PTH), the prosodic organisation of L2 inflection is constrained by the inventory of representations available in the L1. At the same time, this hypothesis does not explicitly limit how freely prosodic representations can be transferred, so that transfer across constructions within the same domain (e.g., verbal domain: L1 tense → L2 agreement) and across domains (e.g., verbal domain: L1 tense → nominal domain: L2 plurals) are both possible in principle. The goal of this study was to determine if the current formulation of the PTH is valid, or must be reined in to exclude transfer across domains in particular. Forty-four Korean learners of English did a spoken sentence-construction task in which they had to produce subject-verb agreement and regular plural inflection. Bayesian hierarchical regression was used to analyse the results. By examining asymmetries in the suppliance of short- vs. long-stemmed inflection, we show that there are no grounds for attaching any stipulations to the PTH along the above lines, as prosodic representations are transferrable not only across constructions but also across domains. [ABSTRACT FROM AUTHOR]

Sonority is a fundamental notion in phonetics and phonology, central to many descriptions of the syllable and various useful predictions in phonotactics. Although widely accepted, sonority lacks a clear basis in speech articulation or perception, given that traditional formal principles in linguistic theory are often exclusively based on discrete units in symbolic representation and are typically not designed to be compatible with auditory perception, sensorimotor control, or general cognitive capacities. In addition, traditional sonority principles also exhibit systematic gaps in empirical coverage. Against this backdrop, we propose the incorporation of symbol‐based and signal‐based models to adequately account for sonority in a complementary manner. We claim that sonority is primarily a perceptual phenomenon related to pitch, driving the optimization of syllables as pitch‐bearing units in all language systems. We suggest a measurable acoustic correlate for sonority in terms of periodic energy, and we provide a novel principle that can account for syllabic well‐formedness, the nucleus attraction principle (NAP). We present perception experiments that test our two NAP‐based models against four traditional sonority models, and we use a Bayesian data analysis approach to test and compare them. Our symbolic NAP model outperforms all the other models we test, while our continuous bottom‐up NAP model is at second place, along with the best performing traditional models. We interpret the results as providing strong support for our proposals: (i) the designation of periodic energy as the acoustic correlate of sonority; (ii) the incorporation of continuous entities in phonological models of perception; and (iii) the dual‐model strategy that separately analyzes symbol‐based top‐down processes and signal‐based bottom‐up processes in speech perception. [ABSTRACT FROM AUTHOR]

Statistical analysis is the tool of choice to turn data into information and then information into empirical knowledge. However, the process that goes from data to knowledge is long, uncertain, and riddled with pitfalls. To be valid, it should be supported by detailed, rigorous guidelines that help ferret out issues with the data or model and lead to qualified results that strike a reasonable balance between generality and practical relevance. Such guidelines are being developed by statisticians to support the latest techniques for Bayesian data analysis. In this article, we frame these guidelines in a way that is apt to empirical research in software engineering. To demonstrate the guidelines in practice, we apply them to reanalyze a GitHub dataset about code quality in different programming languages. The dataset’s original analysis [Ray et al. 2014] and a critical reanalysis [Berger et al. 2019] have attracted considerable attention—in no small part because they target a topic (the impact of different programming languages) on which strong opinions abound. The goals of our reanalysis are largely orthogonal to this previous work, as we are concerned with demonstrating, on data in an interesting domain, how to build a principled Bayesian data analysis and to showcase its benefits. In the process, we will also shed light on some critical aspects of the analyzed data and of the relationship between programming languages and code quality—such as the impact of project-specific characteristics other than the used programming language. The high-level conclusions of our exercise will be that Bayesian statistical techniques can be applied to analyze software engineering data in a way that is principled, flexible, and leads to convincing results that inform the state-of-the-art while highlighting the boundaries of its validity. The guidelines can support building solid statistical analyses and connecting their results. Thus, they can help buttress continued progress in empirical software engineering research. [ABSTRACT FROM AUTHOR]

