Using ancient sedimentary DNA to forecast ecosystem trajectories under climate change

Ecosystem response to climate change is complex. In order to forecast ecosystem dynamics, we need high-quality data on changes in past species abundance that can inform process-based models. Sedimentary ancient DNA (sedaDNA) has revolutionised our ability to document past ecosystems' dynamics. It provides time series of increased taxonomic resolution compared to microfossils (pollen, spores), and can often give species-level information, especially for past vascular plant and mammal abundances. Time series are much richer in information than contemporary spatial distribution information, which have been traditionally used to train models for predicting biodiversity and ecosystem responses to climate change. Here, we outline the potential contribution of sedaDNA to forecast ecosystem changes. We showcase how species-level time series may allow quantification of the effect of biotic interactions in ecosystem dynamics, and be used to estimate dispersal rates when a dense network of sites is available. By combining palaeo-time series, process-based models, and inverse modelling, we can recover the biotic and abiotic processes underlying ecosystem dynamics, which are traditionally very challenging to characterise. Dynamic models informed by sedaDNA can further be used to extrapolate beyond current dynamics and provide robust forecasts of ecosystem responses to future climate change. This article is part of the theme issue 'Ecological novelty and planetary stewardship: biodiversity dynamics in a transforming biosphere'.

Forecasting seasonal influenza activity in Canada-Comparing seasonal Auto-Regressive integrated moving average and artificial neural network approaches for public health preparedness

INTRODUCTION

Public health preparedness is based on timely and accurate information. Time series forecasting using disease surveillance data is an important aspect of preparedness. This study compared two approaches of time series forecasting: seasonal auto-regressive integrated moving average (SARIMA) modelling and the artificial neural network (ANN) algorithm. The goal was to model weekly seasonal influenza activity in Canada using SARIMA and compares its predictive accuracy, based on root mean square prediction error (RMSE) and mean absolute prediction error (MAE), to that of an ANN.

METHODS

An initial SARIMA model was fit using automated model selection by minimizing the Akaike information criterion (AIC). Further inspection of the autocorrelation function and partial autocorrelation function led to 'manual' model improvements. ANNs were trained iteratively, using an automated process to minimize the RMSE and MAE.

RESULTS

A total of 378, 462 cases of influenza was reported in Canada from the 2010-2011 influenza season to the end of the 2019-2020 influenza season, with an average yearly incidence risk of 20.02 per 100,000 population. Automated SARIMA modelling was the better method in terms of forecasting accuracy (per RMSE and MAE). However, the ANN correctly predicted the peak week of disease incidence while the other models did not.

CONCLUSION

Both the ANN and SARIMA models have shown to be capable tools in forecasting seasonal influenza activity in Canada. It was shown that applying both in tandem is beneficial, SARIMA better forecasted overall incidence while ANN correctly predicted the peak week.

Comparing Subjective Similarity of Automated Driving Styles to Objective Distance-Based Similarity

OBJECTIVE

This study explores subjective and objective driving style similarity to identify how similarity can be used to develop driver-compatible vehicle automation.

BACKGROUND

Similarity in the ways that interaction partners perform tasks can be measured subjectively, through questionnaires, or objectively by characterizing each agent's actions. Although subjective measures have advantages in prediction, objective measures are more useful when operationalizing interventions based on these measures. Showing how objective and subjective similarity are related is therefore prudent for aligning future machine performance with human preferences.

METHODS

A driving simulator study was conducted with stop-and-go scenarios. Participants experienced conservative, moderate, and aggressive automated driving styles and rated the similarity between their own driving style and that of the automation. Objective similarity between the manual and automated driving speed profiles was calculated using three distance measures: dynamic time warping, Euclidean distance, and time alignment measure. Linear mixed effects models were used to examine how different components of the stopping profile and the three objective similarity measures predicted subjective similarity.

RESULTS

Objective similarity using Euclidean distance best predicted subjective similarity. However, this was only observed for participants' approach to the intersection and not their departure.

CONCLUSION

Developing driving styles that drivers perceive to be similar to their own is an important step toward driver-compatible automation. In determining what constitutes similarity, it is important to (a) use measures that reflect the driver's perception of similarity, and (b) understand what elements of the driving style govern subjective similarity.

Using cross-validation methods to select time series models: Promises and pitfalls

Vector autoregressive (VAR) modelling is widely employed in psychology for time series analyses of dynamic processes. However, the typically short time series in psychological studies can lead to overfitting of VAR models, impairing their predictive ability on unseen samples. Cross-validation (CV) methods are commonly recommended for assessing the predictive ability of statistical models. However, it is unclear how the performance of CV is affected by characteristics of time series data and the fitted models. In this simulation study, we examine the ability of two CV methods, namely,10-fold CV and blocked CV, in estimating the prediction errors of three time series models with increasing complexity (person-mean, AR, and VAR), and evaluate how their performance is affected by data characteristics. We then compare these CV methods to the traditional methods using the Akaike (AIC) and Bayesian (BIC) information criteria in their accuracy of selecting the most predictive models. We find that CV methods tend to underestimate prediction errors of simpler models, but overestimate prediction errors of VAR models, particularly when the number of observations is small. Nonetheless, CV methods, especially blocked CV, generally outperform the AIC and BIC. We conclude our study with a discussion on the implications of the findings and provide helpful guidelines for practice.

Inferring the evolution of reproductive isolation in a lineage of fossil threespine stickleback, Gasterosteus doryssus

Darwin attributed the absence of species transitions in the fossil record to his hypothesis that speciation occurs within isolated habitat patches too geographically restricted to be captured by fossil sequences. Mayr's peripatric speciation model added that such speciation would be rapid, further explaining missing evidence of diversification. Indeed, Eldredge and Gould's original punctuated equilibrium model combined Darwin's conjecture, Mayr's model and 124 years of unsuccessfully sampling the fossil record for transitions. Observing such divergence, however, could illustrate the tempo and mode of evolution during early speciation. Here, we investigate peripatric divergence in a Miocene stickleback fish, Gasterosteus doryssus. This lineage appeared and, over approximately 8000 generations, evolved significant reduction of 12 of 16 traits related to armour, swimming and diet, relative to its ancestral population. This was greater morphological divergence than we observed between reproductively isolated, benthic-limnetic ecotypes of extant Gasterosteus aculeatus. Therefore, we infer that reproductive isolation was evolving. However, local extinction of G. doryssus lineages shows how young, isolated, speciating populations often disappear, supporting Darwin's explanation for missing evidence and revealing a mechanism behind morphological stasis. Extinction may also account for limited sustained divergence within the stickleback species complex and help reconcile speciation rate variation observed across time scales.

What time-varying network models based on functional analysis tell us about the course of a patient's problem

Background: Relations among psychological variables are assumed to be complex and to vary over time. Personalized networks can model multivariate complex interactions. The development of time-varying networks allows to model the variation of parameters over time. Objectives: We aimed to determine the value of time-varying networks for clinical practice. Methods: We applied time-varying mixed graphical models (TV-MGM) and time-varying vector autoregressive models (TV-VAR) to intensive longitudinal data of nine participants with depressive symptoms (n = 6) or anxiety (n = 3). Results: Most of the participants showed temporal changes in network topology within the assessment period of 30 days. Time-varying networks of participants with small, medium, and large time variability in edge parameters clearly show the different temporal evolvements of dynamic interactions between variables. The case example indicates clinical utility but also limitations to the application of time-varying networks in clinical practice. Conclusion: Time-varying network models provide a data-driven and exploratory approach that could complement current diagnostic standards by reflecting interacting, often mutually reinforcing processes of mental health problems and by accounting for variation over time. They can be used to generate hypotheses for further confirmatory and clinical testing.

Metacell-based differential expression analysis identifies cell type specific temporal gene response programs in COVID-19 patient PBMCs

By profiling gene expression in individual cells, single-cell RNA-sequencing (scRNA-seq) can resolve cellular heterogeneity and cell-type gene expression dynamics. Its application to time-series samples can identify temporal gene programs active in different cell types, for example, immune cells' responses to viral infection. However, current scRNA-seq analysis has limitations. One is the low number of genes detected per cell. The second is insufficient replicates (often 1-2) due to high experimental cost. The third lies in the data analysis-treating individual cells as independent measurements leads to inflated statistics. To address these, we explore a new computational framework, specifically whether "metacells" constructed to maintain cellular heterogeneity within individual cell types (or clusters) can be used as "replicates" for increasing statistical rigor. Toward this, we applied SEACells to a time-series scRNA-seq dataset from peripheral blood mononuclear cells (PBMCs) after SARS-CoV-2 infection to construct metacells, and used them in maSigPro for quadratic regression to find significantly differentially expressed genes (DEGs) over time, followed by clustering expression velocity trends. We showed that such metacells retained greater expression variances and produced more biologically meaningful DEGs compared to either metacells generated randomly or from simple pseudobulk methods. More specifically, this approach correctly identified the known ISG15 interferon response program in almost all PBMC cell types and many DEGs enriched in the previously defined SARS-CoV-2 infection response pathway. It also uncovered additional and more cell type-specific temporal gene expression programs. Overall, our results demonstrate that the metacell-pseudoreplicate strategy could potentially overcome the limitation of 1-2 replicates.

