Browse Articles
Discover research articles across all indexed journals
Ethaline deep eutectic solvent under nanoconfinement: Unveiling structural and dynamical changes
Hybrid nanomaterials incorporating deep eutectic solvents (DESs) in porous hosts or at solid interfaces are gaining increasing attention for their potential interest across a wide range of applications. Under these conditions, the performances of DESs may be influenced by interfacial effects and spatial restrictions. In this study, we examined the effects of nanoconfinement on both the structure and molecular dynamics of the prototypical DES ethaline (a mixture of choline chloride and ethylene glycol) when confined within the cylindrical mesopores of SBA-15 (Dp ≈ 8.1 nm) and MCM-41 (Dp ≈ 3.5 nm) silicas, using neutron diffraction and quasielastic neutron scattering. It demonstrates that ethaline remains structurally homogeneous under confinement, showing no evidence of core–shell segregation within the pore cross section. The molecular dynamics of the confined ethaline preserve the key characteristics observed in its bulk state. Translational diffusion follows a jump-diffusion mechanism, with diffusion coefficients that remain remarkably close to the bulk values, showing only a modest reduction in MCM-41. A more pronounced increase in the residence time τ0 between translational molecular jumps is observed. It corresponds to roughly a factor of 3–8 under confinement in SBA-15 and reaches up to a tenfold enhancement in MCM-41 relative to bulk. Similarly, the characteristic relaxation time, τL, associated with the localized in-cage motion of ethaline, increases by ∼20% in SBA-15 and up to 50% in MCM-41. However, the molecular trajectories, modeled from the elastic incoherent structure factor, remain largely preserved under confinement, showing only a marginal reduction in intra-basin motional amplitudes.
CANet: A color-aware convolutional neural network with clinically grounded chromatic learning for skin lesion classification from dermoscopic images
Benchmarking mixed quantum–classical molecular dynamics for electronic strong coupling
Experiments indicate that collective coupling of molecular ensembles to confined optical modes can modify excited-state dynamics and photochemical reactivity. To describe such cavity-induced effects at atomic resolution, semi-classical molecular dynamics approaches have been developed that treat nuclear motion classically while describing the collective light–matter interaction within the Tavis–Cummings framework of quantum electrodynamics. Here, we benchmark mixed quantum–classical approaches, Ehrenfest dynamics, and Fewest-Switches Surface Hopping (FSSH) for simulating nonadiabatic dynamics of electronically strongly coupled carbon monoxide molecules. Their predictions are compared against numerically exact quantum dynamics simulations performed with the multi-configuration time-dependent Hartree method, which treats both electronic and nuclear degrees of freedom quantum mechanically. We find that the semi-classical approaches reproduce the qualitative features of the full quantum dynamics. Quantitative agreement is best achieved with FSSH when a decoherence correction is included. These results demonstrate that mixed quantum–classical methods provide a computationally efficient and quantitatively reliable alternative to fully quantum simulations for investigating nonadiabatic photochemistry under collective electronic strong coupling in systems beyond the reach of exact quantum treatments.
SignatureGuard: hybrid CNN–transformer model for signature verification and identification across Arabic and English datasets
Abstract Offline signature verification has a persistent Latin-script bias: most systems are built and evaluated on English datasets, while Arabic and other non-Latin scripts are largely absent from the benchmarking literature. SignatureGuard is a three-task framework that evaluates six hybrid CNN–transformer architectures on two offline signature benchmarks (one Arabic, ASVAR; one English, CEDAR) under a single shared preprocessing and training pipeline, enabling direct architectural comparison across writing systems. The three tasks are binary forgery detection, multi-class biometric identification, and forgery source identification. To address the Arabic data gap, we publicly released ASVAR: 3471 images (1712 genuine, 1759 forged) from 70 individuals. Hybrid pairings of EfficientNetB7 or ResNet50 with the Vision Transformer (ViT-B/16) achieve test accuracies of 98.2% and 98.4% on forgery detection, macro-F1 above 0.97, and Cohen’s $$\kappa$$ above 0.96; 95% Wilson confidence intervals ( $${\pm }$$ 1.5 pp) confirm these are not artefacts of finite test-set size. MobileNetV2-based hybrids trail by at most 0.7 percentage points. Architectural rankings are broadly consistent across both datasets. All results are obtained under a seen-writer, image-level 80/10/10 split—a closed-set protocol that supports reproducible architectural comparison but overestimates real-world deployment performance; a signer-disjoint evaluation is identified as the primary follow-on experiment. An information-theoretic argument demonstrates that the hybrid classification head cannot perform worse than either frozen component in isolation. Confusion-matrix analysis, multi-seed validation, and backbone fine-tuning are recommended as extensions.
