Browse Articles

Emergent universal long-range structure in random-organizing systems

Abstract Self-organization through noisy interactions is ubiquitous across physics, mathematics, and machine learning, yet how long-range structure emerges from local noisy dynamics remains poorly understood. Here, we investigate three paradigmatic random-organizing particle systems drawn from distinct domains: models from soft matter physics (random organization, biased random organization) and machine learning (stochastic gradient descent), each characterized by distinct sources of noise. We discover universal long-range behavior across all systems, namely the suppression of long-range density fluctuations, governed solely by the noise correlation between particles. Furthermore, we establish a connection between the emergence of long-range structure and the tendency of stochastic gradient descent to favor flat regions of energy landscape—a phenomenon widely observed in machine learning. To rationalize these findings, we develop a fluctuating hydrodynamic theory that quantitatively captures all observations. Our study resolves long-standing questions about the microscopic origin of noise-induced hyperuniformity, uncovers striking parallels between stochastic gradient descent dynamics on particle system energy landscapes and neural network loss landscapes, and should have wide-ranging applications—from the self-assembly of hyperuniform materials to ecological population dynamics and the design of generalizable learning algorithms.

General risk preference comes up short when predicting risk-taking frequency

Abstract Situations involving risk are common in human affairs, from financial investments, criminal behavior, and health-related decisions. Peoples’ self-reported risk preferences, where individuals in various ways report how they feel about risk, have been considered central to understanding why, and predict when, people behave differently in real-life situations involving risk. However, various other factors have also been suggested as predictors, including age, gender, education, income, anxiety, sensation seeking, impulsivity, and personality traits like neuroticism and extraversion. Still, research is limited on which factors best predict the frequency of risk-taking in real life. In this study, we asked respondents ( n  = 760) to report how often they engaged in various risk-taking behaviors along with the abovementioned variables with the aim of predicting risk-taking frequency. The results from Bayesian multi-model inference analyses showed that the most important predictors of risk-taking frequency were impulsivity, sensation seeking, health and social risk preferences, and gender. The study highlights the importance of examining multiple variables simultaneously when predicting risk-taking frequency. The importance of the concept of general risk-preference, which occupies a central role in many influential theories of risk-taking, should arguably be reconsidered.

Multi-modal data to identify key factors influencing lung injury in ARDS patients undergoing invasive mechanical ventilation: A prospective multi-center observational study protocol

Background Patients with moderate to severe acute respiratory distress syndrome (ARDS) exhibit extremely poor prognoses following mechanical ventilation, with mortality rates as high as 40% to 55%. Despite extensive research into ARDS classification and prognostic assessment, the disease’s pathogenesis remains incompletely understood, and there remains a critical lack of specific biomarkers and effective therapeutic targets for its prevention and management. The core challenges lie in two key areas. First, ARDS demonstrates marked heterogeneity in etiology, pathophysiology, and pathogenesis. Existing research, predominantly reliant on population-level average data, fails to capture inter-individual variability, hindering the precise identification of patient subgroups responsive to specific therapeutic regimens. Second, current definitions of ARDS phenotypes are often confined to clinical symptoms and routine diagnostic indices, lacking integrated analysis of deeper mechanistic indicators, such as key biomarkers and respiratory mechanics parameters, thereby limiting the stability and clinical utility of existing classification systems. Methods/Design We designed a prospective multicenter cohort study incorporating multi-omics analyses. This research aims to investigate the mechanisms underlying the development and progression of ARDS during mechanical ventilation, providing a theoretical foundation and practical guidance for future ARDS therapies. The study plans to enroll over 165 patients with moderate to severe ARDS receiving mechanical ventilation across 10 medical centers. Peripheral blood and bronchoalveolar lavage fluid (BALF) samples will be collected on the first 24 hours after enrollment and at extubation for metagenomic/meta-transcriptomic sequencing, bulk RNA sequencing, single-cell RNA sequencing, proteomics detection, and metabolomics analyses. Concurrently, comprehensive monitoring of physiological indices, electrical impedance tomography, transpulmonary pressure, pulmonary ultrasound findings, and other relevant parameters will be conducted during the enrollment. Study participants will be stratified by survival and mortality outcomes to analyze the dynamic trends of all measured indices and their underlying molecular mechanisms. Biomarkers derived from multi-omics data and clinical baseline characteristics will be evaluated and integrated, followed by multidimensional dimensionality reduction. Predictive models will be subsequently constructed via early or late fusion to identify core prognostic markers, with performance validated using standardized metrics. Discussion Through comparative analysis of multi-omics data, we aim to identify specific markers and risk factors associated with distinct clinical trajectories of ARDS, further clarifying the key determinants of lung injury. Ultimately, this research will reveal critical immune cell subtypes that govern ARDS onset and prognosis, offering novel insights and therapeutic targets to advance precision medicine for ARDS. Study protocol registration ClinicalTrials.gov NCT05922826 .