Anthropogenic particle concentrations and fluxes in an urban river are temporally variable and impacted by storm events

Anthropogenic particles (AP), which include microplastics and other synthetic, semisynthetic, and anthropogenically modified materials, are pollutants of concern in aquatic ecosystems worldwide. Rivers are important conduits and retention sites for AP, and time series data on the movement of these particles in lotic ecosystems are needed to assess the role of rivers in the global AP cycle. Much research assessing AP pollution extrapolates stream loads based on single time point measurements, but lotic ecosystems are highly variable over time (e.g., seasonality and storm events). The accuracy of models describing AP dynamics in rivers is constrained by the limited studies that examine how frequent changes in discharge drive particle retention and transport. This study addressed this knowledge gap by using automated, high-resolution sampling to track AP concentrations and fluxes during multiple storm events in an urban river (Milwaukee River) and comparing these measurements to commonly monitored water quality metrics. AP concentrations and fluxes varied significantly across four storm events, highlighting the temporal variability of AP dynamics. When data from the sampling periods were pooled, there were increases in particle concentration and flux during the early phases of the storms, suggesting that floods may flush AP into the river and/or resuspend particles from the benthic zone. AP flux was closely linked to river discharge, suggesting large loads of AP are delivered downstream during storms. Unexpectedly, AP concentrations were not correlated with other simultaneously measured water quality metrics, including total suspended solids, fecal coliforms, chloride, nitrate, and sulfate, indicating that these metrics cannot be used to estimate AP. These data will contribute to more accurate models of particle dynamics in rivers and global plastic export to oceans. PRACTITIONER POINTS: Anthropogenic particle (AP) concentrations and fluxes in an urban river varied across four storm events. AP concentrations and fluxes were the highest during the early phases of the storms. Storms increased AP transport downstream compared with baseflow. AP concentrations did not correlate with other water quality metrics during storms.

High response diversity and conspecific density-dependence, not species interactions, drive dynamics of coral reef fish communities

Species-to-species and species-to-environment interactions are key drivers of community dynamics. Disentangling these drivers in species-rich assemblages is challenging due to the high number of potentially interacting species (the 'curse of dimensionality'). We develop a process-based model that quantifies how intraspecific and interspecific interactions, and species' covarying responses to environmental fluctuations, jointly drive community dynamics. We fit the model to reef fish abundance time series from 41 reefs of Australia's Great Barrier Reef. We found that fluctuating relative abundances are driven by species' heterogenous responses to environmental fluctuations, whereas interspecific interactions are negligible. Species differences in long-term average abundances are driven by interspecific variation in the magnitudes of both conspecific density-dependence and density-independent growth rates. This study introduces a novel approach to overcoming the curse of dimensionality, which reveals highly individualistic dynamics in coral reef fish communities that imply a high level of niche structure.

The Impact of a Nurse-Led Service on Tobacco Treatment Provision Within a Psychiatric Hospital: A Time Series Study

INTRODUCTION

Psychiatric hospitalization is an opportunity to provide evidence-based tobacco treatment to optimize cessation efforts among people living with mental illnesses (MI). The purpose of this study was to examine the effectiveness of nurse-driven initiatives to enhance tobacco treatment within an inpatient psychiatric setting.

AIMS

We assessed the 4-year impact of implementing a nurse-led tobacco treatment service offered to 11,314 inpatients at admissions in a tobacco-free psychiatric facility in Kentucky.

METHOD

Through a time-series design, we compared the differences in rates of screening for tobacco use and providing treatment from September to December 2015 (prior to implementing the nurse-led tobacco treatment services) to each subsequent year in a 4-year period (2016-2019).

RESULTS

Approximately 60.0% of inpatients were persons using tobacco during the assessment period. Although there were no changes in tobacco use prevalence over the 4-year evaluation duration, there were significant increases in the provision of practical counseling and Food and Drug Administration-approved nicotine replacement therapies for persons using tobacco.

CONCLUSIONS

Our findings support the effectiveness of implementing tobacco treatment programs at the organizational level. Psychiatric hospitalizations provide an opportunity to optimize nurse-driven efforts to deliver tobacco treatment to people with MI. Similar models of nurse-led tobacco treatment services can be adopted within inpatient and other mental and behavioral health settings.

Post-invasion selection acts on standing genetic variation despite a severe founding bottleneck

Invasive populations often have lower genetic diversity relative to the native-range populations from which they derive.1,2 Despite this, many biological invaders succeed in their new environments, in part due to rapid adaptation.3,4,5,6 Therefore, the role of genetic bottlenecks in constraining the adaptation of invaders is debated.7,8,9,10 Here, we use whole-genome resequencing of samples from a 10-year time-series dataset, representing the natural invasion of the Asian honey bee (Apis cerana) in Australia, to investigate natural selection occurring in the aftermath of a founding event. We find that Australia's A. cerana population was founded by as few as one colony, whose arrival was followed by a period of rapid population expansion associated with an increase of rare variants.11 The bottleneck resulted in a steep loss of overall genetic diversity, yet we nevertheless detected loci with signatures of positive selection during the first years post-invasion. When we investigated the origin of alleles under selection, we found that selection acted primarily on the variation introduced by founders and not on the variants that arose post-invasion by mutation. In all, our data highlight that selection on standing genetic variation can occur in the early years post-invasion, even where founding bottlenecks are severe.

Tracking collective emotions in 16 countries during COVID-19: a novel methodology for identifying major emotional events using Twitter

Using messages posted on Twitter, this study develops a new approach to estimating collective emotions (CEs) within countries. It applies time series methodology to develop and demonstrate a novel application of CEs to identify emotional events that are significant at the societal level. The study analyzes over 200 million words from over 10 million Twitter messages posted in 16 countries during the first 120 days of the COVID-19 pandemic. Daily levels of collective anxiety and positive emotions were estimated using Linguistic Inquiry and Word Count's (LIWC) psychologically validated lexicon. The time series estimates of the two collective emotions were analyzed for structural breaks, which mark a significant change in a series due to an external shock. External shocks to collective emotions come from events that are of shared emotional relevance, and this study develops a new approach to identifying them. In the COVID-19 Twitter posts used in the study, analysis of structural breaks showed that in all 16 countries, a reduction in collective anxiety and an increase in positive emotions followed the WHO's declaration of COVID-19 as a global pandemic. Announcements of economic support packages and social restrictions also had similar impacts in some countries. This indicates that the reduction of uncertainty around the evolving COVID-19 situation had a positive emotional impact on people in all the countries in the study. The study contributes to the field of CEs and applied research in collective psychological phenomena.

The Effect of an Electronic Medical Record-Based Clinical Decision Support System on Adherence to Clinical Protocols in Inflammatory Bowel Disease Care: Interrupted Time Series Study

Background

Clinical decision support systems (CDSSs) embedded in electronic medical records (EMRs), also called electronic health records, have the potential to improve the adoption of clinical guidelines. The University of Alberta Inflammatory Bowel Disease (IBD) Group developed a CDSS for patients with IBD who might be experiencing disease flare and deployed it within a clinical information system in 2 continuous time periods.

Objective

This study aims to evaluate the impact of the IBD CDSS on the adherence of health care providers (ie, physicians and nurses) to institutionally agreed clinical management protocols.

Methods

A 2-period interrupted time series (ITS) design, comparing adherence to a clinical flare management protocol during outpatient visits before and after the CDSS implementation, was used. Each interruption was initiated with user training and a memo with instructions for use. A group of 7 physicians, 1 nurse practitioner, and 4 nurses were invited to use the CDSS. In total, 31,726 flare encounters were extracted from the clinical information system database, and 9217 of them were manually screened for inclusion. Each data point in the ITS analysis corresponded to 1 month of individual patient encounters, with a total of 18 months of data (9 before and 9 after interruption) for each period. The study was designed in accordance with the Statement on Reporting of Evaluation Studies in Health Informatics (STARE-HI) guidelines for health informatics evaluations.

Results

Following manual screening, 623 flare encounters were confirmed and designated for ITS analysis. The CDSS was activated in 198 of 623 encounters, most commonly in cases where the primary visit reason was a suspected IBD flare. In Implementation Period 1, before-and-after analysis demonstrates an increase in documentation of clinical scores from 3.5% to 24.1% (P<.001), with a statistically significant level change in ITS analysis (P=.03). In Implementation Period 2, the before-and-after analysis showed further increases in the ordering of acute disease flare lab tests (47.6% to 65.8%; P<.001), including the biomarker fecal calprotectin (27.9% to 37.3%; P=.03) and stool culture testing (54.6% to 66.9%; P=.005); the latter is a test used to distinguish a flare from an infectious disease. There were no significant slope or level changes in ITS analyses in Implementation Period 2. The overall provider adoption rate was moderate at approximately 25%, with greater adoption by nurse providers (used in 30.5% of flare encounters) compared to physicians (used in 6.7% of flare encounters).

