Researchers analyzed 12 years of hospital website data and found pixel tracking in 66% of observations, significantly increasing breach risk. Hospitals using third-party pixels saw a 1.4 percentage point increase in breach probability compared to baseline rates.
A recent study estimates that 22 million Americans exceed safe nitrogen dioxide exposure levels due to gas stove use. Gas stoves are found to be a significant source of indoor nitrogen dioxide emissions, particularly in households with multiple burners and oven usage.
Researchers found that people tend to find less energetic images more attractive, as they require fewer neurons and lower metabolic costs. This study suggests that our attraction to beauty may be an adaptive energy-conserving mechanism to optimize brain function.
Large language models (LLMs) exhibit network formation behaviors similar to humans, including preferential attachment, triadic closure, and homophily. In real-world simulations, LLMs prioritize homophily most strongly, followed by triadic closure and preferential attachment.
A cross-national study found that people tend to be more generous toward those from their own country or culturally similar nations. People from wealthy nations also shared more with less wealthy individuals, possibly to reduce global inequality.
Online experiments found that seeing only poor individuals reduces support for tax rates, while observing both rich and poor increases support for redistribution. However, this increased support can lead to higher polarization and conflict.
Seismic activity in Yellowstone's subsurface microbial communities led to changes in chemical energy and planktonic cell concentrations. The study found that kinetic energy from earthquakes can change the geochemical and microbial compositions of aquifer fluids.
The human costs of climate overshoot are a pressing concern, with policymakers needing to address humanitarian needs and disaster risk management. Assessing the potential impacts post-peak is necessary for identifying adaptation and risk reduction strategies.
A machine learning framework, LifeTracer, has been developed to analyze mass spectrometry data and identify patterns distinguishing abiotic from biotic origins. The framework achieved over 87% accuracy in classifying meteoritic and terrestrial samples, with key predictive features including polycyclic aromatic hydrocarbons.
Launched to promote scientific and technological ideas, PNAS Nexus has brought together experts from various fields. The journal's editorial board, comprising members of the National Academies, has fostered a multidisciplinary convergence in pursuit of useful purposes.
A new analysis of scientific documents found that 60-80% of citation paths led to sources without supporting data, revealing much wider uncertainty ranges for irrigation's impact on grain production and freshwater withdrawals. The study highlights the need to critically evaluate foundational claims in sustainability science.
Researchers have identified and classified 100 biological tilings across various taxonomic groups, providing insights into their functions and dominance. These findings could lead to novel manufacturing design challenges and bio-inspired innovations.
A mega-analysis of 16 neuroimaging studies found that brain activity in reward and social processing regions can predict which messages are well-received by both scanned individuals and broader audiences. Effective messages activate brain responses related to mentalizing, anticipating rewards, and evoking emotions.
Researchers sequenced the genome of the triangle weaver spider to understand its spring-loaded web mechanism. The study found proline-rich protein sequences that enable the web's elasticity and supercontraction.
Learning with AI chatbots versus traditional web search produces shallower knowledge, as evidenced by experiments with thousands of online participants. Participants who used AI chatbots developed less original advice and were less willing to adopt it compared to those using web links.
Researchers found that opt-out policies significantly decrease living donors by 29% due to perceptions of sufficient organ supply. The studies suggest policymakers should consider crowding-out effects when implementing opt-out systems.
A recent study found that offline social connections predict US voting patterns more accurately than online social connections or residential proximity. The research analyzed large-scale data from the Meta Data for Good program, revealing that partisan exposure through physical co-location explained 97% of county-level voting patterns.
A study using the independent cascade model on diverse social networks found that algorithmic outreach creates information gaps, leaving vulnerable nodes without important information. A proposed multi-objective algorithm aims to minimize this effect, resulting in a 6-10% reduction in vulnerable nodes with negligible impact on reach.
Researchers used machine learning to analyze brain connectivity, environmental factors, and behavioral data from 295 older adults with cardiovascular diagnoses. The study identified key predictors of increased physical activity, including access to green spaces, social support, and executive function abilities.
Researchers found that yeast can survive Martian-like conditions by assembling ribonucleoprotein condensates, which protect RNA and affect mRNA fates. The study suggests the importance of understanding RNP condensates in predicting the effects of Martian conditions on life.
A study finds that providing wearable devices to participants from a diverse probability sample yields more equitable and accurate health data compared to convenience samples. This approach can help develop AI health tools that work equally well across all populations.
Researchers derive a scaling law to predict mutant emergence in growing cell populations, considering factors like population size and dimensionality. The law also accounts for intermediate mutants with varying fitness, providing insights into evolutionary biology and biomedicine.
A study of nearly 30,000 Chinese children found that air pollution, particularly nitrogen dioxide and fine particles, is a significant risk factor for childhood myopia. Exposure to poor air quality can cause oxidative stress and inflammation of the eye, potentially harming vision.
A study found that mindfulness meditation training significantly decreased depressive symptoms, Internet Addiction Test scores, and video game cravings in individuals with depression and gaming disorder. Brain imaging revealed a decrease in brain responses to gaming cues and increased activity in the lentiform nucleus.
A recent study using AI and satellite images estimates that fewer than 600,000 wildebeest migrate across the Serengeti-Mara ecosystem annually, contradicting previous aerial surveys. The findings highlight the potential of satellite-based population counts for terrestrial mammals.