Combinatory differentiation of human induced pluripotent stem cells generates functional thymic epithelium driving dendritic cell and CD4/CD8 T cell development

Abstract The thymus educates thymocytes through a selection process mediated by thymic epithelial cells (TECs). Recent advances have made the generation of T lymphocytes from induced pluripotent stem cells (iPSc) a promising therapeutic strategy. However, current approaches often fail to replicate the thymic niche, leading to impaired T cell generation. Here we address the production of functional mature iPSc-derived TECs supporting in vitro T cell generation. We optimize thymic lineage differentiation through an unbiased multifactorial experimental design. By modulating specific signaling pathways, we generate progenitors that mature into medullary and cortical TECs. Co-culture with primary hematopoietic progenitors in a 3D thymic organoid setup induces their differentiation into CD4 + and CD8 + T cells. Importantly, thymic organoids support multilineage differentiation, with dendritic cell populations also emerging. Thus, the presented thymic organoid model provides a practical platform for studying thymic cellular interactions and thymopoiesis in vitro, and opens further research perspectives towards cell-based therapies.

Label-free fluorescence lifetime imaging can distinguish cancer from healthy tissue in spontaneously occurring canine oral tumors

Abstract Post-surgical local recurrence of oral cancer remains unacceptably high across species due to the lack of non-invasive tools capable of accurately delineating tumor from healthy tissue. Label-free fluorescence lifetime imaging (FLIm) has shown moderate-to-high success in human head and neck squamous cell carcinoma, but it is unclear whether diagnostic accuracy can be enhanced by incorporating exogenous fluorophores that selectively accumulate in cancer. This study evaluated the performance of 5-ALA induced Protoporphyrin IX (PpIX) fluorescence and autofluorescence FLIm features to discriminate epithelial cancer and healthy tissues in a spontaneous large animal model of disease (15 pet dogs). Fluorescence emission parameters (e.g. lifetimes, intensity ratios, phasors and Laguerre coefficients) differed significantly ( p  < 0.001) between cancer and healthy tissues in both autofluorescence and PpIX channels. However, autofluorescence features, particularly lifetimes in Channel 1 (390 nm, collagen-sensitive) and intensity ratios in Channel 2 (470 nm, NADH-sensitive), provided the strongest in vivo discrimination. These results demonstrate that label-free FLIm alone is sufficient to distinguish epithelial oral cancers from healthy tissue in dogs, and that the addition of exogenous markers such as 5-ALA–induced PpIX, does not markedly improve diagnostic accuracy enough to warrant incorporation into flourescence imaging approaches.

Formononetin ameliorates SP-induced urticaria in mice via suppressing TAK1/MAK signaling pathway

Background Chronic idiopathic urticaria (CIU) is a condition that significantly impacts patient well-being, requiring effective therapeutic strategies. Formononetin, a natural isoflavone with anti-inflammatory properties, has shown promise in allergic conditions. However, its specific effects and mechanism in CIU are not fully understood. Methods Comparative analyses were conducted between normal and formononetin-treated groups, along with mechanistic investigations into the TAK1/MAPK pathway both in vivo and in vitro . Cell morphology, cytokine secretion, histamine release, and TAK1/MAPK pathway alterations were assessed. Results Formononetin treatment led to a dose-dependent reduction in MC degranulation, histamine release, and secretion of inflammatory cytokines (TNF-α, IL-1β, IL-6). Additionally, formononetin inhibited the phosphorylation of TAK1, p38, ERK1, and JNK similar to a TAK1 inhibitor in murine and cellular models. Conclusion Formononetin shows potential as an anti-allergic agent by alleviating inflammatory responses in CIU through suppression of the TAK1/MAPK pathway in both murine models and MC/9 cells.