Conclusions

This is one of the first studies to investigate the implementation of a CDSS for IBD, designed with a leading EMR software (Epic Systems), providing initial evidence of an improvement over routine care. Several areas for future research were identified, notably the effect of CDSSs on outcomes and how to design a CDSS with greater utility for physicians. CDSSs for IBD should also be evaluated on a larger scale; this can be facilitated by regional and national centralized EMR systems.

Laboratory Data Timeliness and Completeness Improves Following Implementation of an Electronic Laboratory Information System in Côte d'Ivoire: Quasi-Experimental Study on 21 Clinical Laboratories From 2014 to 2020

BACKGROUND

The Ministry of Health in Côte d'Ivoire and the International Training and Education Center for Health at the University of Washington, funded by the United States President's Emergency Plan for AIDS Relief, have been collaborating to develop and implement the Open-Source Enterprise-Level Laboratory Information System (OpenELIS). The system is designed to improve HIV-related laboratory data management and strengthen quality management and capacity at clinical laboratories across the nation.

OBJECTIVE

This evaluation aimed to quantify the effects of implementing OpenELIS on data quality for laboratory tests related to HIV care and treatment.

METHODS

This evaluation used a quasi-experimental design to perform an interrupted time-series analysis to estimate the changes in the level and slope of 3 data quality indicators (timeliness, completeness, and validity) after OpenELIS implementation. We collected paper and electronic records on clusters of differentiation 4 (CD4) testing for 48 weeks before OpenELIS adoption until 72 weeks after. Data collection took place at 21 laboratories in 13 health regions that started using OpenELIS between 2014 and 2020. We analyzed the data at the laboratory level. We estimated odds ratios (ORs) by comparing the observed outcomes with modeled counterfactual ones when the laboratories did not adopt OpenELIS.

RESULTS

There was an immediate 5-fold increase in timeliness (OR 5.27, 95% CI 4.33-6.41; P<.001) and an immediate 3.6-fold increase in completeness (OR 3.59, 95% CI 2.40-5.37; P<.001). These immediate improvements were observed starting after OpenELIS installation and then maintained until 72 weeks after OpenELIS adoption. The weekly improvement in the postimplementation trend of completeness was significant (OR 1.03, 95% CI 1.02-1.05; P<.001). The improvement in validity was not statistically significant (OR 1.34, 95% CI 0.69-2.60; P=.38), but validity did not fall below pre-OpenELIS levels.

CONCLUSIONS

These results demonstrate the value of electronic laboratory information systems in improving laboratory data quality and supporting evidence-based decision-making in health care. These findings highlight the importance of OpenELIS in Côte d'Ivoire and the potential for adoption in other low- and middle-income countries with similar health systems.

A Deep Learning Approach to Analyzing Continuous-Time Cognitive Processes

The dynamics of the mind are complex. Mental processes unfold continuously in time and may be sensitive to a myriad of interacting variables, especially in naturalistic settings. But statistical models used to analyze data from cognitive experiments often assume simplistic dynamics. Recent advances in deep learning have yielded startling improvements to simulations of dynamical cognitive processes, including speech comprehension, visual perception, and goal-directed behavior. But due to poor interpretability, deep learning is generally not used for scientific analysis. Here, we bridge this gap by showing that deep learning can be used, not just to imitate, but to analyze complex processes, providing flexible function approximation while preserving interpretability. To do so, we define and implement a nonlinear regression model in which the probability distribution over the response variable is parameterized by convolving the history of predictors over time using an artificial neural network, thereby allowing the shape and continuous temporal extent of effects to be inferred directly from time series data. Our approach relaxes standard simplifying assumptions (e.g., linearity, stationarity, and homoscedasticity) that are implausible for many cognitive processes and may critically affect the interpretation of data. We demonstrate substantial improvements on behavioral and neuroimaging data from the language processing domain, and we show that our model enables discovery of novel patterns in exploratory analyses, controls for diverse confounds in confirmatory analyses, and opens up research questions in cognitive (neuro)science that are otherwise hard to study.

Late Pleistocene stickleback environmental genomes reveal the chronology of freshwater adaptation

Directly observing the chronology and tempo of adaptation in response to ecological change is rarely possible in natural ecosystems. Sedimentary ancient DNA (sedaDNA) has been shown to be a tractable source of genome-scale data of long-dead organisms1,2,3 and to thereby potentially provide an understanding of the evolutionary histories of past populations.4,5 To date, time series of ecosystem biodiversity have been reconstructed from sedaDNA, typically using DNA metabarcoding or shotgun sequence data generated from less than 1 g of sediment.6,7 Here, we maximize sequence coverage by extracting DNA from ∼50× more sediment per sample than the majority of previous studies1,2,3 to achieve genotype resolution. From a time series of Late Pleistocene sediments spanning from a marine to freshwater ecosystem, we compare adaptive genotypes reconstructed from the environmental genomes of three-spined stickleback at key time points of this transition. We find a staggered temporal dynamic in which freshwater alleles at known loci of large effect in marine-freshwater divergence of three-spined stickleback (e.g., EDA)8 were already established during the brackish phase of the formation of the isolation basin. However, marine alleles were still detected across the majority of marine-freshwater divergence-associated loci, even after the complete isolation of the lake from marine ingression. Our retrospective approach to studying adaptation from environmental genomes of three-spined sticklebacks at the end of the last glacial period complements contemporary experimental approaches9,10,11 and highlights the untapped potential for retrospective "evolve and resequence" natural experiments using sedaDNA.

Influence of N-acetyltransferase 2 polymorphisms and clinical variables on liver function profile of tuberculosis patients

BACKGROUND

Single nucleotide polymorphisms (SNPs) in the N-acetyltransferase 2 (NAT2) gene as well as several other clinical factors can contribute to the elevation of liver function test values in tuberculosis (TB) patients receiving antitubercular therapy (ATT).

RESEARCH DESIGN AND METHODS

A prospective study involving dynamic monitoring of the liver function tests among 130 TB patients from baseline to 98 days post ATT initiation was undertaken to assess the influence of pharmacogenomic and clinical variables on the elevation of liver function test values. Genomic DNA was extracted from serum samples for the assessment of NAT2 SNPs. Further, within this study population, we conducted a case control study to identify the odds of developing ATT-induced drug-induced liver injury (DILI) based on NAT2 SNPs, genotype and phenotype, and clinical variables.

RESULTS

NAT2 slow acetylators had higher mean [90%CI] liver function test values for 8-28 days post ATT and higher odds of developing DILI (OR: 2.73, 90%CI: 1.05-7.09) than intermediate acetylators/rapid acetylators.

CONCLUSION

The current study findings provide evidence for closer monitoring among TB patients with specific NAT2 SNPs, genotype and phenotype, and clinical variables, particularly between the period of more than a week to one-month post ATT initiation for better treatment outcomes.

A hybrid neural ordinary differential equation model of the cardiovascular system

In the human cardiovascular system (CVS), the interaction between the left and right ventricles of the heart is influenced by the septum and the pericardium. Computational models of the CVS can capture this interaction, but this often involves approximating solutions to complex nonlinear equations numerically. As a result, numerous models have been proposed, where these nonlinear equations are either simplified, or ventricular interaction is ignored. In this work, we propose an alternative approach to modelling ventricular interaction, using a hybrid neural ordinary differential equation (ODE) structure. First, a lumped parameter ODE model of the CVS (including a Newton-Raphson procedure as the numerical solver) is simulated to generate synthetic time-series data. Next, a hybrid neural ODE based on the same model is constructed, where ventricular interaction is instead set to be governed by a neural network. We use a short range of the synthetic data (with various amounts of added measurement noise) to train the hybrid neural ODE model. Symbolic regression is used to convert the neural network into analytic expressions, resulting in a partially learned mechanistic model. This approach was able to recover parsimonious functions with good predictive capabilities and was robust to measurement noise.

