Browse Articles

Unlikely chemistry of life may have originated on the ocean floor

Group offers new scenario for how cells came to rely on phosphate molecules

Convolutional neural network and AI technology for colleges and universities students’ achievement performance prediction

Paws for thought

Emergent decadal predictability in Antarctic contribution to sea-level rise

Optimizing the sustainability assessment, mechanical properties and drying shrinkage of concrete containing groundnut shell ash using RSM modeling

The Lucy flyby of (52246) Donaldjohanson: A bilobed asteroid with tumbling rotation

The main belt asteroid (52246) Donaldjohanson (DJ) is a likely member of the Erigone asteroid family. This implies that DJ is a fragment of a larger parent body that was destroyed in a collision about 155 million years ago. We report observations taken during a flyby of DJ by the Lucy spacecraft. We found that DJ is composed of two heavily cratered lobes, connected by a smoother neck, with overall dimensions 8.8 kilometers (km) by 4.4 km by 3.1 km. The crater density is consistent with the Erigone family’s age, except for craters <0.4 km, which have been preferentially erased. DJ rotates slowly in a tumbling state, likely owing to spin-down by radiative forces. Surface spectra show iron-bearing phyllosilicates, indicating moderate aqueous evolution on the parent body.

A visualization tool for eye tracking time series analysis supporting teachers and psychologists

Abstract Eye Tracking (ET) can help improve understanding of visual attention in computer-supported interactive environments. In attention tasks, distinguishing between relevant target objects and distractors is crucial for effective performance, yet the underlying gaze patterns that drive successful task completion remain incompletely understood. Traditional gaze analyses provide limited insight into the temporal dynamics of attention allocation and the relationship between gaze behavior and task performance. When applied to complex visual search scenarios, current gaze analysis methods face several limitations, including the isolation of measurements in dynamic environments, visual stability, search efficiency, and the task-solving processes involved. This paper proposes an analysis tool, VisiTrail , that considers time-series eye-tracking data on task performance and gaze measures; temporal pattern analysis that reveals how attention evolves throughout task performance; object-click sequence tracking that directly links visual attention to user actions; and performance metrics that quantify both accuracy and efficiency of right actions. The proposed analysis is applied to data collected from the Mushroom Hunter serious game, a multilevel visual search task in which participants identify target mushrooms among distractors of increasing complexity across three difficulty levels. This tool focuses on two scenarios: subject-specific analysis and general analysis across all subjects from which a project collected data from Romania and Portugal. Subject-specific analysis uses only one participant’s data and yields two types of results: a first-level overall analysis that provides detailed insights and a multilevel analysis that compares performance across all three levels, where Level 1 presents the easiest task with few distractors, Level 2 increases difficulty by adding more similar distractors, and Level 3 is the hardest with many closely resembling distractors and more complex layouts. The generalized analysis reveals that gaze stability (Fixation %) was broadly consistent across both cohorts. Romanian participants exhibited faster median reaction times than Portuguese participants; however, given the small sample size (N=7 per cohort), the presence of data-quality anomalies in three Portuguese sessions, and the absence of inferential statistical testing, this difference should be regarded as preliminary and hypothesis-generating rather than conclusive. By using standardized parameters, the results reduce accidental interactions and hardware noise, providing a reliable methodological basis for preliminary analysis and highlighting the need for larger sample sizes to establish statistical validity.

Changes in wildlife activity patterns in response to war in Ukraine

Conflict zones are inherently hazardous and inaccessible for researchers, which results in a knowledge gap about the immediate effects of armed conflicts on the environment, particularly wildlife. We used camera-trap detections to investigate the impact of armed conflict on wildlife activity patterns before, during, and after the Russian occupation of the Chornobyl Exclusion Zone (Ukraine) in 2022 and compared it to the same period in 2021. Mammal species responded to armed conflict through immediate behavioral adjustments, including reduced activity during night and on dates when armed-conflict activities intensified. Our results provide insight into wildlife’s behavioral responses to armed conflict in real time and underscore the potential of camera trapping to quantify the ecological effects of war.

Revealing the lagged relationship between groundwater changes and land subsidence in Tianjin based on GRACE and InSAR observations

Ultrathin polymer membranes with locked intrinsic microporosity for hydrocarbon fractionation

Membrane technologies offer an energy-efficient alternative to conventional distillation for hydrocarbon fractionation, but they suffer from a trade-off between fast liquid transport and high molecular selectivity. We report a scalable approach to fabricate polymer membranes with stable interconnected pathways by locking in their intrinsic microporosity. This locking strategy reduces polymer swelling and preserves the subnanometer pore structure in hydrocarbon liquids, resulting in 10-fold higher permeance for synthetic crude oil compared with current state-of-the-art membranes. When applied to Arabian Extra Light crude oil, these membranes achieved excellent size- and class-based separation, removing 99.8% of hydrocarbons containing >15 carbon atoms and 93% of sulfur-containing components. These scalable membranes underpin processes providing rapid and selective hydrocarbon separation, enabling a more sustainable pathway toward crude oil refining.

Probing picometre-scale interlayer deformations via hyperbolic polaritons

Mitigation of transient beam loading in a compact high duty cycle injector via hardware upgrade and multi-objective optimization

CD4 ⁺ T cells impair tumor growth through IL-3 and TNF-dependent vascular damage

Most cancer immunotherapy strategies are focused on direct tumor killing by immune cells, especially T lymphocytes. Clinical and conceptual limitations of these approaches create a need for additional strategies. We identified a tumor stroma–targeting mechanism in which tumor antigen–specific CD4 + T cells inhibit tumor growth through myeloid cell and tumor necrosis factor (TNF)–dependent vascular damage. Multiplex immunofluorescence and single-cell and tissue transcriptomics showed that CD4 + T cells trigger the formation of perivascular myeloid cell clusters containing “classically activated” macrophages that produce TNF in response to T cell–derived interleukin-3. TNF causes intratumoral endothelial damage and blood supply disruption, which are associated with localized tumor cell death. Thus, intratumoral antigen-triggered T cell activation can mediate antitumor effects without direct recognition of living tumor cells, thereby avoiding many of the inhibitory mechanisms that limit anti-tumor immunity.

Deep residual networks for short-term load forecasting: an empirical study on the impact of network depth

A space telescope is falling to Earth. NASA is racing to rescue it

Vehicle will attempt a daring capture-and-boost mission to extend the life of the Swift observatory

A lightweight graph-enhanced deep learning framework for explainable cucumber leaf disease diagnosis

Russia plans deep quest for ‘endless oil’

Soviet-era theory touted by Putin’s former campaign manager claims oil deposits can form without organic matter

Graph attention network-enhanced multi-agent reinforcement learning for dynamic interception task allocation in counter-drone defense

Reanimating ELIZA, the world’s first chatbot <b>Inventing ELIZA</b> <i>Sarah Ciston, David M. Berry, Anthony C. Hay, Mark C. Marino, Peter Millican, Jeff Shrager, Arthur I. Schwarz, Peggy Weil</i> The MIT Press, 2026. 350 pp.

A simple program created in the 1960s laid the groundwork for future conversational computers

Energy harvesting potential of LiHfIrZ (Z = Si, Ge) quaternary Heusler alloys: a first-principles computational study