Identifying the structural dependence between the cryptocurrencies and predicting market trend are fundamental for effective portfolio management in cryptocurrency trading. In this paper, we present a unified Bayesian machine learning framework based on potential field theory and Gaussian Process to characterize the structural dependency of various cryptocurrencies, using historic price information. The following are our significant contributions: (i) Proposed a novel model for cryptocurrency price movements as a trajectory of a dynamical system governed by a time-varying non-linear potential field. (ii) Developed a Bayesian machine learning framework for inferring the non-linear potential function from observed cryptocurrency prices. (iii) Proposed that attractors and repellers inferred from the potential field are reliable cryptocurrency market indicators, surpassing existing attributes in the literature. (iv) Analysis of cryptocurrency market during various Bitcoin crash durations shows that attractors captured the market trend, volatility, and correlation. In addition, attractors aids explainability and visualization. (v) The proposed cryptocurrency market indicators (attractors and repellers) was used to improve the prediction performance of state-of-art deep learning price prediction models. [ABSTRACT FROM AUTHOR]

Speakers often refer to context only implicitly when using lan-guage. The utterance “it’s warm outside” could signal it’swarm relative to other days of the year or just relative to thecurrent season (e.g., it’s warm for winter). Warm vaguely con-veys that the temperature is high relative to some contextualcomparison class, but little is known about how a listener de-cides upon such a standard of comparison. Here, we formalizehow world knowledge and listeners’ internal models of speechproduction can drive the resolution of a comparison class incontext. We introduce a Rational Speech Act model and de-rive two novel predictions from it, which we validate using aparaphrase experiment to measure listeners’ beliefs about thelikely comparison class used by a speaker. Our model makesquantitative predictions given prior world knowledge for thedomains in question. We triangulate this knowledge with afollow-up language task in the same domains, using Bayesiandata analysis to infer priors from both data sets

Clinical intelligence about a patient’s risk of future adverse health events can support clinical decision making in personalized and preventive care. Healthcare predictive analytics using electronic health records offers a promising direction to address the challenging tasks of risk profiling. Patients with chronic diseases often face risks of not just one, but an array of adverse health events. However, existing risk models typically focus on one specific event and do not predict multiple outcomes. To attain enhanced risk profiling, we adopt the design science paradigm and propose a principled approach called Bayesian multitask learning (BMTL). Considering the model development for an event as a single task, our BMTL approach is to coordinate a set of baseline models—one for each event—and communicate training information across the models. The BMTL approach allows healthcare providers to achieve multifaceted risk profiling and model an arbitrary number of events simultaneously. Our experimental evaluations demonstrate that the BMTL approach attains an improved predictive performance when compared with the alternatives that model multiple events separately. We also find that, in most cases, the BMTL approach significantly outperforms existing multitask learning techniques. More importantly, our analysis shows that the BMTL approach can create significant potential impacts on clinical practice in reducing the failures and delays in preventive interventions. We discuss several implications of this study for health IT, big data and predictive analytics, and design science research. [ABSTRACT FROM AUTHOR]

[Display omitted] • Enricherator provides fully Bayesian analysis of targeted enrichment data on genomic pulldowns (such as ChIP-seq) • Enricherator directly reports intuitive and interpretable results – including Bayesian credible intervals on enrichment scores, the Bayesian equivalent of statistical significance testing, correction for multiple hypothesis testing, and effect size estimation – as part of its standard output. • Enricherator properly handles sequencing count data by using a negative binomial likelihood. • Enricherator uses a shrinkage prior on enrichment scores to directly report enrichment estimates that are intuitively useful. • Enricherator uses local covariation in sequencing coverage to achieve more accurate estimates of enrichment and credible intervals. A pervasive question in biological research studying gene regulation, chromatin structure, or genomics is where, and to what extent, does a signal of interest arise genome-wide? This question is addressed using a variety of methods relying on high-throughput sequencing data as their final output, including ChIP-seq for protein-DNA interactions, 1 GapR-seq for measuring supercoiling, 2 and HBD-seq or DRIP-seq for R-loop positioning. 3,4 Current computational methods to calculate genome-wide enrichment of the signal of interest usually do not properly handle the count-based nature of sequencing data, they often do not make use of the local correlation structure of sequencing data, and they do not apply any regularization of enrichment estimates. This can result in unrealistic estimates of the true underlying biological enrichment of interest, unrealistically low estimates of confidence in point estimates of enrichment (or no estimates of confidence at all), unrealistic gyrations in enrichment estimates at very close (<10 bp) genomic loci due to noise inherent in sequencing data, and in a multiple-hypothesis testing problem during interpretation of genome-wide enrichment estimates. We developed a tool called Enricherator to infer genome-wide enrichments from sequencing count data. Enricherator uses the variational Bayes algorithm to fit a generalized linear model to sequencing count data and to sample from the approximate posterior distribution of enrichment estimates (https://github.com/jwschroeder3/enricherator). Enrichments inferred by Enricherator more precisely identify known binding sites in cases where low coverage between binding sites leads to false-positive peak calls in these noisy regions of the genome; these benefits extend to published datasets. [ABSTRACT FROM AUTHOR]