Practical considerations for finite concentration molecular dynamics simulations
Understanding concentrated electrolytes requires a theory that spans local hydration and mesoscale interfacial assembly. We present an integrated workflow—Solvation Characterization via Optimized Probability Ensemble averaging (SCOPE)—that combines (i) enhanced sampling focused on a single Li+ ion, (ii) reweighting of biased trajectories to recover equilibrium microstate probabilities, and (iii) a chemical-potential correction that accounts for the limited reservoir of free water in finite simulation boxes. Applied to LiCl(aq) across 0.5–26M and 283–313 K, this approach reveals a simple organizing principle: solvated ions dominate at low concentrations; contact ion pairs emerge at intermediate strengths; and aggregated Li–xCl clusters become most stable at the solubility limit. The resulting free-energy trends predict temperature-dependent solubility in close agreement with experiment and clarify the role of interfacial nucleation in precipitation. Beyond the simple LiCl(aq) salt considered here, SCOPE offers a transferable strategy for characterizing speciation and phase behavior in concentrated liquid systems where collective coordinates and rare events dominate.
PathInteract: a video analysis suite for capturing and analyzing pathology slide reviewing interactions
Variational estimation of generator invariant subspaces
We present VEGIS (Variational Estimation of Generator Invariant Subspaces), a variational method to approximate invariant subspaces of the infinitesimal generator of reversible diffusion processes. The method represents a trial subspace by neural networks and optimizes a Dirichlet-form trace objective that can be evaluated using only equilibrium samples and gradients of the network outputs. After training, the learned trial space can be diagonalized to recover generator eigenfunctions and eigenvalues or transformed by PCCA+ to obtain membership functions associated with metastable sets. In addition, we introduce a VEGIS-driven sampling strategy in which a rough approximation of the dominant slow mode is used to modify the effective diffusivity while preserving the invariant density. Numerical results on low-dimensional and molecular systems demonstrate the accuracy and flexibility of VEGIS.
Chronic exercise changes the baseline properties of central monoaminergic systems, identified neurons and related behaviors
Abstract Physical exercise influences monoamine-dependent behaviors of mammals and some invertebrates. However, little is known about how a specific neuron adapts its biophysical properties to acute or repeated physical activity. Here we tested the hypothesis that central monoaminergic neurons develop a state of readiness for exercise by changing their baseline biophysical properties during regular training. In the mollusc Lymnaea stagnalis , acute and two-week daily two-hour exercise (crawling in low water) changed the baseline rate of respiration and locomotion. Consumption was activated by regular exercise, while growth rate decreased. Under both acute and baseline chronic conditions (one day after completion of training), the dopaminergic neuron that controls respiration was hyperpolarized, and dopamine levels in the brain were reduced. Two-hour rest after acute exercise abolished these effects. Serotonergic neurons that control locomotion were depolarized both in situ and after being completely isolated from the network under both acute and chronic conditions. Our findings suggest that regular exercise alters the baseline biophysical properties of central monoaminergic neurons, potentially through an anticipatory mechanism preparing the nervous system for subsequent exercise.
Unraveling intramolecular charge transfer coordinate by optical polarization dependent coherent vibrational spectroscopy
Coherent vibrational wavepackets (CVWs) observed in femtosecond time-resolved spectroscopy provide direct access to excited-state structural dynamics and have been widely used to investigate ultrafast photoinduced processes. Their mechanistic utility, however, has been largely limited to reactions occurring within the vibrational coherence time. Here, we demonstrate that polarization-dependent CVW analysis can be extended to weakly reactive, non-ultrafast processes and used to identify vibrations associated with the reaction coordinate. As a model system, the solvation-driven intramolecular charge-transfer (ICT) dynamics of thioflavin T are investigated. In 1-propanol, ICT proceeds more slowly than vibrational dephasing, allowing latent reactivity to remain encoded in the polarization dependence of the CVW. Two vibrational modes exhibit distinct polarization dependence in the stimulated-emission band of the intermediate state, and their assignment as ICT-related modes is supported by quantum-chemical calculations. These results demonstrate that polarization-dependent CVW analysis provides a practical route to identifying reactive vibrations beyond the conventional ultrafast limit and expands the mechanistic scope of coherent vibrational spectroscopy for a broader range of photoinduced dynamics.
