Browse Articles

Discrimination stability and calibration of cardiovascular risk prediction models in the Framingham baseline cohort

Large language models pass a standard three-party Turing test

The Turing test has been widely discussed as a test of machine intelligence, but it also provides a measure of how humans distinguish other humans from machines. We evaluated 4 systems (ELIZA, GPT-4o, LLaMa-3.1-405B, and GPT-4.5) in two randomized, controlled, and preregistered Turing tests on independent populations. Participants had 5 min conversations simultaneously with another human participant and one of these systems before judging which conversational partner they thought was human. When prompted to adopt a humanlike persona, GPT-4.5 was judged to be the human 73% of the time: significantly more often than interrogators selected the real human participant. LLaMa-3.1, with the same prompt, was judged to be the human 56% of the time—not significantly more or less often than the humans it was being compared to. Without these prompts, however, the same models performed significantly worse (38% and 36%), and did not consistently outperform baseline models, ELIZA and GPT-4o (23% and 21%, respectively). A third study replicated these results in 15-min games: two PERSONA-prompted models achieved pass rates of 56% and 59%. The results constitute empirical evidence that artificial systems can pass a standard three-party Turing test. Interrogators’ reasoning focused more on stylistic and socio-emotional aspects of human behavior rather than more traditional notions of intelligence. The results have implications for debates about what kind of intelligence is exhibited by large language models, the social impacts these systems are likely to have, and the aspects of human behavior that people continue to see as unique.

Correction: Comparative evaluation of root canal morphology in mandibular first premolars with deep radicular grooves using direct vision, dental operating microscope, 2D radiographic visualisation and micro-computed tomography

Understanding the influence of training schedules on short-term load forecasting via deep residual networks: An empirical study

Leveraging health care settings to strengthen trust in light-touch benefit outreach

MixNet: A scale-adaptive method for multivariate time series forecasting

Time series forecasting is a critical task with widespread applications in industrial domains and daily life, including weather prediction, long-term energy consumption planning, and marketing analysis. Nevertheless, effectively extracting salient temporal patterns and exploring dependencies within multivariate time series remains a challenge. This paper focuses on multivariate time series forecasting, a common and pivotal issue in numerous analytical tasks. To address the complexity and high variability inherent in multivariate time series, we propose a scale-adaptive multi-head attention mechanism based on a hybrid mixture of experts network. Building on this mechanism, we develop MixNet, a novel architecture designed to achieve flexible feature extraction across diverse types of time series data. Furthermore, to tackle the difficulty in capturing inter-variable dependencies, we introduce a dedicated multivariate time series embedding (MTSE) scheme integrated with learnable positional encoding. This approach aims to comprehensively model the dependencies among variables, thereby enhancing overall forecasting performance. Experimental results demonstrate that MixNet outperforms several state-of-the-art methods on seven benchmark datasets from primary domains.

High performance Raman amplifier: applications in optical communication and biomedical devices

Spatially tunable multiomic sequencing using light-driven combinatorial barcoding of molecules in tissues

Mapping the molecular identities and functions of cells within their spatial context is key to understanding the complex interplay within and between tissue neighborhoods. A wide range of methods have recently enabled spatial profiling of cellular anatomical contexts, some offering single-cell resolution. These use different barcoding schemes to encode either the location or the identity of target molecules. However, all these technologies face a trade-off between spatial resolution, depth of profiling, and scalability. Here, we present B arcoding by A ctivated L inkage of I ndexes (BALI), a method that uses light to write combinatorial spatial molecular barcodes directly onto target molecules in situ, enabling multiomic profiling by next generation sequencing. A unique feature of BALI is that the user can define the number, size, shape, and resolution of the spatial locations to be interrogated, with the potential to profile millions of distinct regions with subcellular precision. As a proof of concept, we used BALI to capture the transcriptome, chromatin accessibility, or both simultaneously, from distinct areas of the mouse brain in single tissue sections, demonstrating strong concordance with publicly available datasets. We also developed an integrated instrument that automates combinatorial barcode writing on tissue sections, enabling high-throughput profiling. BALI therefore combines high spatial resolution, high throughput, compatibility with standard histological pipelines, and workflow accessibility to enable tunable spatial multi-omic profiling.