Automatic Scoring of Rhizoctonia Crown and Root Rot Affected Sugar Beet Fields from Orthorectified UAV Images Using Machine Learning

Rhizoctonia crown and root rot (RCRR), caused by Rhizoctonia solani, can cause severe yield and quality losses in sugar beet. The most common strategy to control the disease is the development of resistant varieties. In the breeding process, field experiments with artificial inoculation are carried out to evaluate the performance of genotypes and varieties. The phenotyping process in breeding trials requires constant monitoring and scoring by skilled experts. This work is time demanding and shows bias and heterogeneity according to the experience and capacity of each individual person. Optical sensors and artificial intelligence have demonstrated great potential to achieve higher accuracy than human raters and the possibility to standardize phenotyping applications. A workflow combining red-green-blue and multispectral imagery coupled to an unmanned aerial vehicle (UAV), as well as machine learning techniques, was applied to score diseased plants and plots affected by RCRR. Georeferenced annotation of UAV-orthorectified images was carried out. With the annotated images, five convolutional neural networks were trained to score individual plants. The training was carried out with different image analysis strategies and data augmentation. The custom convolutional neural network trained from scratch together with pretrained MobileNet showed the best precision in scoring RCRR (0.73 to 0.85). The average per plot of spectral information was used to score the plots, and the benefit of adding the information obtained from the score of individual plants was compared. For this purpose, machine learning models were trained together with data management strategies, and the best-performing model was chosen. A combined pipeline of random forest and k-nearest neighbors has shown the best weighted precision (0.67). This research provides a reliable workflow for detecting and scoring RCRR based on aerial imagery. RCRR is often distributed heterogeneously in trial plots; therefore, considering the information from individual plants of the plots showed a significant improvement in UAV-based automated monitoring routines.

Foraging habits of Northwest Atlantic hooded seals over the past 30 years: Future habitat suitability under global warming

The Arctic is a global warming 'hot-spot' that is experiencing rapid increases in air and ocean temperatures and concomitant decreases in sea ice cover. These environmental changes are having major consequences on Arctic ecosystems. All Arctic endemic marine mammals are highly dependent on ice-associated ecosystems for at least part of their life cycle and thus are sensitive to the changes occurring in their habitats. Understanding the biological consequences of changes in these environments is essential for ecosystem management and conservation. However, our ability to study climate change impacts on Arctic marine mammals is generally limited by the lack of sufficiently long data time series. In this study, we took advantage of a unique dataset on hooded seal (Cystophora cristata) movements (and serum samples) that spans more than 30 years in the Northwest Atlantic to (i) investigate foraging (distribution and habitat use) and dietary (trophic level of prey and location) habits over the last three decades and (ii) predict future locations of suitable habitat given a projected global warming scenario. We found that, despite a change in isotopic signatures that might suggest prey changes over the 30-year period, hooded seals from the Northwest Atlantic appeared to target similar oceanographic characteristics throughout the study period. However, over decades, they have moved northward to find food. Somewhat surprisingly, foraging habits differed between seals breeding in the Gulf of St Lawrence vs those breeding at the "Front" (off Newfoundland). Seals from the Gulf favoured colder waters while Front seals favoured warmer waters. We predict that foraging habitats for hooded seals will continue to shift northwards and that Front seals are likely to have the greatest resilience. This study shows how hooded seals are responding to rapid environmental change and provides an indication of future trends for the species-information essential for effective ecosystem management and conservation.

Predicting Functional Consequences of Recent Natural Selection in Britain

Ancient DNA can directly reveal the contribution of natural selection to human genomic variation. However, while the analysis of ancient DNA has been successful at identifying genomic signals of selection, inferring the phenotypic consequences of that selection has been more difficult. Most trait-associated variants are noncoding, so we expect that a large proportion of the phenotypic effects of selection will also act through noncoding variation. Since we cannot measure gene expression directly in ancient individuals, we used an approach (Joint-Tissue Imputation [JTI]) developed to predict gene expression from genotype data. We tested for changes in the predicted expression of 17,384 protein coding genes over a time transect of 4,500 years using 91 present-day and 616 ancient individuals from Britain. We identified 28 genes at seven genomic loci with significant (false discovery rate [FDR] < 0.05) changes in predicted expression levels in this time period. We compared the results from our transcriptome-wide scan to a genome-wide scan based on estimating per-single nucleotide polymorphism (SNP) selection coefficients from time series data. At five previously identified loci, our approach allowed us to highlight small numbers of genes with evidence for significant shifts in expression from peaks that in some cases span tens of genes. At two novel loci (SLC44A5 and NUP85), we identify selection on gene expression not captured by scans based on genomic signatures of selection. Finally, we show how classical selection statistics (iHS and SDS) can be combined with JTI models to incorporate functional information into scans that use present-day data alone. These results demonstrate the potential of this type of information to explore both the causes and consequences of natural selection.

Clustering Analysis of Time Series of Affect in Dyadic Interactions

An important goal when analyzing multivariate time series is the identification of heterogeneity, both within and across individuals over time. This heterogeneity can represent different ways in which psychological processes manifest, either between people or within a person across time. In many instances, those differences can have systematic patterns that can be related to future outcomes. In close relationships, for example, the daily exchange of affect between two individuals in a couple can contain a particular structure that is different across people and can result in varying levels of relationship satisfaction. In this paper we use Louvain, a clustering method, as a tool to characterize heterogeneity in multivariate time series data. Using affect measures from dyadic interactions, we first determine that Louvain is adept at detecting homogeneous patterns that are distinct from one another. Additionally, these homogeneous points are linked, at some level, by time. Thus, we find that clustering via Louvain is useful to find time periods of stable, reoccurring patterns. However, using measures founded on information theory reveals that there is some level of information loss that is inevitable when clustering on levels of variable expression. Finally, we evaluate the predictive validity of the clustering method by examining the relation between the identified clusters of affect and measures outside the time series (i.e., relationship satisfaction and breakup taken one and two years later).

The contribution of gene flow, selection, and genetic drift to five thousand years of human allele frequency change

Genomic time series from experimental evolution studies and ancient DNA datasets offer us a chance to directly observe the interplay of various evolutionary forces. We show how the genome-wide variance in allele frequency change between two time points can be decomposed into the contributions of gene flow, genetic drift, and linked selection. In closed populations, the contribution of linked selection is identifiable because it creates covariances between time intervals, and genetic drift does not. However, repeated gene flow between populations can also produce directionality in allele frequency change, creating covariances. We show how to accurately separate the fraction of variance in allele frequency change due to admixture and linked selection in a population receiving gene flow. We use two human ancient DNA datasets, spanning around 5,000 y, as time transects to quantify the contributions to the genome-wide variance in allele frequency change. We find that a large fraction of genome-wide change is due to gene flow. In both cases, after correcting for known major gene flow events, we do not observe a signal of genome-wide linked selection. Thus despite the known role of selection in shaping long-term polymorphism levels, and an increasing number of examples of strong selection on single loci and polygenic scores from ancient DNA, it appears to be gene flow and drift, and not selection, that are the main determinants of recent genome-wide allele frequency change. Our approach should be applicable to the growing number of contemporary and ancient temporal population genomics datasets.

Impact of the Narcotics Information Management System on Opioid Use Among Outpatients With Musculoskeletal and Connective Tissue Disorders: Quasi-Experimental Study Using Interrupted Time Series

BACKGROUND

Opioids have traditionally been used to manage acute or terminal pain. However, their prolonged use has the potential for abuse, misuse, and addiction. South Korea introduced a new health care IT system named the Narcotics Information Management System (NIMS) with the objective of managing all aspects of opioid use, including manufacturing, distribution, sales, disposal, etc.

OBJECTIVE

This study aimed to assess the impact of NIMS on opioid use.

METHODS

We conducted an analysis using national claims data from 45,582 patients diagnosed with musculoskeletal and connective tissue disorders between 2016 and 2020. Our approach included using an interrupted time-series analysis and constructing segmented regression models. Within these models, we considered the primary intervention to be the implementation of NIMS, while we treated the COVID-19 outbreak as the secondary event. To comprehensively assess inappropriate opioid use, we examined 4 key indicators, as established in previous studies: (1) the proportion of patients on high-dose opioid treatment, (2) the proportion of patients receiving opioid prescriptions from multiple providers, (3) the overlap rate of opioid prescriptions per patient, and (4) the naloxone use rate among opioid users.

RESULTS

During the study period, there was a general trend of increasing opioid use. After the implementation of NIMS, significant increases were observed in the trend of the proportion of patients on high-dose opioid treatment (coefficient=0.0271; P=.01) and in the level of the proportion of patients receiving opioid prescriptions from multiple providers (coefficient=0.6252; P=.004). An abrupt decline was seen in the level of the naloxone use rate among opioid users (coefficient=-0.2968; P=.04). While these changes were statistically significant, their clinical significance appears to be minor. No significant changes were observed after both the implementation of NIMS and the COVID-19 outbreak.

CONCLUSIONS

This study suggests that, in its current form, the NIMS may not have brought significant improvements to the identified indicators of opioid overuse and misuse. Additionally, the COVID-19 outbreak exhibited no significant influence on opioid use patterns. The absence of real-time monitoring feature within the NIMS could be a key contributing factor. Further exploration and enhancements are needed to maximize the NIMS' impact on curbing inappropriate opioid use.

