Bayesian Mixed-effects Model Research Articles

Public Access for Pheasant Hunters: Understanding an Emerging Need

ABSTRACTRing‐necked pheasant (Phasianus colchicus; i.e., pheasant) hunting participation is declining across North America, reflecting a larger downward trend in American hunting participation and threatening benefits to grassland conservation and rural economies. To stabilize and expand the pheasant hunting population, we must first identify factors that influence pheasant hunter participation. We used an extensive in‐person hunter survey to test the hypothesis that hunter demographics interact with social‐ecological traits of hunting locations to affect hunter decisions, outcomes, and perceptions. We built a series of Bayesian mixed effects models to parse variation in demographics, perceptions, and hunt outcomes of pheasant hunters interviewed at public access hunting sites across 3 regions in Nebraska, USA, that varied in pheasant abundance and proximity to urban population centers. Among pheasant hunters in Nebraska, access to private lands was negatively related to the human population density of a pheasant hunter's home ZIP code and the distance a hunter had traveled to reach a hunting location. Pheasant hunters interviewed closer to metropolitan areas tended to be more urban and travel shorter distances, and their parties were more likely to include youth but less likely to include dogs. Hunter satisfaction was positively associated with seeing and harvesting pheasants and hunting with youth. Whereas youth participation and the number of pheasants seen varied by study region, hunter satisfaction did not differ across regions, suggesting that hunters may calibrate their expectations and build their parties based on where they plan to hunt. The variation in hunter demographics across hunting locations and disconnects between social and ecological correlates of hunter satisfaction suggests that diverse pheasant hunting constituencies will be best served by diverse pheasant hunting opportunities. © 2019 The Wildlife Society.

Partitioning the risk of tuberculosis transmission in household contact studies.

Household contact studies of tuberculosis (TB) are a common way to study disease transmission dynamics. However these studies lack a mechanism for accounting for community transmission, which is known to be significant, particularly in high burden settings. We illustrate a statistical approach for estimating both the correlates with transmission of TB in a household setting and the probability of community transmission using a modified Bayesian mixed-effects model. This is applied to two household contact studies in Vitória, Brazil from 2008–2013 and Kampala, Uganda from 1995–2004 that enrolled households with an individual that was recently diagnosed with pulmonary TB. We estimate the probability of community transmission to be higher in Uganda (ranging from 0.21 to 0.69, depending on HHC age and HIV status of the index case) than in Brazil (ranging from 0.13 for young children to 0.50 in adults). These estimates are consistent with a higher overall burden of disease in Uganda compared to Brazil. Our method also estimates an increasing risk of community-acquired TB with age of the household contact, consistent with existing literature. This approach is a useful way to integrate the role of the community in understanding TB disease transmission dynamics in household contact studies.

Speed-Accuracy Tradeoffs in Brain and Behavior: Testing the Independence of P300 and N400 Related Processes in Behavioral Responses to Sentence Categorization.

Although the N400 was originally discovered in a paradigm designed to elicit a P300 (Kutas and Hillyard, 1980), its relationship with the P300 and how both overlapping event-related potentials (ERPs) determine behavioral profiles is still elusive. Here we conducted an ERP (N = 20) and a multiple-response speed-accuracy tradeoff (SAT) experiment (N = 16) on distinct participant samples using an antonym paradigm (The opposite of black is white/nice/yellow with acceptability judgment). We hypothesized that SAT profiles incorporate processes of task-related decision-making (P300) and stimulus-related expectation violation (N400). We replicated previous ERP results (Roehm et al., 2007): in the correct condition (white), the expected target elicits a P300, while both expectation violations engender an N400 [reduced for related (yellow) vs. unrelated targets (nice)]. Using multivariate Bayesian mixed-effects models, we modeled the P300 and N400 responses simultaneously and found that correlation between residuals and subject-level random effects of each response window was minimal, suggesting that the components are largely independent. For the SAT data, we found that antonyms and unrelated targets had a similar slope (rate of increase in accuracy over time) and an asymptote at ceiling, while related targets showed both a lower slope and a lower asymptote, reaching only approximately 80% accuracy. Using a GLMM-based approach (Davidson and Martin, 2013), we modeled these dynamics using response time and condition as predictors. Replacing the predictor for condition with the averaged P300 and N400 amplitudes from the ERP experiment, we achieved identical model performance. We then examined the piecewise contribution of the P300 and N400 amplitudes with partial effects (see Hohenstein and Kliegl, 2015). Unsurprisingly, the P300 amplitude was the strongest contributor to the SAT-curve in the antonym condition and the N400 was the strongest contributor in the unrelated condition. In brief, this is the first demonstration of how overlapping ERP responses in one sample of participants predict behavioral SAT profiles of another sample. The P300 and N400 reflect two independent but interacting processes and the competition between these processes is reflected differently in behavioral parameters of speed and accuracy.