Unravelling the nucleation–elongation mechanism of one-pot catenation

Neutrophil extracellular traps as biomarkers for predicting prognosis and chemotherapy response in colorectal cancer

A cross-national study examining imaginary companions and face pareidolia in British and Chinese adults

Although imaginary companions are created by children and sometimes adults around the world, the prevalence of this play behaviour varies. Cross-nationally, imaginary companions are reported more frequently in Western countries. These imaginary entities have been speculated to be similar to hallucination-like-experiences, based on evidence for elevated top-down auditory processing in children who report them. Face pareidolia tasks engage visual top-down processing, and performance on them does not tend to vary across cultures. This study asked if: 1) there would be cross-national differences in imaginary companion creation in childhood and adulthood between Chinese and British adults, 2) whether those creating imaginary companions would see more face pareidolia and 3) if there would be cross-national differences in face pareidolia. 291 participants (185 Chinese) completed a questionnaire on their imagination followed by a face pareidolia task consisting of 36 image trials (24 containing face pareidolia). Results showed that including all participants (Chinese and British) 11% of the adults currently had an imaginary companion. Chinese adults were significantly less likely than British adults to report a childhood, but not adulthood, imaginary companion. There were significantly more reports of face pareidolia from participants with a current imaginary companion, but not those who remembered a companion in childhood. The pareidolia hits did not differ between country, but false alarms were experienced significantly more by the Chinese participants. Taken together, the results provide more information around imaginary companion creation in China and the UK as well as the role top-down processing may play in imaginary companion interactions.

Endothelial stem cells of the retinal vasculature reside in the optic nerve

A proteomics and redox proteomics approach to understanding ARDS heterogeneity

Abstract Acute Respiratory Distress Syndrome (ARDS) is a severe and heterogeneous critical illness characterized by systemic inflammation, lung injury, and profound hypoxemia. To investigate the temporal evolution of molecular features underlying ARDS heterogeneity, we applied advanced proteomics and redox proteomics to matched plasma and bronchoalveolar lavage (BAL) fluid samples collected longitudinally from 16 intensive care unit (ICU) ARDS patients during hospitalization. Exploratory, data-driven hierarchical clustering (Ward method) identified three distinct molecular patterns across patients represented as Groups A, B, and C. This framework was associated with illness severity at study enrollment (Group A profiling patients with more severe illness at enrollment), demonstrated temporal stability across sampling timepoints, and revealed molecular features associated with clinical improvement during hospitalization. Key pathways distinguishing the molecular patterns and consistent with prior findings included the production and detoxification of reactive oxygen species (ROS), Liver X receptor–Retinoid X receptor (LXR/RXR) activation and 24-dehydrocholesterol reductase (DHCR24) signaling, interleukin-12 (IL-12) signaling and production in macrophages, and neutrophil degranulation. Although plasma proteomic profiles were generally consistent with findings in BAL fluid, BAL fluid data were more mechanistically informative and enabled clearer and more consistent interrogation of ARDS molecular heterogeneity. The results highlight the potential value of lung–focused, temporal studies to improve patient stratification and guide future therapeutic strategies. However, the modest cohort size and exploratory nature of this study necessitate cautious interpretation of pathway-level inferences. Future longitudinal studies in larger, independent ARDS cohorts will be required to validate these molecular groups and assess their clinical relevance.

Effectiveness of roadside alcohol testing in reducing fatal accidents and fatal drinking-driving accidents: A multi-city study in China

This study examines the dynamic relationship between roadside alcohol check rates and traffic mortality across 248 cities in mainland China from 2014 to 2020. Using a dataset comprising 365,753 roadside check arrests, 227,896 traffic deaths, and 21,036 DUI-related fatalities, we applied both traditional time series analysis (Vector Autoregression and Impulse Response Functions) and machine learning techniques (XGBoost) to explore temporal and nonlinear patterns. The time series analysis revealed an initial positive association between enforcement intensity and mortality rates, likely reflecting reactive increases in checks following fatality rises. However, a longer-term beneficial effect emerged after approximately eight months, particularly pronounced in smaller cities. Machine learning models revealed that roadside check rates are more strongly associated with overall traffic mortality than with DUI-specific deaths. Enforcement was found to have a greater impact in smaller cities compared to larger ones. Notably, the patterns reveal diminishing returns in enforcement effectiveness, with the marginal benefits tapering off around 0.002 per 100,000 people per month in large cities, while remaining evident up to about 0.006 in small cities. These findings suggest that increasing roadside checks in low population density areas may lead to the most significant reductions in traffic fatalities at the national level. However, limitations include data exclusions due to non-disclosure and the inability to determine causal mechanisms. Overall, the study offers valuable insights for optimizing DUI enforcement strategies, highlighting the importance of tailored approaches based on city size and enforcement thresholds.