Time series prediction of insect pests in tea gardens

BACKGROUND

Tea-garden pest control is crucial to ensure tea quality. In this context, the time-series prediction of insect pests in tea gardens is very important. Deep-learning-based time-series prediction techniques are advancing rapidly but research into their use in tea-garden pest prediction is limited. The current study investigates the time-series prediction of whitefly populations in the Tea Expo Garden, Jurong City, Jiangsu Province, China, employing three deep-learning algorithms, namely Informer, the Long Short-Term Memory (LSTM) network, and LSTM-Attention.

RESULTS

The comparative analysis of the three deep-learning algorithms revealed optimal results for LSTM-Attention, with an average root mean square error (RMSE) of 2.84 and average mean absolute error (MAE) of 2.52 for 7 days' prediction length, respectively. For a prediction length of 3 days, LSTM achieved the best performance, with an average RMSE of 2.60 and an average MAE of 2.24.

CONCLUSION

These findings suggest that different prediction lengths influence model performance in tea garden pest time series prediction. Deep learning could be applied satisfactorily to predict time series of insect pests in tea gardens based on LSTM-Attention. Thus, this study provides a theoretical basis for the research on the time series of pest and disease infestations in tea plants. © 2024 Society of Chemical Industry.

Time-integrated BMP signaling determines fate in a stem cell model for early human development

How paracrine signals are interpreted to yield multiple cell fate decisions in a dynamic context during human development in vivo and in vitro remains poorly understood. Here we report an automated tracking method to follow signaling histories linked to cell fate in large numbers of human pluripotent stem cells (hPSCs). Using an unbiased statistical approach, we discover that measured BMP signaling history correlates strongly with fate in individual cells. We find that BMP response in hPSCs varies more strongly in the duration of signaling than the level. However, both the level and duration of signaling activity control cell fate choices only by changing the time integral. Therefore, signaling duration and level are interchangeable in this context. In a stem cell model for patterning of the human embryo, we show that signaling histories predict the fate pattern and that the integral model correctly predicts changes in cell fate domains when signaling is perturbed. Our data suggest that mechanistically, BMP signaling is integrated by SOX2.

Analysis of the yearly transition function in measles disease modeling

Globally, there were an estimated 9.8 million measles cases and 207 500 measles deaths in 2019. As the effort to eliminate measles around the world continues, modeling remains a valuable tool for public health decision-makers and program implementers. This study presents a novel approach to the use of a yearly transition function that formulates mathematically the vaccine schedules for different age groups while accounting for the effects of the age of vaccination, the timing of vaccination, and disease seasonality on the yearly number of measles cases in a country. The methodology presented adds to an existing modeling framework and expands its analysis, making its utilization more adjustable for the user and contributing to its conceptual clarity. This article also adjusts for the temporal interaction between vaccination and exposure to disease, applying adjustments to estimated yearly counts of cases and the number of vaccines administered that increase population immunity. These new model features provide the ability to forecast and compare the effects of different vaccination timing scenarios and seasonality of transmission on the expected disease incidence. Although the work presented is applied to the example of measles, it has potential relevance to modeling other vaccine-preventable diseases.

Inequalities in access to prenatal care during the COVID-19 pandemic: Analysis of a population-based cohort

BACKGROUND

Before the COVID-19 pandemic, access to prenatal care was lower among some socio-demographic groups. This pandemic caused disruptions to routine preventative care, which could have increased inequalities.

OBJECTIVES

To investigate if the COVID-19 pandemic increased inequalities in access to prenatal care among those who are younger, live in rural areas, have a lower socio-economic situation (SES) and are recent immigrants.

METHODS

We used linked administrative datasets from ICES to identify a population-based cohort of 455,245 deliveries in Ontario from January 2018 to December 2021. Our outcomes were first-trimester prenatal visits, first-trimester ultrasound and adequacy of prenatal care. We used joinpoint analysis to examine outcome time trends and identify trend change points. We stratified analyses by age, rural residence, SES and recent immigration, and examined risk differences (RD) with 95% confidence intervals (CI) between groups at the beginning and end of the study period.

RESULTS

For all outcomes, we noted disruptions to care beginning in March or April 2020 and returning to previous trends by November 2020. Inequalities were stable across groups, except recent immigrants. In July 2017, 65.0% and 69.8% of recent immigrants and non-immigrants, respectively, received ultrasounds in the first trimester (RD -4.8%, 95% CI -8.0, -1.5). By October 2020, this had increased to 75.4%, with no difference with non-immigrants (RD 0.4%, 95% CI -2.4, 3.2). Adequacy of prenatal care showed more intensive care as of November 2020, reflecting a higher number of visits.

CONCLUSIONS

We found no evidence that inequalities between socio-economic groups that existed prior to the pandemic worsened after March 2020. The pandemic may be associated with increased access to care for recent immigrants. The introduction of virtual visits may have resulted in a higher number of prenatal care visits.

Active learning strategies for robotic tactile texture recognition tasks

Accurate texture classification empowers robots to improve their perception and comprehension of the environment, enabling informed decision-making and appropriate responses to diverse materials and surfaces. Still, there are challenges for texture classification regarding the vast amount of time series data generated from robots' sensors. For instance, robots are anticipated to leverage human feedback during interactions with the environment, particularly in cases of misclassification or uncertainty. With the diversity of objects and textures in daily activities, Active Learning (AL) can be employed to minimize the number of samples the robot needs to request from humans, streamlining the learning process. In the present work, we use AL to select the most informative samples for annotation, thus reducing the human labeling effort required to achieve high performance for classifying textures. We also use a sliding window strategy for extracting features from the sensor's time series used in our experiments. Our multi-class dataset (e.g., 12 textures) challenges traditional AL strategies since standard techniques cannot control the number of instances per class selected to be labeled. Therefore, we propose a novel class-balancing instance selection algorithm that we integrate with standard AL strategies. Moreover, we evaluate the effect of sliding windows of two-time intervals (3 and 6 s) on our AL Strategies. Finally, we analyze in our experiments the performance of AL strategies, with and without the balancing algorithm, regarding f1-score, and positive effects are observed in terms of performance when using our proposed data pipeline. Our results show that the training data can be reduced to 70% using an AL strategy regardless of the machine learning model and reach, and in many cases, surpass a baseline performance. Finally, exploring the textures with a 6-s window achieves the best performance, and using either Extra Trees produces an average f1-score of 90.21% in the texture classification data set.

Offering an auto-play feature likely increases total gambling activity at online slot-machines: preliminary evidence from an interrupted time series experiment at a real-life online casino

Auto-play is a ubiquitous feature in online casino gambling and virtual slot machines especially, allowing gamblers to initiate spin sequences of pre-set length and value. While theoretical accounts diverge on the hypothesized causal effect on gambling behavior of using the auto-play feature, observational findings show that this feature is used to a higher degree by problem and/or high-intensity gamblers, suggesting that banning this feature may constitute a global responsible gambling measure. Direct, experimental research on causal effects of offering auto-play at online casinos is however lacking. Here, we report the findings of an interrupted time series experiment, conducted at a real-life online casino in Sweden, in which the auto-play feature was made available during a pre-set duration on 40 online slot machines, with 40 matched slots serving as control. Aggregated time series on daily betted amount, spins and net losses were analyzed using a structural Bayesian framework that compared observed developments during the peri-intervention period to modeled counterfactual estimates. Results suggest that offering an auto-play feature on online casinos likely increases total gambling activity in terms of betted amount (approx.+ 7-9%) and (perhaps) number of spins (approx. +3%) but has no effect on net losses. Limitations of studying auto-play effects on a population-level, as well as the complexities of banning this feature within a complex ecosystem of non-perfect channelization to licensed providers, are discussed, including suggestions for future research.

Introducing Region Based Pooling for handling a varied number of EEG channels for deep learning models

Introduction

A challenge when applying an artificial intelligence (AI) deep learning (DL) approach to novel electroencephalography (EEG) data, is the DL architecture's lack of adaptability to changing numbers of EEG channels. That is, the number of channels cannot vary neither in the training data, nor upon deployment. Such highly specific hardware constraints put major limitations on the clinical usability and scalability of the DL models.

Methods

In this work, we propose a technique for handling such varied numbers of EEG channels by splitting the EEG montages into distinct regions and merge the channels within the same region to a region representation. The solution is termed Region Based Pooling (RBP). The procedure of splitting the montage into regions is performed repeatedly with different region configurations, to minimize potential loss of information. As RBP maps a varied number of EEG channels to a fixed number of region representations, both current and future DL architectures may apply RBP with ease. To demonstrate and evaluate the adequacy of RBP to handle a varied number of EEG channels, sex classification based solely on EEG was used as a test example. The DL models were trained on 129 channels, and tested on 32, 65, and 129-channels versions of the data using the same channel positions scheme. The baselines for comparison were zero-filling the missing channels and applying spherical spline interpolation. The performances were estimated using 5-fold cross validation.

Results

For the 32-channel system version, the mean AUC values across the folds were: RBP (93.34%), spherical spline interpolation (93.36%), and zero-filling (76.82%). Similarly, on the 65-channel system version, the performances were: RBP (93.66%), spherical spline interpolation (93.50%), and zero-filling (85.58%). Finally, the 129-channel system version produced the following results: RBP (94.68%), spherical spline interpolation (93.86%), and zero-filling (91.92%).