Combining multiple data sets to unravel the spatiotemporal dynamics of a data-limited fish stock

The biological status of many commercially exploited fishes remains unknown, mostly due to a lack of data necessary for their assessment. Investigating the spatiotemporal dynamics of such species can lead to new insights into population processes and foster a path towards improved spatial management decisions. Here, we focused on striped red mullet (Mullus surmuletus), a widespread yet data-limited species of high commercial importance. Aiming to quantify range dynamics in this data-poor scenario, we combined fishery-dependent and -independent data sets through a series of Bayesian mixed-effects models designed to capture monthly and seasonal occurrence patterns near the species’ northern range limit across 20 years. Combining multiple data sets allowed us to cover the entire distribution of the northern population of M. surmuletus, exploring dynamics at different spatiotemporal scales and identifying key environmental drivers (i.e., sea surface temperature, salinity) that shape occurrence patterns. Our results demonstrate that even when process and (or) observation uncertainty is high, or when data are sparse, if we combine multiple data sets within a hierarchical modelling framework, accurate and useful spatial predictions can still be made.

Herd and environmental determinants of reproductive performance in Swedish dairy herds, 2001–2009

This was a retrospective cohort study of Swedish dairy herds. Summary measures of production and reproductive performance, details of soil, moss mineral concentrations, and temperature and rainfall measurements at each herd location were available for the period September 2001 to August 2009. A Bayesian mixed-effects regression model including spatial and non-spatial heterogeneity terms was developed to quantify associations between hypothesised explanatory variables and mean herd breeding interval, defined as the difference between mean calving to last service interval and mean calving to first service interval for each fiscal year.Mean herd breeding intervals were shorter in herds with greater than 80% Swedish Red Cattle, herds with lower mean age at first calving, herds comprised of older cows and in larger herds. None of the soil composition or moss mineral concentration estimates were associated with mean herd breeding interval and the effect of temperature and rainfall on mean herd breeding interval was small. We conclude that environmental conditions (soil composition, moss mineral concentrations, environmental temperature and rainfall) had relatively minor effects on dairy herd reproductive performance in Sweden between 2001 and 2009.

HIV-Associated TB Among Children and Adolescents in TB/HIV High Burden Settings: Trends and Risk Factors in the Era of Increasing ART Coverage