Cue-based retrieval theories of sentence processing posit that long-distance dependency formation is guided by a cue-based retrieval mechanism: dependents are retrieved via retrieval cues associated with a verb. When retrieval cues match multiple similar items in memory, this leads to cue-based retrieval interference. A landmark study by Van Dyke and McElree tested interference from sentence-external items: retrieval cues were manipulated to (mis-)match semantically similar items presented prior to a target dependency. The support for interference of this type is weak, and only comes from English object cleft constructions. Our study provides a cross-linguistic investigation of interference from sentence-external items: Three eyetracking studies in English, German and Russian tested interference in the online processing of filler-gap dependencies under varying task demands. A fourth study attempted to replicate the Van Dyke and McElree study using self-paced reading. Bayes factors analyses show cross-linguistic evidence against interference from sentence-external items. A broader implication from these data is that cue-based retrieval interference is driven by sentence-internal distracting items, suggesting that a cue-based search is restricted to the current linguistic context. • Cross-linguistic study on cue-based interference from sentence-external items. • Three larger-sample eye-tracking studies in English, German and Russian. • A fourth self-paced reading study attempts to replicate the original Van Dyke study. • Data show evidence against interference from sentence-external items. [ABSTRACT FROM AUTHOR]

With increasing traffic demand in urban areas of metropolises, many tunnels have been constructed to improve road capacity and traffic mobility. The distance between two consecutive tunnels is relatively short which usually forms a weaving section, leading to considerable traffic conflicts. The objective of this study is to evaluate the safety performance of such inter-tunnel sections. Conflict prediction models based on negative binomial regression were developed to identify influential factors. Field data were collected at ten selected sites in Nanjing, China, and used for calibrating and validating the proposed models. Two types of inter-tunnel weaving sections (type 1 and type 2) were found in the field with distinct lane markings and operation rules. The unique lane markings in type 1 weaving sections are designed to isolate weaving traffic flows and thus reduce conflicts, but in practice, contradictory to its design intention, lead to more traffic conflicts compared with type 2 weaving sections. In addition, the length of the diverging section, merging section, and whole weaving section are found to be significant influencing factors on the conflict occurrence. The findings in the present study are expected to help engineer better design inter-tunnel sections. [ABSTRACT FROM AUTHOR]

Habitual sentences (e.g. Bill smokes.) generalize an event overtime, but how do you know when a habitual sentence is true?We develop a computational model and use this to guide exper-iments into the truth conditions of habitual language. In Ex-pts. 1 & 2, we measure participants’ prior expectations aboutthe frequency with which an event occurs and validate thepredictions of the model for when a habitual sentence is ac-ceptable. In Expt. 3, we show that habituals are sensitive totop-down moderators of expected frequency: It is the expec-tation of future tendency that matters for habitual language.This work provides the mathematical glue between our intu-itive theories’ of others and events and the language we useto talk about them.