Conclusion

In conclusion, RBP obtained similar results to spherical spline interpolation, and superior results to zero-filling. We encourage further research and development of DL models in the cross-dataset setting, including the use of methods such as RBP and spherical spline interpolation to handle a varied number of EEG channels.

Cancer mortality in childhood and adolescence: analysis of trends and spatial distribution in the 133 intermediate Brazilian regions grouped by macroregions

OBJECTIVE

To assess the magnitude, trend, and spatial patterns of childhood and adolescent cancer mortality between 1996 and 2017 in 133 Brazilian intermediate regions by using socioeconomic and healthcare services indicators.

METHODS

This is an ecological study for analyzing the trend of mortality from cancer in childhood and adolescence through time series. Data on deaths were extracted from the Brazilian Mortality Information System. Data on population were extracted from the 1991, 2000, and 2010 demographic censuses of the Brazilian Institute of Geography and Statistics, with interpolation for intercensal years. Time series were delineated for mortality by type of cancer in each intermediate region. Such regions were grouped by macroregions to present the results. The calculation and interpretation of mortality trends use the Prais-Winsten autoregression procedure.

RESULTS

Mortality rates for all neoplasms were higher in the Northern region (7.79 deaths per 100 thousand population), while for leukemias, they were higher in the Southern region (1.61 deaths per 100 thousand population). In both regions, mortality was higher in boys and in the 0-4 age group. The trend was decreasing (annual percent change [APC] - -2.11 [95%CI: -3.14; - 1.30]) for all neoplasms in the Brazilian regions and stationary (APC - -0.43 [95%CI: -1.61; 2.12]) for leukemias in the analyzed period.

CONCLUSION

The mortality rate for all neoplasms showed higher values in regions with smaller numbers of ICU beds in the public healthcare system.

Depression and Anxiety Symptoms Among Cisgender Gay and Bisexual Men During the Onset of the COVID-19 Pandemic: Time Series Analysis of a US National Cohort Study

BACKGROUND

The onset of the COVID-19 pandemic in the United States in March 2020 caused a dramatic change in the way many people lived. Few aspects of daily life were left undisrupted by the pandemic's onset as well as the accompanying policies to control the spread of the disease. Previous research has found that the pandemic may have significantly impacted the mental health of lesbian, gay, bisexual, transgender, and queer (LGBTQ) individuals-potentially more so than other individuals. However, the pandemic did not affect all areas of the United States at the same time, and there may be regional variation in the impact of the onset of the pandemic on depressive symptoms among LGBTQ individuals.

OBJECTIVE

To assess regional variation of the impact of the pandemic, we conducted a time series analysis stratified by US geographic region to examine symptoms of depression and anxiety among a sample of primarily cisgender gay and bisexual men before and after the onset of the COVID-19 pandemic in the United States.

METHODS

In total, 5007 participants completed assessments as part of the Together 5000 study, an ongoing prospective cohort study. Depressive and anxiety symptoms were measured using the Patient Health Questionnaire-4. Patient Health Questionnaire-4 scores were graphed as a function of days from March 15, 2020. Locally estimated scatterplot smoothing trend lines were applied. A sieve-bootstrap Mann-Kendall test for monotonic trend was conducted to assess the presence and direction of trends in the scatterplots. We then compared the observed trends to those observed for 1 year prior (2018-2019) to the pandemic onset using data collected from the same sample.

RESULTS

Significant positive trends were detected for the Northeast (P=.03) and Midwest (P=.01) regions of the United States in the 2020 assessment, indicating that symptoms of anxiety and depression were increasing in the sample in these regions immediately prior to and during the onset of the pandemic. In contrast, these trends were not present in data from the 2018 to 2019 assessment window.

CONCLUSIONS

Symptoms of anxiety and depression increased among the study population in the Northeast and Midwest during the beginning months of the COVID-19 pandemic, but similar increase was not observed in the South and West regions. These trends were also not found for any region in the 2018 to 2019 assessment window. This may indicate region-specific trends in anxiety and depression, potentially driven by the burden of the pandemic and policies that varied from region to region. Future studies should consider geographic variation in COVID-19 spread and policies as well as explore potential mechanisms by which this could influence the mental health of LGBTQ individuals.

Assessing Internet Search Models in Predicting Daily New COVID-19 Cases and Deaths in South Korea

Search data were found to be useful variables for COVID-19 trend prediction. In this study, we aimed to investigate the performance of online search models in state space models (SSMs), linear regression (LR) models, and generalized linear models (GLMs) for South Korean data from January 20, 2020, to July 31, 2021. Principal component analysis (PCA) was run to construct the composite features which were later used in model development. Values of root mean squared error (RMSE), peak day error (PDE), and peak magnitude error (PME) were defined as loss functions. Results showed that integrating search data in the models for short- and long-term prediction resulted in a low level of RMSE values, particularly for SSMs. Findings indicated that type of model used highly impacts the performance of prediction and interpretability of the model. Furthermore, PDE and PME could be beneficial to be included in the evaluation of peaks.

Patterns of Animal Rabies Prevalence in Northern South Africa between 1998 and 2022

Rabies is endemic in South Africa and rabies cycles are maintained in both domestic and wildlife species. The significant number of canine rabies cases reported by the World Organization for Animal Health Reference Laboratory for Rabies at Onderstepoort suggests the need for increased research and mass dog vaccinations on specific targeted foci in the country. This study aimed to investigate the spatiotemporal distribution of animal rabies cases from 1998 to 2017 in northern South Africa and environmental factors associated with highly enzootic municipalities. A descriptive analysis was used to investigate temporal patterns. The Getis-Ord Gi statistical tool was used to exhibit low and high clusters. Logistic regression was used to examine the association between the predictor variables and highly enzootic municipalities. A total of 9580 specimens were submitted for rabies diagnosis between 1998 and 2022. The highest positive case rates were from companion animals (1733 cases, 59.71%), followed by livestock (635 cases, 21.88%) and wildlife (621 cases, 21.39%). Rabies cases were reported throughout the year, with the majority occurring in the mid-dry season. Hot spots were frequently in the northern and eastern parts of Limpopo and Mpumalanga. Thicket bush and grassland were associated with rabies between 1998 and 2002. However, between 2008 and 2012, cultivated commercial crops and waterbodies were associated with rabies occurrence. In the last period, plantations and woodlands were associated with animal rabies. Of the total number of municipalities, five consistently and repeatedly had the highest rabies prevalence rates. These findings suggest that authorities should prioritize resources for those municipalities for rabies elimination and management.

The contribution of gene flow, selection, and genetic drift to five thousand years of human allele frequency change

Genomic time series from experimental evolution studies and ancient DNA datasets offer us a chance to directly observe the interplay of various evolutionary forces. We show how the genome-wide variance in allele frequency change between two time points can be decomposed into the contributions of gene flow, genetic drift, and linked selection. In closed populations, the contribution of linked selection is identifiable because it creates covariances between time intervals, and genetic drift does not. However, repeated gene flow between populations can also produce directionality in allele frequency change, creating covariances. We show how to accurately separate the fraction of variance in allele frequency change due to admixture and linked selection in a population receiving gene flow. We use two human ancient DNA datasets, spanning around 5,000 years, as time transects to quantify the contributions to the genome-wide variance in allele frequency change. We find that a large fraction of genome-wide change is due to gene flow. In both cases, after correcting for known major gene flow events, we do not observe a signal of genome-wide linked selection. Thus despite the known role of selection in shaping long-term polymorphism levels, and an increasing number of examples of strong selection on single loci and polygenic scores from ancient DNA, it appears to be gene flow and drift, and not selection, that are the main determinants of recent genome-wide allele frequency change. Our approach should be applicable to the growing number of contemporary and ancient temporal population genomics datasets.

Quantifying changes in individual-specific template-based representations of center-of-mass dynamics during walking with ankle exoskeletons using Hybrid-SINDy

Ankle exoskeletons alter whole-body walking mechanics, energetics, and stability by altering center-of-mass (CoM) motion. Controlling the dynamics governing CoM motion is, therefore, critical for maintaining efficient and stable gait. However, how CoM dynamics change with ankle exoskeletons is unknown, and how to optimally model individual-specific CoM dynamics, especially in individuals with neurological injuries, remains a challenge. Here, we evaluated individual-specific changes in CoM dynamics in unimpaired adults and one individual with post-stroke hemiparesis while walking in shoes-only and with zero-stiffness and high-stiffness passive ankle exoskeletons. To identify optimal sets of physically interpretable mechanisms describing CoM dynamics, termed template signatures, we leveraged hybrid sparse identification of nonlinear dynamics (Hybrid-SINDy), an equation-free data-driven method for inferring sparse hybrid dynamics from a library of candidate functional forms. In unimpaired adults, Hybrid-SINDy automatically identified spring-loaded inverted pendulum-like template signatures, which did not change with exoskeletons (p > 0.16), except for small changes in leg resting length (p < 0.001). Conversely, post-stroke paretic-leg rotary stiffness mechanisms increased by 37-50% with zero-stiffness exoskeletons. While unimpaired CoM dynamics appear robust to passive ankle exoskeletons, how neurological injuries alter exoskeleton impacts on CoM dynamics merits further investigation. Our findings support Hybrid-SINDy's potential to discover mechanisms describing individual-specific CoM dynamics with assistive devices.