Phonon-driven wavefunction localization enhances room-temperature single-photon purity in large hybrid lead halide perovskite quantum dots

Abstract In lead halide perovskites (APbX 3 ), the effect of the A-site cation on optical and electronic properties has initially been thought to be marginal. Yet, evidence of beneficial effects on solar-cell performance and light emission is accumulating. Here, we report that the A-site cation in soft APbBr 3 colloidal quantum dots (QDs) controls the phonon-induced localization of the exciton wavefunction. Insights from ab-initio molecular-dynamics simulations and single-particle fluorescence spectroscopy demonstrate that anharmonic crystal vibrations and the resulting disorder act as an additional confinement potential. Avoiding the trade-off between single-photon purity and optical stability faced by downsizing conventional QDs into the strong confinement regime, dynamical phonon-induced confinement in large organic-inorganic perovskite QDs enables bright (10 6 photons/s), stable ( > 1 h), and pure (> 95%) single-photon emission tunable across a wide spectral range (495-745 nm). Strong electron-phonon interaction in soft perovskite QDs provides an unconventional route toward developing scalable room-temperature quantum-light sources.

Multiobjective starfish optimization algorithm for engineering design and optimal power flow problems

Abstract This paper presents a robust multi-objective optimization approach—the multi-objective starfish optimization algorithm (MOSFOA)—designed to address complex challenges in engineering design and optimal power flow analysis. As an advanced extension of the starfish optimization algorithm (SFOA), MOSFOA leverages biological inspiration from starfish behaviors such as exploration, predation, and regeneration to balance global exploration and local exploitation. The proposed MOSFOA employs elitist non-dominated sorting (NDS) and crowding distance (CD) mechanisms to preserve solution diversity and guide convergence toward the Pareto-optimal front. The effectiveness of MOSFOA is validated on standard ZDT and DTLZ benchmark suites and further demonstrated on real-world applications, including engineering design tasks and the IEEE 30-bus power system. Performance comparisons with ten state-of-the-art multi-objective algorithms, using metrics such as inverted generational distance (IGD) and hypervolume (HV), confirm the strength of MOSFOA in achieving a well-balanced trade-off between convergence and diversity. Additionally, the KKT proximity metric (KKTPM) is employed to assess convergence. The results demonstrate that MOSFOA significantly outperforms its counterparts in terms of both IGD and HV, achieving superior convergence and diversity performance. These findings underscore MOSFOA’s robustness, scalability, and stability across runs. Moreover, its strong performance in handling constrained engineering problems highlights its practical potential for real-world decision-making and optimization tasks in power systems and complex design optimization, making MOSFOA a promising tool for both theoretical research and industrial applications. Source code of MOSFOA are publicly available at https://www.mathworks.com/matlabcentral/fileexchange/183090-mosfoa-multi-objective-starfish-optimization-algorithm .

Recovery of expected salary estimated by facial emotion scores against computer-based landscape data

The perceived recovery of expected salary (RES) matters for work efficacy at a given amount of wage investment. A total of 31 industrial parks (IPs) were randomly chosen from North China. Employees’ facial photos were obtained from social networks and analyzed for happy, sad, and neutral emotion scores. Green spaces were analyzed as surface feature heights and area in 950m-buffer areas at every IP location. Green view index (GVI) was rated using a pre-trained machine-learning model on street view images (SVIs) crawled from the Baidu map. The Simpson diversity index was calculated by recognizing woody plant species in each SVI. RES was estimated as the difference of recruitment wage (mean ± standard deviation, 8625.62 ± 2735.54 CNY M -1 ) minus satisfactory salary (SS) (8153.77 ± 971.28 CNY M -1 ), which was positively impacted by GVI but a negative effect from Simpson plant diversity index. Although the green space area impaired happy score and perception of SS, it enforced a tiny contribution to RES with a negative contribution from the longitude of IPs.