People’s beliefs about everyday events are both of theoreti-cal interest in their own right and an important ingredient inmodel building—especially in Bayesian cognitive models ofphenomena such as logical reasoning, future predictions, andlanguage use. Here, we explore several recently used methodsfor measuring subjective beliefs about unidimensional contigu-ous properties, such as the likely price of a new watch. Asa first step towards a way of assessing and comparing beliefelicitation methods, we use hierarchical Bayesian modeling forinferring likely population-level beliefs as the central tendencyof participants’ individual-level beliefs. Three different depen-dent measures are considered: (i) slider ratings of (relative)likelihood of intervals of values, (ii) a give-a-number task, and(iii) choice of the more likely of two intervals of values. Ourresults suggest that using averaged normalized slider ratingsfor binned quantities is a practical and fairly good approxima-tor of inferred population-level beliefs.

The meanings of natural language utterances depend heavily on context. Yet, what counts as context is often only implicit in conversation. The utterance it's warm outside signals that the temperature outside is relatively high, but the temperature could be high relative to a number of different comparison classes: other days of the year, other weeks, other seasons, etc. Theories of context sensitivity in language agree that the comparison class is a crucial variable for understanding meaning, but little is known about how a listener decides upon the comparison class. Using the case study of gradable adjectives (e.g., warm), we extend a Bayesian model of pragmatic inference to reason flexibly about the comparison class and test its qualitative predictions in a large‐scale free‐production experiment. We find that human listeners infer the comparison class by reasoning about the kinds of observations that would be remarkable enough for a speaker to mention, given the speaker and listener's shared knowledge of the world. Further, we quantitatively synthesize the model and data using Bayesian data analysis, which reveals that usage frequency and a preference for basic‐level categories are two main factors in comparison class inference. This work presents new data and reveals the mechanisms by which human listeners recover the relevant aspects of context when understanding language. [ABSTRACT FROM AUTHOR]

Summary: This paper proposes the novel past‐faults fault prediction algorithm Linespots, based on the Bugspots algorithm. We analyse the predictive performance and runtime of Linespots compared with Bugspots with an empirical study using the most significant self‐built dataset as of now, including high‐quality samples for validation. As a novelty in fault prediction, we use Bayesian data analysis and Directed Acyclic Graphs to model the effects. We found consistent improvements in the predictive performance of Linespots over Bugspots for all seven evaluation metrics. We conclude that Linespots should be used over Bugspots in all cases where no real‐time performance is necessary. [ABSTRACT FROM AUTHOR]

Motivated by firms' increasing use of new media technology for investor communications, we investigate how alignment between company image and communication platform affects investor judgment and decision making. In our first experiment, we demonstrate that investors expect alignment between firm image and the perception of the new media communication platform managers choose for investor relations. In a second experiment, we examine how this alignment affects investor judgment and decision making. We predict and find that greater platform-image alignment leads investors to experience subjective ease of processing, but does not change investment amounts. Additionally, we demonstrate an approach to conducting an explicit test of a null hypothesis by evaluating the convergence of null hypothesis significance testing (NHST) and Bayesian methods. Our findings have implications for researchers, firms, and investors, and add to a growing literature on new media disclosure. [ABSTRACT FROM AUTHOR]

This article explores how beliefs affect preferences leading to policy instrument choices of elite actors. Beliefs are general attitudes regarding a given policy field, for example toward the role of the state or the urgency of a problem. Both beliefs and preferences are central for applications of Sabatier's Advocacy Coalition Framework, but their interrelationship has remained undertheorized. Understanding how beliefs and preferences are linked can provide important insights into policy instrument choice, while improving the comparability of studies across policy subsystems. The article compares the relative contribution of beliefs to shaping instrument choices of elite actors in the domain of Swiss renewable energy policy. Results suggest that beliefs are likely to play a prominent role in shaping instrument choice. We find that policy core beliefs translate into preferences through a process involving two main pathways. First, some policy beliefs primarily influence the preferred characteristics of the overall instrument mix. Second, some policy beliefs are primarily associated with preferences for specific instruments. Some policy beliefs are influential via both pathways. These, therefore, emerge as especially important factors shaping the policy process. Our results offer insights for policymakers into how potential future conflicts in negotiations can be attenuated. [ABSTRACT FROM AUTHOR]