Background: Children and adolescents living with human immune-deficiency virus (C/ALHIV) have increased tuberculosis (TB) risk compared to HIV-negative peers. Anti-retroviral therapy (ART) reduces TB risk among adults living with HIV, but data is limited in C/ALHIV. Methods: Patient data from seven pediatric HIV centers that adhere to national guidelines in Botswana, Eswatini, Lesotho, Malawi, Tanzania, and Uganda were analyzed. Standardized HIV and TB classification and outcomes were reported following WHO guidelines. A Bayesian mixed effects model assessed association between ART coverage and TB period prevalence. Findings: The analytic period (2013-2017) encompassed 57,525 patient-years and captured 1,217 cases of HIV-associated child and adolescent TB. Favorable TB outcomes were associated with increased length of time in care and early ART initiation. Severe immunosuppression at TB diagnosis was associated with death. The TB incidence was 2,017/100,000 patient-years. With every 10% increase in C/ALHIV receiving ART, there was a 2.33% (95% Credible interval; 0.58%-4.40%) reduction in TB prevalence. Clinics with lower TB prevalence experienced less dramatic declines in TB incidence. Interpretation: These findings support continued need to advocate for policies and implement practices that optimize integrated pediatric and adolescent HIV and TB services. Although ART use is associated with decreased TB prevalence among C/ALHIV, novel TB interventions are likely required to achieve TB elimination and HIV epidemic control in TB/HIV high-burden settings. Funding Statement: This material has been supported by the U.S. President’s Emergency Plan for AIDS Relief through the United States Agency for International Development (USAID) under the terms of USAID Award Number AID-674-A-16-00003. In addition, funds were leveraged from Texas Children’s Hospital to support completion of this research and dissemination of findings. Declaration of Interests: The authors declare no competing interests. Ethics Approval Statement: No individual participant consent was required for this retrospective analysis of programmatic data. Approval was obtained from all necessary ethical bodies in each country including the Baylor College of Medicine Children’s Foundation or Trust, the national ethics committees in each country, and the Baylor College of Medicine Institutional Review Board.

MIMIX: A Bayesian Mixed-Effects Model for Microbiome Data From Designed Experiments

Recent advances in bioinformatics have made high-throughput microbiome data widely available, and new statistical tools are required to maximize the information gained from these data. For example, analysis of high-dimensional microbiome data from designed experiments remains an open area in microbiome research. Contemporary analyses work on metrics that summarize collective properties of the microbiome, but such reductions preclude inference on the fine-scale effects of environmental stimuli on individual microbial taxa. Other approaches model the proportions or counts of individual taxa as response variables in mixed models, but these methods fail to account for complex correlation patterns among microbial communities. In this article, we propose a novel Bayesian mixed-effects model that exploits cross-taxa correlations within the microbiome, a model we call microbiome mixed model (MIMIX). MIMIX offers global tests for treatment effects, local tests and estimation of treatment effects on individual taxa, quantification of the relative contribution from heterogeneous sources to microbiome variability, and identification of latent ecological subcommunities in the microbiome. MIMIX is tailored to large microbiome experiments using a combination of Bayesian factor analysis to efficiently represent dependence between taxa and Bayesian variable selection methods to achieve sparsity. We demonstrate the model using a simulation experiment and on a 2 × 2 factorial experiment of the effects of nutrient supplement and herbivore exclusion on the foliar fungal microbiome of Andropogon gerardii, a perennial bunchgrass, as part of the global Nutrient Network research initiative. Supplementary materials for this article, including a standardized description of the materials available for reproducing the work, are available as an online supplement.

Positional variation rather than salt stress dominates changes in three-dimensional leaf shape patterns in cucumber canopies

Abstract Leaf shape plays a key role in the interaction of a plant with its environment, best-known in the plant’s light harvest. Effects of the environment on the interplay of canopy architecture and physiological functioning can be estimated using functional-structural plant models (FSPMs). In order to reduce the complexity of canopy simulations, leaf shape models used in FSPMs are often simple prototypes scaled to match current leaf area. L-Cucumber is such an FSPM, whose leaf prototype mimics average real leaf shape of unstressed cucumber plants well. However, adaptation processes or stress responses may lead to non-proportional changes in leaf geometries, which, for example, could affect length to width ratios or curvatures. The current leaf shape model in L-Cucumber is static and hence does not incorporate changes in leaf shape within or between plants. Thus, the aim of this study was to estimate leaf shape variation and exemplarily study its effects on FSPM simulations. Three-dimensional leaf coordinate data from a salt stress study were analysed with a robust Bayesian mixed-effects model for estimating leaf shape depending on rank, size and salinity. Results showed that positional and size variation rather than salinity levels dominated 3D leaf shape patterns of cucumber. Considering variable leaf shapes in relation to this main sources of variation in L-Cucumber simulations, only minor effects compared to a realistic, yet static average shape were found. However, with similar computational demands variation in shapes other studies highly sensitive to shape dynamics, for example, pesticide spraying might be affected more strongly.