Viral dsRNA triggers human fetal membrane miR-146a-3p to be packaged into small extracellular vesicles which in turn drives inflammation through activation of Toll-like Receptor 7 and 8

Maternal infection and chorioamnionitis are one of the leading causes of preterm birth and neonatal morbidity. The relationship and mechanisms linking bacterial infections and preterm labor are well researched, however, less is known about the mechanisms involved in how viral infections contribute to preterm labor. Previous work from our group demonstrated that following bacterial triggers, fetal membranes (FMs) express elevated miR-146a-3p which in turn acts as an intermediate danger signal by activating TLR8 to induce a robust inflammatory response. Using an established FM explant model system, the role of this and other TLR7/8-activating miRs in the propagation of viral-induced inflammation was investigated. Following exposure to the viral dsRNA mimic and TLR3 agonist, Poly(I:C), expression of FM tissue TLR7/8-activating miRs (miR-146a-3p, miR-21a, miR-29a, and Let7b) were not elevated. Despite this, in response to Poly(I:C), elevated FM secretion of pro-inflammatory IL-6 and IL-8, and IL-1β was TLR7- and TLR8-dependent. To investigate alternative methods of miR delivery, small extracellular vesicles (sEVs) from FM supernatants were isolated and found to contain elevated levels of miR-146a-3p and miR-21a under Poly(I:C) conditions. Furthermore, Poly(I:C)-induced IL-6 and IL-8 responses were reduced in the presence of an inhibitor of sEV biogenesis/release, and IL-6 and IL-1β production was reduced in the presence of a miR-146a-3p inhibitor. Together, these data suggests that sEVs produced from virally-stimulated human FMs contain and deliver elevated miR-146a-3p which acts as a danger signal to drive perpetuate inflammation via TLR7 and TLR8 activation. This work demonstrates a novel and important role for sEV packaged TLR7/8 activating-miR-146a-3p in FM inflammatory responses to viral infections.

Unveiling a novel role for p19Arf (alternative reading frame) in mESC differentiation toward the pancreatic lineage

A systematic review and meta-analysis on achievement emotions, working memory and student-teacher relationship during second language learning in primary school

Learning is a multidimensional process resulting from the interaction between cognitive and emotional factors within the learning context; in this respect the quality of the student-teacher relationship plays a significant role. Although the literature suggests that cognitive processes and emotions experienced during learning and task performing play a central role in academic achievement, it remains unclear how these factors interact with socio-affective factors in explaining academic performance, particularly in second language (L2) learning from primary school. This systematic review and meta-analysis focused on the studies that jointly or individually investigated the role of emotional factors (achievement emotions), socio-affective factors (student-teacher relationship) and cognitive factors (working memory) in L2 learning during primary school. Our sample contained 19 primary studies with 5,340 participants involved in at least one of the factors of our interest. 16 out of 19 studies were included in the meta-analysis. Our results showed a positive correlation between working memory and L2 learning, differentiated effects of achievement emotions, with a significant negative association with anxiety, and a small but positive association with enjoyment. The student-teacher relationship was supported only by qualitative evidence, however, showing a protective effect of emotional closeness to the teacher in the learning process in the presence of negative emotions such as anxiety. Findings support the importance of integrating cognitive, emotional, and relational factors to understand L2 learning in primary school. Further empirical research focusing on positive emotions and relational dynamics in different educational contexts is needed.

Chirality and width effects on elastic and fracture properties of nanoribbons with coronene edges

Abstract Coronenes are a class of polycyclic aromatic hydrocarbons consisting of fused benzene rings arranged in a specific pattern. The structure of coronenes can vary depending on the number of fused benzene rings, with the most common form being hexabenzocoronene, which consists of six benzene rings fused in a circular arrangement. In 2012, coronene-based nanoribbons (CNRs) were successfully synthesized. Although several properties of CNRs have already been studied, a complete characterization of the mechanical properties of CNRs and the effect of the edges are lacking. This study examines how the width and chirality of CNRs affect their mechanical properties under tensile stress. Nanoribbons of rectangular shape and the same width as the CNRs are also studied for the determination of the edge effects on these properties. CNRs of different widths are analyzed, from which the elastic constants are obtained. Results indicate a transition in the mechanical properties from quasi-1D to 2D systems, with the most significant change observed between the narrowest and second-narrowest CNRs considered here. The peak stress is approximately 40 GPa larger for AC-CNRs than for ZZ-CNRs. The response to strain also differs between CNRs with armchair (AC-CNR) and zigzag (ZZ-CNR) edges due to variations in bonding orientations, with AC-CNRs sustaining larger strains than ZZ-CNRs. Fracture patterns are similar for all widths except for the narrowest nanoribbons. Furthermore, the elastic constants of CNRs exhibit an increase with width, approaching the graphene value of 1 TPa, as expected. However, an interesting trend of the Young’s modulus of CNRs with increasing width is observed: the first five AC-CNRs of smallest widths present larger Young’s moduli than those of the first five ZZ-CNRs of smallest widths, and invert for the CNRs with larger values of width. This result is then discussed in terms of the differences in the shape of the edges between AC and ZZ CNRs.