High throughput sequencing (HTS)-based technology enables identifying and quantifying non-culturable microbial organisms in all environments. Microbial sequences have enhanced our understanding of the human microbiome, the soil and plant environment, and the marine environment. All molecular microbial data pose statistical challenges due to contamination sequences from reagents, batch effects, unequal sampling, and undetected taxa. Technical biases and heteroscedasticity have the strongest effects, but different strains across subjects and environments also make direct differential abundance testing unwieldy. We provide an introduction to a few statistical tools that can overcome some of these difficulties and demonstrate those tools on an example. We show how standard statistical methods, such as simple hierarchical mixture and topic models, can facilitate inferences on latent microbial communities. We also review some nonparametric Bayesian approaches that combine visualization and uncertainty quantification. The intersection of molecular microbial biology and statistics is an exciting new venue. Finally, we list some of the important open problems that would benefit from more careful statistical method development. [ABSTRACT FROM AUTHOR]

• Municipal waste data with repeated observations over time were considered. • A multilevel Bayesian framework for prediction was proposed. • Prediction models at aggregated and disaggregated levels were compared. • Models were tested on real-life data of yearly waste generation rates from Denmark. • The multilevel model outperformed traditional models in our case study. Prediction of waste production is an essential part of the design and planning of waste management systems. The quality and applicability of such predictions depend heavily on model assumptions and the structure of the collected data. Ordinarily, municipal waste generation data are organized in hierarchical structures with municipal or county levels, and multilevel models can be used to generalize linear regression by directly incorporating the structure into the model. However, small amounts of data can limit the applicability of multilevel models and provide biased estimates. To cope with this problem, Bayesian estimation is often recommended as an alternative to frequentist estimation, such as least squares or maximum likelihood estimation. This paper proposes a multilevel framework under a Bayesian approach to model municipal waste generation with hierarchical data structures. Using a real-world dataset of municipal waste generation in Denmark, the predictive accuracy of multilevel models is compared to aggregated and disaggregated Bayesian models using socio-economic external variables. Results show that Bayesian multilevel models outperform the other models in prediction accuracy, based on the leave-one-out information criterion. A comparison of the Bayesian approach with its frequentist alternative shows that the Bayesian model is more conservative in coefficient estimation, with estimates shrinking to the grand mean and broader credible intervals, in contrast with narrower confidence intervals produced by the frequentist models. [ABSTRACT FROM AUTHOR]

Abstract Hyperbaric exposure (1.3 atmospheres absolute with 20.9% O2) for 1 h a day was shown to improve exercise capacity. The present study was designed to reveal whether the daily exposure time affects exercise performance and metabolism in skeletal and cardiac muscles. Male mice in the training group were housed in a cage with a wheel activity device for 7 weeks from 5 weeks old. Trained mice were then subjected to hybrid training (HT, endurance exercise for 30 min followed by sprint interval exercise for 30 min). Hyperbaric exposure was applied following daily HT for 15 min (15HT), 30 min (30HT), or 60 min (60HT) for 4 weeks. In the endurance capacity test, maximal work values were significantly increased by 30HT and 60HT. In the left ventricle (LV), activity levels of 3‐hydroxyacyl‐CoA‐dehydrogenase, citrate synthase, and carnitine palmitoyl transferase (CPT) 2 were significantly increased by 60HT. CPT2 activity levels were markedly increased by hyperbaric exposure in red gastrocnemius (Gr) and plantaris muscle (PL). Pyruvate dehydrogenase complex activity values in PL were enhanced more by 30HT and 60HT than by HT. Protein levels of N‐terminal isoform of PGC1α (NT‐PGC1α) protein were significantly enhanced in three hyperbaric exposed groups in Gr, but not in LV. These results indicate that hyperbaric exposure for 30 min or longer has beneficial effects on endurance, and 60‐min exposure has the potential to further increase performance by facilitating fatty acid metabolism in skeletal and cardiac muscles in highly trained mice. NT‐PGC1α may have important roles for these adaptations in skeletal muscle.

