Search for: All records

Abstract Epithelial cells stabilize ion concentrations and volume through coordinated membrane pumps, ion channels, and paracellular pathways, which can be modeled by classical single-compartment pump-leak equations (PLEs). Many epithelial functions, however, depend on the interaction between a cell and an enclosed luminal space, a geometry that cannot be captured by classical PLEs. To address this, we develop a two-compartment model consisting of an intracellular compartment coupled to a luminal compartment through the apical membrane, with both compartments interfacing an infinite extracellular bath and connected to it through the basolateral membrane and a paracellular pathway. Building on the five-dimensional single-cell PLEs, we formulate a ten-dimensional PLE system for this geometry and derive analytical equilibria and steady-state formulas for both the passive system and the Na⁺/K⁺-ATPase (NKA) driven active system. We characterize how these states depend on physiologically relevant parameters, analyze local stability across wide parameter ranges, and apply global sensitivity and robustness methods to identify the principal determinants of ion and volume homeostasis. Our focus is on closed epithelial systems in which the lumen volume relaxes to a steady state, rather than on fluid-secreting epithelia with time-varying luminal volume. We quantify apical and basolateral membrane potentials at steady state; the basolateral potential varies widely across parameter regimes, whereas the apical potential remains comparatively small in magnitude. The model reveals fundamental differences between basolateral and apical placement of the NKA, including the onset of luminal volume expansion when apical potassium recycling is insufficient. More broadly, this framework provides a mathematically tractable and physiologically grounded foundation for studying epithelial transport and for predicting conditions under which pump localization and conductance changes lead to stable function or pathological lumen expansion. 

This research uses virtual reality (VR) to immerse human subjects in an active school shooting scenario in order to generate ecologically valid models of school shooter movement and behavior. Historically, data recovered from U.S. school shootings has lacked the fidelity needed to model shooter movement; consequently, simulations of school shootings have had to rely on significant and unsupported assumptions about the movements of the shooter and/or victims. We asked human subjects to act as school shooters in a virtual reality simulation. We then recorded their movements, observations, and actions. Our results show that participant shooters are statistically equivalent to historical incidents with respect to aggregate engagement metrics (shot rate, victim rate, and accuracy) across most scenarios. Moreover, empirical models trained on participant data reduced prediction error by at least 15.5% compared to heuristic baselines and 16.7% compared to models trained on pedestrian data. Overall, this work provides a reproducible framework for data-driven modeling of shooter movement, supporting controlled simulation-based evaluation of response strategies. 

This paper uses virtual reality (VR) to immerse human subjects in an active school shooting scenario to generate ecologically valid models of school shooter movement and behavior. Historically, data recovered from U.S. school shootings has lacked the fidelity needed to model shooter movement; consequently, simulations have made significant assumptions about the movements of the shooter or victims. We therefore recruited participants in a human-subject study and asked them to act as a school shooter in a virtual reality simulation. We recorded movements, observations, and actions. Our results show that the behavior of our participants (n = 90) was statistically equivalent to that of real-world shooters in most scenarios, using bounds from historical data. Additionally, we found that participant data can be used to fit an empirical model that more accurately predicts shooter movements in both unseen simulated (−11.8%) and real (−24.4%) data (in terms of final path error) 

We introduce a new method for identifying genomic regions with introgressed DNA. The method does not search for introgressed haplotypes on individual chromosomes and does not rely on linkage disequilibrium. For this reason, it has more power than existing methods at deep time scales. It works with or without a genome from the donor population. We apply the method to three previously-documented episodes of admixture: α, in which Neanderthals contributed DNA to modern Eurasians, β, in which superarchaics contributed to Denisovans, and δ, in which superarchaics contributed to the ancestors of Neanderthals and Denisovans. For all three episodes, introgression is reduced in exons, promoter-like sequences (PRS), and proximal enhancer-like sequences (pELS) suggesting that selection has opposed introgression in these areas. On the other hand, introgression is elevated in distal enhancer-like sequences. Like previous studies, we find regions--"admixture deserts"--within which introgression is low. Those identified here are similar across all three episodes of admixture. Thus, selection seems to have opposed introgression in the same genomic regions three different times. Previous research has shown that different methods disagree about which regions are admixture deserts. Our method finds introgression in regions previously identified as deserts. 

Evacuation robots hold promise for facilitating efficient and safe building evacuations during emergencies. While studies have demonstrated that people will follow an evacuation robot [Nayyar 2019 28th IEEE International Conference on Robot and Human Interactive Communication (RO-MAN):1–6, 1, Robinette et al. 2016 11th ACM/IEEE International Conference On:101–108, 2], many practical hurdles remain. This paper uses a human subject study in a physical environment to investigate robot-guided evacuation of individuals versus small groups and considers different strategies for using robots to guide evacuees to exits. We further show that the data collected from these human subject studies can be used to train evacuee motion models which accurately predict the movement of the evacuee (9.9 cm mean error) while following the robot. We further show that this model can be used to predict the motion of an evacuee in a different environment and show that the accuracy of our model is superior to the more standard social forces model (SFM). The results from this research will contribute to the creation of evacuation robots and to the modeling and prediction of evacuee behavior.