A Time-Varying Dynamic Partial Credit Model to Analyze Polytomous and Multivariate Time Series Data

The accessibility to electronic devices and the novel statistical methodologies available have allowed researchers to comprehend psychological processes at the individual level. However, there are still great challenges to overcome as, in many cases, collected data are more complex than the available models are able to handle. For example, most methods assume that the variables in the time series are measured on an interval scale, which is not the case when Likert-scale items were used. Ignoring the scale of the variables can be problematic and bias the results. Additionally, most methods also assume that the time series are stationary, which is rarely the case. To tackle these disadvantages, we propose a model that combines the partial credit model (PCM) of the item response theory framework and the time-varying autoregressive model (TV-AR), which is a popular model used to study psychological dynamics. The proposed model is referred to as the time-varying dynamic partial credit model (TV-DPCM), which allows to appropriately analyze multivariate polytomous data and nonstationary time series. We test the performance and accuracy of the TV-DPCM in a simulation study. Lastly, by means of an example, we show how to fit the model to empirical data and interpret the results.

Spatial and temporal analysis of juvenile blacktip reef shark (Carcharhinus melanopterus) demographies identifies critical habitats

Reef shark species have undergone sharp declines in recent decades, as they inhabit coastal areas, making them an easy target in fisheries (i.e., sharks are exploited globally for their fins, meat, and liver oil) and exposing them to other threats (e.g., being part of by-catch, pollution, and climate change). Reef sharks play a critical role in coral reef ecosystems, where they control populations of smaller predators and herbivorous fishes either directly via predation or indirectly via behavior, thus protecting biodiversity and preventing potential overgrazing of corals. The urgent need to conserve reef shark populations necessitates a multifaceted approach to policy at local, federal, and global levels. However, monitoring programmes to evaluate the efficiency of such policies are lacking due to the difficulty in repeatedly sampling free-ranging, wild shark populations. Over nine consecutive years, we monitored juveniles of the blacktip reef shark (Carcharhinus melanopterus) population around Moorea, French Polynesia, and within the largest shark sanctuary globally, to date. We investigated the roles of spatial (i.e., sampling sites) and temporal variables (i.e., sampling year, season, and month), water temperature, and interspecific competition on shark density across 10 coastal nursery areas. Juvenile C. melanopterus density was found to be stable over 9 years, which may highlight the effectiveness of local and likely federal policies. Two of the 10 nursery areas exhibited higher juvenile shark densities over time, which may have been related to changes in female reproductive behavior or changes in habitat type and resources. Water temperatures did not affect juvenile shark density over time as extreme temperatures proven lethal (i.e., 33°C) in juvenile C. melanopterus might have been tempered by daily variation. The proven efficiency of time-series datasets for reef sharks to identify critical habitats (having the highest juvenile shark densities over time) should be extended to other populations to significantly contribute to the conservation of reef shark species.

Application of Time-Series Analysis and Expert Judgment in Modeling and Forecasting Blood Donation Trends in Zimbabwe

Background. Blood cannot be artificially manufactured, and there is currently no substitute for human blood. The supply of blood in transfusion facilities requires constant and timely collection of blood from donors. Modeling and forecasting trends in blood collections are critical for determining both the current and future capacity requirements and appropriate models of adequate blood provision. Objectives. The objective of this study is to determine blood collection or donation patterns and develop time-series models that can be updated and refined in predicting future blood donations in Zimbabwe when given the historical data. Materials and Methods. Monthly blood donation data for the period 2009 to 2019 were collected retrospectively from the National Blood Service Zimbabwe database. Time-series models (i.e., the Seasonal Autoregressive Integrated Moving Average [SARIMA] and Error, Trend and Seasonal [ETS]) models were applied and compared. The models were chosen because of their ability to handle the seasonality and other time-series components evident in the blood donation data. Expert opinions and experience were used in selecting the models and in making inferences in the analysis. Results. Time-series plots of blood donations showed seasonal patterns, with significant drops in blood donations in months associated with Zimbabwe's school holidays (April, August, and December) and public holidays. During these holidays, there is a reduced number of school donors, while at about the same time, there is increasing blood demand as a result of road accidents. Model identification procedures established the SARIMA ( 1 , 1 , 2 ) ( 0 , 1 , 1 ) 12 model as the appropriate model for forecasting total blood donation in Zimbabwe. The results and forecasts show an upward trend in blood donations. According to the accuracy measures used, the SARIMA model outperforms the ETS model. Conclusions. Expert knowledge in the blood donation process, coupled with statistical models, can help explain trends exhibited in blood donation data in Zimbabwe. These findings help the blood authorities plan for blood donor campaign drives. The findings are key indicators of where to allocate more resources toward blood donation and when to collect more blood units. The increasing blood donation projections ensure a stable blood bank inventory in the near future.

Highlights

A SARIMA model can be used to predict the flow of blood donations in Zimbabwe.The seasonal blood donation pattern peaks in the months of March, June/July, and September.The donations troughs are in the months of April, August, December, and January. These are the months coinciding with school holidays in Zimbabwe.Both the SARIMA and ETS models provided similar forecasts, but measures of fit and expert knowledge gave a slight preference to the SARIMA ( 1 , 1 , 2 ) ( 0 , 1 , 1 ) 12 model in predicting the flow of blood donations in Zimbabwe.These model results are useful for guiding allocation of blood donation resources and blood donation drive timing.

In-Advance Prediction of Pressure Ulcers via Deep-Learning-Based Robust Missing Value Imputation on Real-Time Intensive Care Variables

Pressure ulcers (PUs) are a prevalent skin disease affecting patients with impaired mobility and in high-risk groups. These ulcers increase patients' suffering, medical expenses, and burden on medical staff. This study introduces a clinical decision support system and verifies it for predicting real-time PU occurrences within the intensive care unit (ICU) by using MIMIC-IV and in-house ICU data. We develop various machine learning (ML) and deep learning (DL) models for predicting PU occurrences in real time using the MIMIC-IV and validate using the MIMIC-IV and Kangwon National University Hospital (KNUH) dataset. To address the challenge of missing values in time series, we propose a novel recurrent neural network model, GRU-D++. This model outperformed other experimental models by achieving the area under the receiver operating characteristic curve (AUROC) of 0.945 for the on-time prediction and AUROC of 0.912 for 48h in-advance prediction. Furthermore, in the external validation with the KNUH dataset, the fine-tuned GRU-D++ model demonstrated superior performances, achieving an AUROC of 0.898 for on-time prediction and an AUROC of 0.897 for 48h in-advance prediction. The proposed GRU-D++, designed to consider temporal information and missing values, stands out for its predictive accuracy. Our findings suggest that this model can significantly alleviate the workload of medical staff and prevent the worsening of patient conditions by enabling timely interventions for PUs in the ICU.

A novel approach for sports injury risk prediction: based on time-series image encoding and deep learning

The rapid development of big data technology and artificial intelligence has provided a new perspective on sports injury prevention. Although data-driven algorithms have achieved some valuable results in the field of sports injury risk assessment, the lack of sufficient generalization of models and the inability to automate feature extraction have made it challenging to deploy research results in the real world. Therefore, this study attempts to build an injury risk prediction model using a combination of time-series image encoding and deep learning algorithms to address this issue better. This study used the time-series image encoding approach for feature construction to represent relationships between values at different moments, including Gramian Angular Summation Field (GASF), Gramian Angular Difference Field (GADF), Markov Transition Field (MTF), and Recurrence Plot (RP). Deep Convolutional Auto-Encoder (DCAE) learned the image-encoded data for representation to obtain features with good discrimination, and the classifier was performed using Deep Neural Network (DNN). The results from five repeated experiments show that the GASF-DCAE-DNN model is overall better in the training (AUC: 0.985 ± 0.001, Gmean: 0.930 ± 0.007, Sensitivity: 0.997 ± 0.003, Specificity: 0.868 ± 0.013) and test sets (AUC: 0.891 ± 0.026, Gmean: 0.830 ± 0.027, Sensitivity: 0.816 ± 0.039, Specificity: 0.845 ± 0.022), with good discriminative power, robustness, and generalization ability. Compared with the best model reported in the literature, the AUC, Gmean, Sensitivity, and Specificity of the GASF-DCAE-DNN model were higher by 23.9%, 27.5%, 39.7%, and 16.2%, respectively, which confirmed the validity and practicability of the model in injury risk prediction. In addition, differences in injury risk patterns between the training and test sets were identified through shapley additivity interpretation. It was also found that the training volume was an essential factor that affected injury risk prediction. The model proposed in this study provides a powerful injury risk prediction tool for future sports injury prevention practice.

