3821 papers
This paper demonstrates that a parameter-conditioned graph neural network can learn calibrated, transferable node-level infection risk across diverse network topologies to effectively guide targeted interventions, although its performance is limited by significant structural heterogeneity in unseen graphs.
This study challenges the traditional view of gradual brain expansion by demonstrating that significant increases in hominin encephalization occurred late in human evolution through a stepwise pattern that tightly co-evolved with accelerating cultural complexity during the Middle to Late Pleistocene.
This study utilized phase-anchored Gaussian decomposition to characterize bilateral in-shoe plantar loading during the volleyball spike jump's final plant phase, finding that while a double-Gaussian model effectively represented the loading waveforms, no significant associations were identified between these derived features and performance or kinematic outcomes.
This study reveals that the unusual, exact threefold 8-residue repeat in the N-terminal region of the Arabidopsis FLK protein is a lineage-restricted architectural feature that is statistically improbable given local amino acid composition alone, suggesting a potential functional role for precise residue order beyond bulk composition.
This paper establishes an analytical feasibility boundary demonstrating that while externally imposed confinement can theoretically influence closed mitosis, the extreme sensitivity of the resulting pressure thresholds to spindle-force calibration and load transmission parameters renders specific numerical predictions unreliable, thereby identifying a critical identifiability limit that necessitates precise experimental measurement of these quantities before confinement can be quantitatively assigned a biological role.
This paper presents and experimentally validates a four-stage, two-layer coarse-grained semantic-geometry framework that enforces physical feasibility constraints to successfully reproduce standard photosynthetic light-response curves while explicitly rejecting nonphysical parameter estimates.
This study found that while sweet basil oil exhibited moderate antibacterial activity against *Pseudomonas aeruginosa*, hemp oil showed negligible effects, and their combination did not enhance efficacy compared to sweet basil oil alone or the standard antibiotic streptomycin.
This preregistered study demonstrates that while individual kinetic parameter recovery from Ganzfeld-tACS EEG using Jansen–Rit models is practically unfeasible under Welch PSD, hybrid models incorporating explicit aperiodic terms significantly outperform null baselines, revealing that current data supports observation-layer generator modeling rather than physiological profiling and highlighting that specific datasets like ds004902 are unsuitable Ganzfeld proxies.
The study introduces FibrilNet, a network framework that integrates protein interactions with semantic context to demonstrate that systemic amyloidoses share a conserved molecular core while maintaining distinct, precursor- and tissue-specific environments.
This systematic review evaluates computer vision methods for fine-grained behavioral phenotyping in Autism Spectrum Disorder, highlighting their potential for severity assessment and differential diagnosis while identifying critical gaps in standardized datasets, evaluation metrics, and longitudinal research.