Attention, memory, and executive functions profile in a prospective cohort of patients with malignant glioma

A prospective cohort study was structured to examine patient’s tumor mass factors that could determine a specific cognitive function profile in patients with Malignant Glioma who survived after the Stupp protocol, while adjusting for previously explored confounding factors. Neuropsi - Attention and Memory (Second edition) was applied after tumor resection, but before any other oncological treatment. Follow-up neuropsychological evaluation was conducted when patients completed the Stupp protocol treatment scheme. To estimate the mean difference between follow-up and baseline Neuropsi subtest scores, a Bayesian Mixed Effects Model was fitted with three random-effect regressors representing tumor locations, co-administered medications and by-patient cognitive performance variability. Sixteen patients were included. Temporal lobe tumors were the most frequent (31.2%) and more than three-fourths of those were located in the left lobe; such results agree with the clinical and statistical significance that supports an impairment in the Semantic Verbal Fluency subtest (Mean score change = −5.97, 95% Credible Interval = [−8.01, −3.80]). The findings suggest that any potential variants in a determined cognitive dysfunction profile could be more influenced by tumor location than any other variable related to the patient’s tumor mass or administered oncological medications; however, because the limited sample representativeness, additional studies are required.

Getting more from heterogeneous HIV-1 surveillance data in a high immigration country: estimation of incidence and undiagnosed population size using multiple biomarkers.

BackgroundMost HIV infections originate from individuals who are undiagnosed and unaware of their infection. Estimation of this quantity from surveillance data is hard because there is incomplete knowledge about (i) the time between infection and diagnosis (TI) for the general population, and (ii) the time between immigration and diagnosis for foreign-born persons.MethodsWe developed a new statistical method for estimating the incidence of HIV-1 and the number of undiagnosed people living with HIV (PLHIV), based on dynamic modelling of heterogeneous HIV-1 surveillance data. The methods consist of a Bayesian non-linear mixed effects model using multiple biomarkers to estimate TI of HIV-1-positive individuals, and a novel incidence estimator which distinguishes between endogenous and exogenous infections by modelling explicitly the probability that a foreign-born person was infected either before or after immigration. The incidence estimator allows for direct calculation of the number of undiagnosed persons. The new methodology is illustrated combining heterogeneous surveillance data from Sweden between 2003 and 2015.ResultsA leave-one-out cross-validation study showed that the multiple-biomarker model was more accurate than single biomarkers (mean absolute error 1.01 vs ≥1.95). We estimate that 816 [95% credible interval (CI) 775-865] PLHIV were undiagnosed in 2015, representing a proportion of 10.8% (95% CI 10.3-11.4%) of all PLHIV.ConclusionsThe proposed methodology will enhance the utility of standard surveillance data streams and will be useful to monitor progress towards and compliance with the 90–90-90 UNAIDS target.

Realistic loss of plant species diversity decreases soil quality in a Tibetan alpine meadow