Fisher-Shannon Investigation of the Effect of Nonlinearity of Discrete Langevin Model on Behavior of Extremes in Generated Time Series

Diverse forms of nonlinearity within stochastic equations give rise to varying dynamics in processes, which may influence the behavior of extreme values. This study focuses on two nonlinear models of the discrete Langevin equation: one with a fixed diffusion function (M1) and the other with a fixed marginal distribution (M2), both characterized by a nonlinearity parameter. Extremes are defined according to the run theory with thresholds based on percentiles. The behavior of inter-extreme times and run lengths is examined by employing Fisher's Information Measure and the Shannon Entropy. Our findings reveal a clear relationship between the entropic and informational measures and the nonlinearity of model M1-these measures decrease as the nonlinearity parameter increases. Similar relationships are evident for the M2 model, albeit to a lesser extent, even though the background data's marginal distribution remains unaffected by this parameter. As thresholds increase, both the values of Fisher's Information Measure and the Shannon Entropy also increase.

Impact of the Spanish smoke-free laws on cigarette sales by brands, 2000-2021: Evidence from a club convergence approach

INTRODUCTION

In January 2006, the Spanish government enacted a tobacco control law that banned the advertising, promotion and sponsorship of tobacco. In January 2011, further legislation on this matter was adopted to provide a more restrictive specification of the ban. In this study, we analyze the effect produced on cigarette sales by these two prohibitions. We address this problem using a cluster time-series analysis to test whether the sales of cigarettes by brands have been homogenized with the prohibition of advertising, promotion, and sponsorship.

METHODS

The data source used was the official data on legal sales of cigarettes by brands in Spain, from January 2005 to December 2021 (excluding the Canary Islands and the Autonomous Communities of the cities of Ceuta and Melilla). To achieve our objective, we used log(t) test statistics to check if there is global convergence in the three selected periods according to the regulatory changes that have occurred in Spain (2005-2021, 2005-2010 and 2011-2021). Second, once absolute convergence is rejected, we applied a clustering algorithm to test for the existence of subgroup convergence.

RESULTS

The cigarette brands that have been marketed during the period 2005-2021 (n=40), can only be grouped into three groups according to the behavior of their sales. When we focus on the period 2005-2010 (n=74), cigarette brands are grouped into five groups according to their sales behavior. Finally, the cigarette brands marketed during the period 2011-2021 (n=67) are grouped into three groups according to the temporal evolution of their sales. These results suggest a greater homogenization of cigarette sales after the application of the law of January 2011.

CONCLUSIONS

Act 42/2010 (total ban on tobacco advertising, promotion, and sponsorship actions) was associated with greater homogenization of cigarette sales than the application of Act 28/2005 (partial ban). This finding supports what is established in the previous literature that indicates that Act 42/2010 provided a more restrictive specification of the ban than Act 28/2005.

Forecasting local hospital bed demand for COVID-19 using on-request simulations

Accurate forecasting of hospital bed demand is crucial during infectious disease epidemics to avoid overwhelming healthcare facilities. To address this, we developed an intuitive online tool for individual hospitals to forecast COVID-19 bed demand. The tool utilizes local data, including incidence, vaccination, and bed occupancy data, at customizable geographical resolutions. Users can specify their hospital's catchment area and adjust the initial number of COVID-19 occupied beds. We assessed the model's performance by forecasting ICU bed occupancy for several university hospitals and regions in Germany. The model achieves optimal results when the selected catchment area aligns with the hospital's local catchment. While expanding the catchment area reduces accuracy, it improves precision. However, forecasting performance diminishes during epidemic turning points. Incorporating variants of concern slightly decreases precision around turning points but does not significantly impact overall bed occupancy results. Our study highlights the significance of using local data for epidemic forecasts. Forecasts based on the hospital's specific catchment area outperform those relying on national or state-level data, striking a better balance between accuracy and precision. These hospital-specific bed demand forecasts offer valuable insights for hospital planning, such as adjusting elective surgeries to create additional bed capacity promptly.

Response trait diversity and species asynchrony underlie the diversity-stability relationship in Romanian bird communities

Biodiversity-stability relationships have frequently been studied in ecology, with the recent integration of traits to explain community stability over time. Classical theory underlying the biodiversity-stability relationship posits that different species' responses to the environment should stabilise community-level properties (e.g. biomass or abundance) through compensatory dynamics. However, functional response traits, which aim to predict how species respond to environmental change, are still rarely integrated into studies of ecological stability. Such traits should mechanistically drive community stability, both in terms of community abundance (functional variability) and composition (compositional variability). In turn, whether and how functional or compositional stability scales to affect temporal variation in functional effect traits (a proxy for ecosystem functioning) remains largely unknown, but is key to consistent ecosystem functioning under environmental change. Here, we explore the diversity-stability relationship in bird communities using annual survey data across 98 sites in central Romania, in combination with global trait databases and structural equation models. We show that higher response trait diversity promotes compositional variability directly, and functional variability indirectly via species asynchrony. In turn, functional variability impacts the temporal stability of effect trait diversity. Multiple facets of diversity and community stability differ between natural forests and agricultural or human-dominated survey sites, and the relationship between response diversity and functional variability is mediated by land cover. Further integration of response-and-effect trait frameworks into studies of community stability will enhance understanding of the drivers of biodiversity change, allowing targeted conservation decision-making with a focus on stable ecosystem functioning in the face of global environmental change.

Trends in mental health inequalities for people with disability, Australia 2003 to 2020

OBJECTIVE

Cross-sectional studies have demonstrated that people with disability have substantial inequalities in mental health compared to people without disability. However, it is not known if these inequalities have changed over time. This study compared the mental health of people with and without disability annually from 2003 to 2020 to investigate time trends in disability-related mental health inequalities.

METHODS

We use annual data (2003-2020) of the Household, Income and Labour Dynamics in Australia Survey. Mental health was measured using the five-item Mental Health Inventory. For each wave, we calculated population-weighted age-standardised estimates of mean Mental Health Inventory scores for people with and without disability and calculated the mean difference in Mental Health Inventory score to determine inequalities. Analyses were stratified by age, sex and disability group (sensory or speech, physical, intellectual or learning, psychological, brain injury or stroke, other).

RESULTS

From 2003 to 2020, people with disability had worse mental health than people without disability, with average Mental Health Inventory scores 9.8 to 12.1 points lower than for people without disability. For both people with and without disability, Mental Health Inventory scores decreased, indicating worsening mental health, reaching the lowest point for both groups in 2020. For some subpopulations, including young females and people with intellectual disability, brain injury or stroke, mental health inequalities worsened.

CONCLUSION

This paper confirms that people with disability experience worse mental health than people without disability. We add to previous findings by demonstrating that disability-related inequalities in mental health have been sustained for a long period and are worsening in some subpopulations.

Short-term effects of air pollution exposure on the risk of preterm birth in Xi'an, China

INTRODUCTION

Long-term exposure to air pollution is known to be harmful to preterm birth (PTB), but little is known about the short-term effects. This study aims to quantify the short-term effect of particulate matter with aerodynamic diameter ≤ 2.5 μm (PM_2.5), ≤10 μm (PM₁₀) and nitrogen dioxide (NO₂) on PTB.

MATERIALS AND METHODS

A total of 18,826 singleton PTBs were collected during the study period. Poisson regression model combined with the distributed lag non-linear model was applied to evaluate the short-term effects of PTBs and air pollutants.

RESULTS

Maternal exposure to NO₂ was significantly associated increased risk of PTB at Lag1 (RR: 1.025, 95%CI: 1.003-1.047). In the moving average model, maternal exposure to NO₂ significantly increased the risk of PTB at Lag01 (RR: 1.029, 95%CI: 1.004-1.054). In the cumulative model, maternal exposure to NO₂ significant increased the risk of PTB at Cum01 (RR:1.026, 95%CI: 1.002-1.051), Cum02 (RR: 1.030, 95%CI: 1.003-1.059), and Cum03 (RR: 1.033, 95%CI: 1.002-1.066). The effects of PM_2.5, PM₁₀ and NO₂ on PTB were significant and greater in the cold season than the warm season.

CONCLUSIONS

Maternal exposure to NO₂, PM_2.5 and PM₁₀ before delivery has a significant risk for PTB, particularly in the cold season.Key messagesMaternal exposure to NO₂ was significant associated with an increased risk of preterm birth at the day 1 before delivery.Particle matter (PM_2.5 and PM₁₀) showed a significant short-term effect on preterm birth in the cold season.The effects of air pollutants on preterm birth was greater in the cold season compared with the warm season.