Attempts to formalize inspection and monitoring strategies in industry have struggled to combine evidence from multiple sources (including subject matter expertise) in a mathematically coherent way. The perceived requirement for large amounts of data are often cited as the reason that quantitative risk-based inspection is incompatible with the sparse and imperfect information that is typically available to structural integrity engineers. Current industrial guidance is also limited in its methods of distinguishing quality of inspections, as this is typically based on simplified (qualitative) heuristics. In this paper, Bayesian multi-level (partial pooling) models are proposed as a flexible and transparent method of combining imperfect and incomplete information, to support decision-making regarding the integrity management of in-service structures. This work builds on the established theoretical framework for computing the expected value of information, by allowing for partial pooling between inspection measurements (or groups of measurements). This method is demonstrated for a simulated example of a structure with active corrosion in multiple locations, which acknowledges that the data will be associated with some precision, bias, and reliability. Quantifying the extent to which an inspection of one location can reduce uncertainty in damage models at remote locations has been shown to influence many aspects of the expected value of an inspection. These results are considered in the context of the current challenges in risk based structural integrity management.

Hyperbaric exposure (1.3 atmospheres absolute with 20.9% O2) for 1 h a day was shown to improve exercise capacity. The present study was designed to reveal whether the daily exposure time affects exercise performance and metabolism in skeletal and cardiac muscles. Male mice in the training group were housed in a cage with a wheel activity device for 7 weeks from 5 weeks old. Trained mice were then subjected to hybrid training (HT, endurance exercise for 30 min followed by sprint interval exercise for 30 min). Hyperbaric exposure was applied following daily HT for 15 min (15HT), 30 min (30HT), or 60 min (60HT) for 4 weeks. In the endurance capacity test, maximal work values were significantly increased by 30HT and 60HT. In the left ventricle (LV), activity levels of 3‐hydroxyacyl‐CoA‐dehydrogenase, citrate synthase, and carnitine palmitoyl transferase (CPT) 2 were significantly increased by 60HT. CPT2 activity levels were markedly increased by hyperbaric exposure in red gastrocnemius (Gr) and plantaris muscle (PL). Pyruvate dehydrogenase complex activity values in PL were enhanced more by 30HT and 60HT than by HT. Protein levels of N‐terminal isoform of PGC1α (NT‐PGC1α) protein were significantly enhanced in three hyperbaric exposed groups in Gr, but not in LV. These results indicate that hyperbaric exposure for 30 min or longer has beneficial effects on endurance, and 60‐min exposure has the potential to further increase performance by facilitating fatty acid metabolism in skeletal and cardiac muscles in highly trained mice. NT‐PGC1α may have important roles for these adaptations in skeletal muscle. [ABSTRACT FROM AUTHOR]

A shortcoming of the typical use of the frequentist probabilistic approach in reliability analysis of limit state equations that appear in the geotechnical literature is that estimates of the uncertainty in reliability index (probability of failure) are not made. A Bayesian approach is demonstrated in the paper to overcome this shortcoming using the example of the pullout limit state for internal stability of mechanically stabilised earth (MSE) walls. The paper shows that in the Bayesian context, reliability index or probability of failure are modelled as probability distributions, and therefore readily incorporate the uncertainty in their estimates. The general approach provides a tool to communicate the level of risk to stakeholders that follows from the choice of design method and guidance to reduce uncertainty on the estimate of reliability. For practical purposes, the predictive reliability index may be used as an alternative measure of margin of safety that takes this uncertainty into account. [ABSTRACT FROM AUTHOR]

This paper explores how the development of digital solutions for communication and daily interaction between government and its citizens influences citizens' satisfaction, trust and perceived importance of the information received from government. It illuminates this effect by drawing on a survey-experimental design in a Danish research setting. With digitization of public services happening so quickly, it leaves the question of how this transformation is actually viewed by the citizens. The change in medium from traditional communication by standard mail to digital communication can be argued to have both positive and negative effects. Some citizens might find it easy to rely on digital communication, others might perceive digital solutions as a challenge and an obstacle when receiving essential information, which again might foster greater dissatisfaction. Results of Bayesian statistical analysis suggest that the digitization of communication form has little effect on citizens' trust and satisfaction with the information received. Further, results do indicate that the perceived importance of the information received is lower for information received digitally. [ABSTRACT FROM AUTHOR]