With rapid changes in climate conditions and grazing regimes, a decline in soil quality has become a main challenge for sustainability of Tibetan rangeland. Increasing evidence highlights the significance of plant biomass, but few studies test the role of realistic loss of plant diversity in impacting soil quality. We hypothesized that loss of species diversity will decrease soil quality, not only through lost standing biomass (i.e., “mass effect”) but also directly via the loss of the “diversity effect”. Considering the realistic loss of species diversity following overgrazing, grazing exclusion and climate change, we experimentally removed grass, sedge and various forb species from a Tibetan alpine meadow to test this hypothesis. After target plant species had been removed for five years, we examined the effect of the species removal on the plant community and soil quality with Bayesian linear mixed effect models and inferred potentially causal linkages between plants and soil with a Bayesian network. Results showed (i) overall, plant species removal not only significantly decreased plant diversity and standing biomass but also decreased soil organic matter and nutrients; (ii) the standing biomass decreased more pronouncedly in response to the removal of sedge and grass species, but plant diversity and soil nutrients decreased more pronouncedly in response to the removal of various forb species; and (iii) the loss of species diversity rather than standing biomass was directly associated with the decline in the soil quality, especially a decrease in soil available P. In short, our results indicate that realistic loss of plant species diversity can decrease soil quality in a Tibetan alpine meadow through lost the “diversity effect”. The functional significance of plant species diversity in maintaining soil quality supports the expectation of local pastoralists that diverse plants, promoted by traditionally rotational grazing, can ensure the sustainability of alpine soil and Tibetan rangeland.

Identifying ecological drivers of interspecific variation in song complexity in songbirds (Passeriformes, Passeri)

A diversity of selective pressures and stochastic processes have likely created substantial variation in song structure, creating difficulties in quantifying the influence of specific ecological factors. This problem is further compounded by differences in study taxa and methods of data analysis between studies. Large comparative studies offer the potential to mitigate some of these methodological difficulties by maximizing the power of statistical analyses and minimizing the probability of misidentifying the magnitude and direction of relationships between independent and dependent variables. In this study, we quantified song complexity for 367 species of globally distributed songbirds (Passeriformes, Passeri). We quantified eight individual acoustic variables that have previously been linked to audio complexity which we analyzed independently, and after applying multivariate statistics to the variables. We used Bayesian linear mixed effect models to test multiple hypotheses regarding song complexity: that it should be greater in open habitats, in migratory species, for sexually monomorphic species, at higher latitudes and altitudes, and that it should co‐vary with clutch size characteristics. Our results challenge perceptions of the effect of habitat structure on song complexity; for instance, counter to expectation, we found songs in closed environments to have reduced syllable diversity. Additionally, our results suggest song complexity may not be ubiquitously a means of communicating male quality, with no significant difference between recordings from monomorphic and dimorphic species. By estimating song complexity in multiple ways, and quantifying these over large taxonomic and spatial scales, we are able to gain a more nuanced understanding of how song complexity is potentially affected by a range of biotic and abiotic factors. Our results also suggest that caution is required when making generalized statements about the relative influence of different factors on song complexity; more densely‐sampled, group‐specific studies are necessary complements to this taxonomically broad analysis.

Development of a human milk concentrate with human milk lyophilizate for feeding very low birth weight preterm infants: A preclinical experimental study.

Breast milk is considered the gold standard nutritional resource for very low birth weight (VLBW) infants in terms of nutrients and protective factors. If mother's milk is not available, the second choice is donated and fortified human milk (HM) from the Human Milk Bank (HMB). This study hypothesized that HM could be lyophilized and used as an additive to increase the levels of macronutrients and micronutrients available to VLBW infants. This study aimed to constitute a lyophilized HM concentrate and determine the osmolality and the concentration of macronutrients and micronutrients in HM samples at “baseline” and in “HM concentrates”, analyzed immediately (HMCI), and after 3 (HMC3m) and 6 (HMC6m) months of freezing. Osmolality was verified using the freezing point osmometric method. Macronutrient quantification was performed using the MIRIS Human Milk Analyzer. Micronutrients were determined by Flame Atomic Absorption Spectrophotometry and by the automated colorimetric method. Bayesian linear mixed effect models were adjusted using OpenBUGS to estimate mean differences and 95% credibility intervals (CrI) of osmolality and of macro- and micronutrients between the types of HM samples. A comparison of dosage values showed a significant increase between HM baseline and HMCI, HMC3m, and HMC6m. Comparing HM baseline and HMCI highlighted the increase in energy content and the concentration of carbohydrates and total lipids. The Ca and P contents increased and the levels of energy, total lipids, and Cu were reduced in HMC3m compared to HMCI. Ca, Mg, K, Zn, and P increased and the levels of energy, total lipids, and Cu were reduced in HMC6m, compared to HMCI. The present study confirms the possibility of formulation and utilization of the immediate concentrate. Partial stability of HM concentrates generated from freeze-drying of donated milk do not recommend storage.