Electoral politics influence expansion and titling of informal urban settlements

Understanding what and who shapes urban growth and how it unfolds in the Global South is critical to ensure theories and predictive models of urban expansion and, consequently, global environmental change, are salient to contemporary contexts. Much of the urban expansion in the Global South occurs in informal settlements where urbanization is not officially permitted and residents often lack access to basic services and infrastructure. As a result, formalizing settlements becomes an asset for political transactions where politicians provide titles and improve conditions in exchange for residents’ votes. Yet the cumulative effect of these political transactions on land change remains unknown, rendering electoral dynamics absent from land system science models of urbanization. Here we use panel regression to examine the influence of electoral cycles and voting patterns on urban expansion and formalization of property titles from 1997–2015 in Mexico City. We find the distribution of land titles to residents increases in the months leading up to local elections, and more titles are given to core voting neighborhoods of the historically dominant party. Furthermore, urban built-up area increases in districts with higher electoral competition. These results demonstrate that electoral politics influence when and where the legal city boundary expands, who can access secure tenure, and potentially which settlements grow. These results imply sociopolitical dynamics shape urban sustainability. Incorporating quantified links between political dynamics and urban expansion could improve models of future urban growth and identify relevant levers to improve urban sustainability.

Correction: Can laboratory-based XAFS compete with XRD and Mössbauer spectroscopy as a tool for quantitative species analysis? Critical evaluation using the example of a natural iron ore

Ameliorative effects of curcumin and resveratrol following abamectin-induced liver injury in rats

Abstract Abamectin (ABA), a macrocyclic lactone of the avermectin family, is widely used in agriculture and animal husbandry for its insecticidal and anthelmintic properties. The present study investigated the therapeutic and antioxidant potential of curcumin (CUR) and resveratrol (RES) following abamectin-induced hepatotoxicity in Wistar albino rats. Fifty-eight male rats were randomly assigned to six groups: control (olive oil), CUR, RES, ABA, ABA + CUR, and ABA + RES. Treatments were administered orally for seven days, and liver tissues were collected on day eight for biochemical and histopathological evaluation. ABA exposure significantly decreased hepatic glutathione (GSH), total antioxidant capacity (TAC), and glutathione S-transferase (GST) activity, while markedly increasing lipid peroxidation (LPO), total oxidant status (TOS), and oxidative stress index (OSI) (p < 0.001). CUR and RES supplementation effectively mitigated these alterations, restoring antioxidant parameters and reducing oxidative stress. Histopathological analysis confirmed severe hepatic injury and elevated TNF-α expression in the ABA group, whereas CUR and RES treatments improved liver architecture and reduced injury scores. These results indicate that CUR and RES effectively attenuate ABA-induced oxidative liver damage and may represent promising adjunctive candidates for managing pesticide-related hepatotoxicity.

A functional map of the human intrinsically disordered proteome

Intrinsically disordered regions (IDRs) represent at least one-third of the human proteome and defy the established structure–function paradigm. Because IDRs often have limited positional sequence conservation, the functional classification of IDRs using standard bioinformatics is generally not possible. Here, we show that evolutionarily conserved molecular features of IDRs enable clustering of the human disordered proteome (IDRome) into a map with strong functional enrichments. We quantify how conserved IDR features correlate with functional terms and, for a subset of terms, provide proteome-wide predictions of annotations for IDRs. Further, we show that conserved features of IDRs can predict protein localization to different biomolecular condensates and underlie elevated intracluster connectivity in condensate-associated IDRs, as well as enrich for short-linear motif-binding domains among interaction partners. We highlight patterns of conservation in disordered proteins with unknown function and in clusters enriched for proteins encoded by disease-risk genes. Our map of the human IDR-ome should be a valuable resource that aids in the discovery of new IDR biology.