Polar discontinuities, emergent conductivity, and critical twist-angle-dependent behaviour at wafer-bonded ferroelectric interfaces

Abstract Probing novel properties, arising from twisted interfaces, has traditionally relied on the stacking of exfoliated two-dimensional materials and the spontaneous formation of van der Waals bonds. So far, investigations involving intimate covalent or ionic bonds have not been a focus. Yet, we show here that an established technique, involving thermocompressional wafer bonding, works well for creating twisted non-van der Waals interfaces. We have successfully bonded z-cut lithium niobate single crystals to create ferroelectric oxide interfaces with strong polar discontinuities and have mapped the associated emergent interfacial conductivity. In some instances, a dramatic change in microstructure occurs, involving local dipolar switching. A twist-induced collapse in the capability of the system to effec8tively screen interfacial bound charge is implied. Importantly, this only occurs around specific moiré twist angles with sparse coincident lattices and associated short-range aperiodicity. In quasicrystals, aperiodicity is known to induce pseudo-bandgaps and we suspect a similar phenomenon here.

CKAAKN peptide-conjugated long-circulating nanoliposomes for the targeted delivery of oridonin to pancreatic cancers

HEDGES co-prevents both SARS-CoV-2 and pandemic influenza infection in mice by rapid, durable co-production of twelve different anti-pandemic monoclonal antibodies

Despite all currently available anti-pandemic monoclonal-antibodies (mAbs) and vaccines, subsequently emerging pandemic-infections will likely become more pan-resistant-, -transmissible and/or -lethal. We have created HEDGES generation-2, a significantly more-combinatorial, -synergistic version of our generation-1 HEDGES DNA vector-based platform. We previously published that one safe intravenous injection of a HEDGES generation-1 DNA vector encoding one of three different FDA-approved mAbs produced durable therapeutic serum mAb levels as well as critical therapeutic endpoints in immunocompetent mice. Here we show one safe, intravenous administration of a 2 nd -generation HEDGES DNA vector co-encoding four different anti-SARS-CoV-2 mAbs rapidly then durably co-produces high anti-SARS-CoV-2 mAb serum levels that effectively block SARS-CoV-2 virus binding to the ACE-2 spike protein in immunocompetent mice. In addition, four weekly intravenous HEDGES generation-2 DNA vector administrations co-encoding a total of ten-different anti-SARS-CoV-2 mAbs, 5J8, plus an anti-1918 pandemic influenza mAb and mepolizumab, an FDA-approved anti-IL-5 mAb, durably co-produce highly-neutralizing 5J8 anti-pandemic influenza mAb serum levels, as well as durably block SARS-CoV-2 virus-ACE-2 receptor binding in mice. Furthermore, unlike vaccines and mAbs, HEDGES does not require an intact cold chain and is readily freeze dried, enabling its prolonged storage at ambient temperatures worldwide, even in equatorial regions. Also, HEDGES can create, then deploy novel, more effective anti-pandemic mAbs ~three weeks after their identification. Conversely, vaccines require ~three months to deploy, recombinant-mAbs ~nine months. By rapidly then durably co-producing many different highly-neutralizing, highly-synergistic anti-pandemic mAbs, HEDGES may effectively co-prevent both SARS-CoV-2 and pandemic-influenza infections. HEDGES may also prevent even more-transmissible, -pan-resistant and/or -lethal pandemic diseases that subsequently-emerge.

Detecting anthropogenically induced changes in extreme and seasonal evapotranspiration observations

Abstract Increasing temperature and radiation drive an increase in evaporative demand. However, it is still uncertain whether the increase in demand has led to an increase in evapotranspiration (ET) in observational products, as this increase is at odds with a limited water supply over land. Here, we examine changes in high ET extremes and seasonal mean ET using climate models as well as observational data. High ET extremes are driven by periods with high incoming surface radiation and temperatures. In line with physical understanding, these events are intensified by anthropogenic climate change. We detect robust changes in extreme and seasonal ET in two observational data sets. Regionally, seasonal mean ET shows mixed increases and decreases from 1980 to 2023, while extreme ET universally increases or shows no significant change. Although the drivers for these changes can vary regionally, we expect that regions with strong extreme ET trends are at increased risk of flash droughts.

A deep reinforcement learning approach to dance movement analysis