1762. The Adjusted Ranking Metric (ARM) and Its Use in Composite Measures for HAI Prevention in the National Healthcare Safety Network (NHSN)

BackgroundThe National Healthcare Safety Network (NHSN), developed and used by the Centers for Disease Control and Prevention (CDC) for surveillance of healthcare-associated infections (HAIs), provides benchmark measures, such as standardized infection ratio (SIRs), that CDC and its partners in healthcare and public health use for prevention purposes. NHSN provides benchmarks for each HAI measure separately, but a composite HAI measure could provide a more rounded assessment of HAI problems and prevention opportunities.MethodsSeveral issues must be addressed to produce a sound HAI composite measure, the most of which is that the SIR can be inaccurate for facilities with low HAI exposure (e.g., low device days, operative procedure volume). We remedy this issue with the Adjusted Ranking Metric (ARM), a new measure that reliability-adjusts the SIR using a Bayesian mixed effects model. The ARM is particularly useful in the production of a composite measure because ARMs are well-suited to comparison between facilities. The composite was therefore produced by applying adjustments to the ARMs to account for (1) differences between exposure to separate HAI types within facilities and (2) differences in frequency and severity between HAIs. The composite is calculated for 6 HAIs based on 2015 data.ResultsCase studies of 3 facilities (table) show that the new composite measure provides a meaningful measure of overall facility performance that is less prone to the biases that afflict simple combinations of SIRs.ConclusionWe introduce a framework for calculating a composite HAI measure that is flexible, customizable, and transparent. The current implementation of the framework is intended to assist in prevention efforts and can be easily modified to include cost weights, if desired. Flexibility in weighting the HAIs provides an opportunity for different stakeholders to customize the composite measure to their own needs.Table:Example of SIRs, ARMs for 6 HAIs with Composite Measures for 3 Sample FacilitiesDisclosures All authors: No reported disclosures.

Local adaptation of phenology revealed in outcrosses between spawning segments of a salmonid population

Local adaptation has been demonstrated in spatially or temporally distant animal populations but seldom in proximate populations. To address the scale of local adaptation in Pacific salmon (Oncorhynchus spp.), two generations of hybrids between temporally separated spawning segments were made in a population of pink salmon (O.gorbuscha) and compared with controls to evaluate the genetic architecture underlying adult migration time and to test for declines in marine survival that resulted from outbreeding depression. Bayesian mixed-effects models revealed that adult migration times in hybrid lines were intermediate to those of controls and that additive sources of genetic variation were significant, thereby indicating that local adaptation has acted on additive genetic variation in shaping this trait. Similarly, a line cross analysis revealed that an additive model best described the genetic architecture of adult migration time. In contrast, marine survival was generally similar between control and hybrid lines, which suggested that the effect of outbreeding upon marine survival was minimal at such a fine scale of genetic divergence. The implications of these results are that (a) local adaptation can facilitate genetic divergence of life history traits between proximate subpopulations; (b) artificial relaxation of natural barriers to gene flow can cause maladaptive shifts in life history traits; and (c) wild populations may harbour fine-scale adaptive variation that supports productivity and sustainability.