Medicines prescribed in pregnancy: Protocol for a signal detection study using routinely collected data in England

Introduction Medication use in pregnancy is common and increasing. Pregnant women are often excluded from drug trials, therefore the safety to mother and baby of many medications used in pregnancy is unknown. Routine health data have been used to look for evidence of harm and safety of specific medications, but not to systematically look across all prescribed medications in pregnancy. Aims To systematically identify medications prescribed during pregnancy requiring further research into their potential harms to mother or baby. Methods We will describe trends in primary care prescribing in pregnancy over time, and socio-demographic patterning of prescribing, using data from pregnancies in the Clinical Practice Research Datalink (CPRD). We will identify and categorise maternal, fetal, and infant adverse outcomes reported in the CPRD and linked hospital and mortality data. We will use Bayesian signal detection models to analyse all medication/outcome pairs to identify where an outcome occurs significantly more frequently in mothers prescribed a specific medication than in mothers not prescribed it. Potential confounders will not be accounted for at this stage. Published evidence on the identified medication/outcome pairs will be reviewed and incorporated with the results from trends and signal detection analyses. These data will be used by Patient Involvement and Study Advisory Board workshops to prioritise medications for further research, with medications more commonly prescribed for underserved groups being given precedence. For three prioritised signals, cohort analyses will compare the occurrence of adverse outcomes in those prescribed compared to those not prescribed the medication of interest, adjusting for maternal morbidity, co-medications, and other measured confounders. Discussion This project will provide a list of medications used in pregnancy that should be prioritised for further safety research. Results of three detailed retrospective cohort studies will contribute to informed decision making for patients and clinicians. Study grant number: NIHR207172; CPRD accepted protocol number: 24_004518.

Retraction Note: Structural configuration of sustainable sports industry based on deep learning and genetic algorithm

Natural variations in small RNA origin loci generate circuit diversity underlying temperature acclimation in <i> <i>Caenorhabditis elegans</i> </i>

Individuals respond differently to environmental cues because of inherent variations in genome sequences. This study demonstrates that natural variation in small RNA (sRNA) within embryos epigenetically shapes adaptive neural circuits underlying diverse animal temperature acclimation. We performed comparative genome profiling of Caenorhabditis elegans variants with differential temperature acclimation and identified smrn-1 , a member of a novel gene family predicted to be conserved within nematode genomes. Unexpectedly, determination of full-length smrn-1 sequence by long-read sequencing revealed 1791 orthologues present in human genome. smrn-1 was among the most abundant sRNA-accumulating genes downstream of the helicase ERI-6/7 functioning in embryos. smrn-1 -derived sRNAs were loaded onto the Argonaute protein HRDE-1, thereby regulating axonogenesis of O 2 -sensing BAG neurons through upregulation of an axonogenesis regulator protein. O 2 information from BAG influences the thermal responsiveness of temperature-sensing neurons, resulting in individual diversity in temperature acclimation. Thus, natural variation in embryonic sRNA epigenetically forms adaptive neural circuits associated with diverse temperature acclimation. We propose that accumulation of gene polymorphisms producing epigenetic regulators during nematode evolution generates adaptive neural circuits for temperature acclimation.

The perils of pay variability: Determinants of worker aversion to variable compensation in low- and middle-wage jobs

A substantial proportion of the labor force in low- and middle-wage jobs is prone to pay variability, or variance in earnings from paycheck to paycheck. Emerging research suggests that pay variability can be detrimental to workers’ well-being and enhance their likelihood of exiting their job. Using several iterations of qualitative and quantitative data, we identify and evidence critical determinants of workers’ experience of pay variability and their likelihood of voluntary turnover. Combining insights from exploratory interviews and surveys with ride-hailing drivers (N = 63) and cognitive appraisal theory, we predicted that pay variability is most likely to result in voluntary turnover when (a) workers were more frequently earning lower-than-average paychecks, (b) felt limited control over their earnings, and (c) their household was dependent upon their paycheck. Using matched survey and archival data from a sample of truck drivers (N = 711) from a national transportation company, we subsequently found that pay variability is positively associated with the likelihood of turnover, but this relationship significantly varied with the predicted contextual moderators. Additional quantitative analyses demonstrated that this relationship was observed even among the highest performing drivers and, surprisingly, likely resulted in less annual pay for those who left. Supplemental qualitative data provided further support for the predicted moderators. Taken together, our findings explain why and when variable compensation represents a source of precarity for low-and middle-wage workers that can motivate turnover. Theoretical and practical implications are discussed.