Hopelessness is an important symptom of adolescent depression, being associated with both risk of suicide and poor treatment response, but predictors of hopelessness are understudied. Conflict with parents is common in adolescent depression, but parents and adolescents often disagree when reporting conflict severity. Discrepancy in reporting may be an indicator of the parent–adolescent dyad lacking a shared representation of the state of their relationship. This could make conflicts seem unresolvable to the adolescent, leading to expectations of persistent stress and lack of support, increasing hopelessness. This study employed latent difference scores, ordinal regression and cross-validation to evaluate the hypothesis that discrepancy in report of parent–adolescent conflict would predict hopelessness among depressed adolescents. Parents reporting less conflict than the adolescent was associated with increased adolescent hopelessness, giving preliminary support to the hypothesis. [ABSTRACT FROM AUTHOR]

Attempts to formalize inspection and monitoring strategies in industry have struggled to combine evidence from multiple sources (including subject matter expertise) in a mathematically coherent way. The perceived requirement for large amounts of data are often cited as the reason that quantitative risk-based inspection is incompatible with the sparse and imperfect information that is typically available to structural integrity engineers. Current industrial guidance is also limited in its methods of distinguishing quality of inspections, as this is typically based on simplified (qualitative) heuristics. In this paper, Bayesian multi-level (partial pooling) models are proposed as a flexible and transparent method of combining imperfect and incomplete information, to support decision-making regarding the integrity management of in-service structures. This work builds on the established theoretical framework for computing the expected value of information, by allowing for partial pooling between inspection measurements (or groups of measurements). This method is demonstrated for a simulated example of a structure with active corrosion in multiple locations, which acknowledges that the data will be associated with some precision, bias, and reliability. Quantifying the extent to which an inspection of one location can reduce uncertainty in damage models at remote locations has been shown to influence many aspects of the expected value of an inspection. These results are considered in the context of the current challenges in risk based structural integrity management. [ABSTRACT FROM AUTHOR]

Purpose: The purpose of this study is estimating and comparing the three different dimensions of the EEG and studying the trials variability for two auditory and visually oddball tasks in the healthy subjects. They include regional as the region of the brain, longitudinal as the repetition of the stimuli, and functional as whole curve of Evoked Related Potential (ERP), dimensions. Materials and Methods: The sample size is seventeen, with six females, in this three-trial study with standard and target stimuli per task. The dataset was downloaded from the internet and preprocessed. The Hybrid Principal Component Analysis (HPCA) decomposed the ERPs and estimated eigen components of three dimensions. The 95% Bayesian credible sets and trial effects as random effects of the first eigen component of each dimensions studied with the Generalized Additive Mixed Model (GAMM). Results: The p-values of the interaction effects between time and stimuli, repeats and stimuli and regions and stimuli are 0.05. The p-value of trial effects are

Breast cancer patients with metastatic disease have a higher incidence of deaths from breast cancer than patients with early-stage cancers. Recent findings suggest that there are differences in immune cell function between metastatic and non-metastatic cases, even years before diagnosis. We have analyzed whole blood gene expression by Illumina bead chips in blood samples taken using the PAXgene blood collection system up to two years before diagnosis. The final study sample included 197 breast cancer cases and 197 age-matched controls. We defined a causal directed acyclic graph to guide a Bayesian data analysis to estimate the risk of metastasis associated with the expression of all genes and with relevant sets of genes. We ranked genes and gene sets according to the sign probability for excess risk. Among the screening detected cancers, 82% were without metastasis, compared to 53% of between-screening detected cancers. Among the highest ranking genes and gene sets associated with metastasis risk, we identified plasmacytiod dentritic cell function, the SLC22 family of transporters, and glutamine metabolism as potential links between the immune system and metastasis. We conclude that there may be potentially wide-reaching differences in blood gene expression profiles between metastatic and non-metastatic breast cancer cases up to two years before diagnosis, which warrants future study.