Researchers created Bacterial Reef Ink, a living material that stimulates coral larvae to attach and settle. The ink's customized bacterial strains increased coral settlement by fivefold in test tanks.
Microbial communities around alkali waste barrels have low diversity and are dominated by alkalophilic bacteria, similar to those found at hydrothermal vents. The ecological effects of caustic alkaline waste dumping are expected to take thousands of years to cease.
Empirical evidence shows that land sharing and land sparing strategies alone are insufficient for balancing agricultural production with biodiversity conservation. Context-specific combinations of both approaches yield the best results, according to a meta-analysis published in PNAS Nexus.
Most people consistently choose settings featuring more people of their own ethnicity, age, and educational level. Individuals without a college degree were the only exception, showing willingness to interact with those of higher education levels.
A survey of over 3,000 US citizens found that Donald Trump's rhetoric about his prosecution had minimal effect on increasing support for him or reducing opposition to democratic norms.
Research reveals that women and girls exhibit higher levels of direct aggression towards their siblings compared to boys and men. In contrast, women and girls tend to engage in reputational aggression with non-siblings, similar to men and boys.
A large-scale study found that people who made successful guesses in a sequence of fair coin tosses rated their chances of further success higher and were more willing to bet on it. In contrast, unlucky guesses led to pessimistic predictions about future performance, with a stronger negative effect than the positive one.
A neural network trained on the World Values Survey predicts an individual's country of origin with 90% accuracy from their attitudes and values. The top predictors include political attitudes, environmental attitudes, family values, and interpersonal relationships.
A study found that in-utero exposure to extreme heat increases neonatal mortality rates, with a significant impact on economically disadvantaged and less educated mothers. Targeted interventions such as community health workers, education, and economic development are recommended to mitigate this effect.
Researchers developed a machine learning model to predict WNV risk based on weather, land cover, and demographic data. The model shows highest risk in urban areas with high temperature and precipitation during summer months.
A novel experimental rig accurately reproduces human circulatory system behavior, revealing the reason for cuff-based systolic underestimation. The study proposes improved calibration methods or refinements to the cuff protocol.
The study chronicles 220 extinctions in Macaronesia, with half linked to human arrival and a post-human extinction rate exceeding 12 times higher than pre-human periods. Conservation efforts are crucial to mitigate the extinction crisis in these unique island ecosystems.
The article highlights the risks of deepfakes and manipulated media, with expert Hany Farid advocating for legal requirements, imperceptible watermarks, and digital media literacy to mitigate these harms. He also explores policy landscape and authentication techniques to sort real from synthetic content.
Research combines AI-generated disinformation with the repetitive truth effect to create persuasive content. A study found that while neither tactic convinced participants of conspiracy theories, AI-generated disinformation increased perceived consensus among both parties.
A study analyzed social determinants' impact on mental health, identifying variables that predict problems, such as physical health, loneliness, and satisfaction. Implementing policies like financial stability, local community engagement, and volunteering may reduce mental illness rates by up to 3 percentage points.
A study of the genetic basis of plague immunity in prairie dogs has identified key genes linked to survival. The research provides insights into rapid adaptation to a globally introduced virulent pathogen in a keystone species.
Research on record-low years of Antarctic sea ice reveals warming of the Southern Ocean, increased ice-shelf calving, and stronger phytoplankton blooms. Low sea-ice area negatively affects krill and fish populations, ultimately impacting their predators and humans operating in Antarctica.
Researchers developed a magnetic resonance technique to detect fentanyl hydrochloride in sealed packages, offering a non-invasive solution. The technique uses Nuclear Quadrupole Resonance (NQR) spectroscopy, which can identify key target nuclei in the drug, allowing for effective detection and potential diversion of large quantities.
Combinative and transformative creativity are recognized by Julio M. Ottino as essential components of creative breakthroughs. Historical figures such as Filippo Brunelleschi and Niels Bohr demonstrate creative fluidity, while AI's limitations in transformationally creative tasks are discussed.
A new study predicts cognitive abilities from brain scans by integrating multiple MRI modalities through a technique called stacking. The method achieves substantial predictive accuracy and improved test-retest reliability, suggesting a more reliable neural marker of cognitive function.
A study on Americans' behavior after the COVID-19 shutdowns found that people often delay returning to lost pleasures. Participants reported waiting longer to satisfy their first time back, indicating a desire for a special experience.
Natural climate solutions like wetland restoration and reforestation face various barriers, including lack of funding, insufficient information, and ineffective policies. The study highlights the need for collaborations across sectors and disciplines to unlock their full potential.
Global survey of 4,351 participants reveals that people in more equal and wealthy societies are more likely to condemn excessive wealth. Moral condemnation is linked to values related to equality, purity, and younger age, suggesting concerns about the corrupting influence of wealth.
A non-DNA based test using a nanopore device and AI model can identify live viruses in patients within minutes. The system showed 100% sensitivity and 71.4% specificity for detecting herpesvirus, with improved accuracy when conditions were adjusted.
A recent study evaluates GPT-4's performance in solving letter-string analogies, a task requiring basic counting and relational understanding. The model excels at this task, demonstrating emergent analogical reasoning capabilities that may be supported by structured operations and relational representations.