On the indeterministic nature of star formation on the cloud scale

Molecular clouds are turbulent structures whose star formation efficiency (SFE) is strongly affected by internal stellar feedback processes. In this paper we determine how sensitive the SFE of molecular clouds is to randomised inputs in the star formation feedback loop, and to what extent relationships between emergent cloud properties and the SFE can be recovered. We introduce the yule suite of 26 radiative magnetohydrodynamic (RMHD) simulations of a 10,000 solar mass cloud similar to those in the solar neighbourhood. We use the same initial global properties in every simulation but vary the initial mass function (IMF) sampling and initial cloud velocity structure. The final SFE lies between 6 and 23 percent when either of these parameters are changed. We use Bayesian mixed-effects models to uncover trends in the SFE. The number of photons emitted early in the cluster's life and the length of the cloud provide are the strongest predictors of the SFE. The HII regions evolve following an analytic model of expansion into a roughly isothermal density field. The more efficient feedback is at evaporating the cloud, the less the star cluster is dispersed. We argue that this is because if the gas is evaporated slowly, the stars are dragged outwards towards surviving gas clumps due to the gravitational attraction between the stars and gas. While star formation and feedback efficiencies are dependent on nonlinear processes, statistical models describing cloud-scale processes can be constructed.

Abstract A03: Evaluating health disparities in prostate cancer: A neighborhood-wide association study (NWAS) in African Americans

Abstract Background: African American (AA) men are more than twice as likely to be diagnosed with and die of prostate cancer (PCa) compared to White men. Risk factors from social influences to biology could be contributing to these disparities. However, empiric methods to assess the role of social or neighborhood factors in PCa are limited. In a previous study, we developed a neighborhood-wide association study (NWAS) approach, which is analogous to a genome-wide association study (GWAS), and empirically identified neighborhood characteristics from the U.S. Census associated with PCa in White men. In this study, we conducted an NWAS in AA men and preliminarily compared neighborhood and PCa effects by race. Methods: Pennsylvania State Cancer Registry data from 1995-2005 were linked to the 2000 U.S. Census data. In a successively more stringent multiphase approach, we evaluated the association between neighborhood exposures (n = 14,663, standard deviation adjusted census variables) and prostate cancer aggressiveness (n = 911 aggressive Stage>3/Gleason grade >7 cases vs. n = 8632 non-aggressive Stage<3/Gleason grade<7 controls) in AA men. Specifically, we used generalized estimating equations in Phase 1 and Bayesian mixed effects models in Phase 2 to calculate odds ratios (OR) and credible intervals (CI), adjusting for age, year of diagnosis, spatial correlation, and multiple testing. Variables meeting statistically significant thresholds were identified and used in subsequent phases. In Phase 3, principal components analysis grouped together correlated variables, and the most significant variable within each component was identified as the study top hit. Results: Seven variables were most significantly associated with prostate cancer aggressiveness in AA men. The top hit related to crowding in renter-occupied housing units (OR = 1.10; CI = 1.001-1.12). In White men, 17 variables were significant, and the top hit related to workers' mode of transportation (OR = 1.05; CI = 1.001-1.09). Similar to the NWAS in White men, significant associations were seen between PCa and neighborhood employment and transportation, but the specific variables used to represent these domains differed by race. Further, neighborhood education was significant in the AA analysis, but not in White men. Conclusions: These preliminary findings suggest that associations between neighborhood circumstances and aggressive PCa may differ by race. Although further testing is needed, NWAS is a novel, empiric approach that could provide new insights into health disparities research. Citation Format: Shannon M. Lynch, Elizabeth Handorf, Craig Newcomb, Charnita Zeigler-Johnson, Timothy R. Rebbeck. Evaluating health disparities in prostate cancer: A neighborhood-wide association study (NWAS) in African Americans [abstract]. In: Proceedings of the Tenth AACR Conference on the Science of Cancer Health Disparities in Racial/Ethnic Minorities and the Medically Underserved; 2017 Sep 25-28; Atlanta, GA. Philadelphia (PA): AACR; Cancer Epidemiol Biomarkers Prev 2018;27(7 Suppl):Abstract nr A03.

Chimpanzees demonstrate individual differences in social information use

Studies of transmission biases in social learning have greatly informed our understanding of how behaviour patterns may diffuse through animal populations, yet within-species inter-individual variation in social information use has received little attention and remains poorly understood. We have addressed this question by examining individual performances across multiple experiments with the same population of primates. We compiled a dataset spanning 16 social learning studies (26 experimental conditions) carried out at the same study site over a 12-year period, incorporating a total of 167 chimpanzees. We applied a binary scoring system to code each participant’s performance in each study according to whether they demonstrated evidence of using social information from conspecifics to solve the experimental task or not (Social Information Score—‘SIS’). Bayesian binomial mixed effects models were then used to estimate the extent to which individual differences influenced SIS, together with any effects of sex, rearing history, age, prior involvement in research and task type on SIS. An estimate of repeatability found that approximately half of the variance in SIS was accounted for by individual identity, indicating that individual differences play a critical role in the social learning behaviour of chimpanzees. According to the model that best fit the data, females were, depending on their rearing history, 15–24% more likely to use social information to solve experimental tasks than males. However, there was no strong evidence of an effect of age or research experience, and pedigree records indicated that SIS was not a strongly heritable trait. Our study offers a novel, transferable method for the study of individual differences in social learning.

Memory and attention recovery in patients with High Grade Glioma who completed the Stupp protocol: A before-after study

ObjectivesAdvances in therapeutic schemes for High Grade Gliomas, such as the introduction of Stupp protocol, have raised interest in elucidating its effects on quality of life and cognitive function. This study aims to examine executive functions, memory and attention in patients with High Grade Gliomas, before initiation and after completion of Stupp protocol. Patients and methodsA before-after study was executed. Patients were selected from a consecutive sample of subjects with High Grade Gliomas who had not initiated Stupp protocol. Neuropsi - Attention and Memory test (Second edition) was employed to evaluate subjects. Inferential analysis was conducted with Bayesian Mixed Effects Model for repeated measures, using a contrast coding scheme for estimating the change between mean neurocognitive scores obtained before and after Stupp protocol. Model was fitted employing random-effect predictors accounting for concomitant administration of bevacizumab and anticonvulsants, and by-subject performance variability. As excluding patients who did not complete Stupp protocol might bias results, complementary analyses were conducted for determining if those were consistent in alternative scenarios. ResultsTwenty-nine patients were included; 17 (58.6%) completed Stupp protocol and both neuropsychological evaluations. Inside Fixed Effects Model, 95% credible intervals for all cognitive functions crossed the zero threshold (No change); however, in all Mixed Effects Models, credible intervals for memory and attention did not cross zero. ConclusionResults showed significant recovery in memory and attention functions after treatment and suggest a confounding role for co-administered treatments and other factors related to subject cognitive performance variability. Nevertheless, these findings need to be corroborated with additional studies including more representative samples.

Divergent trait and environment relationships among parallel radiations in Pelargonium (Geraniaceae): a role for evolutionary legacy?

Functional traits in closely related lineages are expected to vary similarly along common environmental gradients as a result of shared evolutionary and biogeographic history, or legacy effects, and as a result of biophysical tradeoffs in construction. We test these predictions in Pelargonium, a relatively recent evolutionary radiation. Bayesian phylogenetic mixed effects models assessed, at the subclade level, associations between plant height, leaf area, leaf nitrogen content and leaf mass per area (LMA), and five environmental variables capturing temperature and rainfall gradients across the Greater Cape Floristic Region of South Africa. Trait-trait integration was assessed via pairwise correlations within subclades. Of 20 trait-environment associations, 17 differed among subclades. Signs of regression coefficients diverged for height, leaf area and leaf nitrogen content, but not for LMA. Subclades also differed in trait-trait relationships and these differences were modulated by rainfall seasonality. Leave-one-out cross-validation revealed that whether trait variation was better predicted by environmental predictors or trait-trait integration depended on the clade and trait in question. Legacy signals in trait-environment and trait-trait relationships were apparently lost during the earliest diversification of Pelargonium, but then retained during subsequent subclade evolution. Overall, we demonstrate that global-scale patterns are poor predictors of patterns of trait variation at finer geographic and taxonomic scales.