Experimental analysis and optimization of driving smoothness simulation of the SUV vehicle based on ADAMS/car
Abstract Sport utility vehicles (SUVs) often encounter challenges in simultaneously balancing handling stability and ride comfort when driving on rough roads. In order to ensure driver safety and improve ride comfort, this study utilizes the multibody dynamics simulation software Automatic Dynamic Analysis of Mechanical Systems (ADAMS) to build a comprehensive vehicle dynamics model of a family SUV. Initially, a front suspension parallel wheel hop test and a rear suspension reverse wheel hop simulation test are conducted to verify the accuracy of the suspension simulation models. Subsequently, simulation analyses of the complete vehicle’s handling stability under steady-state cornering conditions and driving smoothness under random road excitations are performed. Finally, the Latin hypercube experimental design method in ADAMS/Insight is adopted to conduct 256 sets of multi-objective optimization iterative simulations on the stiffness and damping parameters of the front and rear suspensions. In response to the reviewers’ comments, the optimization target is clarified as the total frequency-weighted acceleration RMS used for ride-comfort evaluation, rather than an unsupported linear weighting of the three vibration components. The 60 km/h Class B-road condition is also defined as the nominal design condition, while the other speeds are interpreted as off-design robustness checks rather than evidence of global Pareto optimality over the complete speed range. The experimental and simulation results indicate: (1) The established multibody dynamic model effectively simulates the structural and mechanical constraints of the actual vehicle. Key alignment parameters, including kingpin inclination angle, kingpin caster angle, camber angle, and toe angle of the front suspension, are all within the permissible design ranges. Furthermore, the relative errors for the rear suspension stiffness and roll center height are both within 3%, effectively reflecting the real vehicle’s operational status. (2) The optimized suspension parameters improve the ride comfort of the SUV, achieving a maximum observed 39.9% reduction in the total weighted acceleration RMS at 100 km/h in the reported simulation case. The handling-stability indices change only slightly after optimization: the steering gradient changes from $$U=-23.912$$ to $$U=-24.081$$ and the roll angle gradient from $$K_{\varphi }=3.862$$ to 3.861. However, because the negative steering-gradient values indicate an oversteer tendency under the adopted sign convention, the optimization should be interpreted as preserving the reported handling indices within this simulation condition rather than resolving the baseline oversteer characteristic.
An automated clustering workflow for molecular simulation data
Efficient analysis methods are needed that are able to cope with the huge size of modern molecular simulation datasets. Typical analysis workflows involve dimensionality reduction algorithms that improve the comprehensibility of the dataset by lowering its dimensions. Other common elements of the analysis are clustering algorithms that divide the dataset into groups according to a predefined characteristic, often with an emphasis on structural homogeneity. The application of such a clustering algorithm can help identify important metastable states of the system. In this paper, we revisit a clustering workflow described by Hunkler et al. [J. Chem. Phys. 158, 144109 (2023)] and provide a refined and automated version. We apply this workflow to a dataset of atomistic simulations of the 76 amino-acid residue protein ubiquitin (Ub) to illustrate its strengths and characteristics. The automated clustering workflow combines two dimensionality reduction algorithms, cc_analysis and EncoderMap, with the clustering algorithm HDBSCAN and a root mean square deviation-based sorting criterion. Due to an iterative approach, it is especially suitable for highly efficient categorization of large datasets of structures into structurally homogeneous clusters of different densities and sizes. Small adaptations to the original clustering workflow ensure considerable improvements in the quality and homogeneity of the obtained clusters.
Brief use of a passive shoulder exoskeleton modifies muscle activation after removal
Abstract Wearable exoskeletons are increasingly implemented in industrial environments to alleviate physical strain; however, their impact on motor control strategies, particularly following device removal, remains insufficiently understood. This exploratory study examines the short-term post-effects of a passive upper-limb exoskeleton, the Paexo Shoulder, which uses spring-based assistance to support arm elevation. Twenty healthy participants were assigned to either an Exposed group (n = 10), who performed an overhead reaching task while wearing the exoskeleton for 25 minutes, or a Baseline group (n = 10), who completed the same task without assistance. Kinematic and muscular activity were assessed using motion capture and electromyography before and after the intervention. Post-removal analyses revealed limited changes in classical performance and kinematic outcomes during the first 10 minutes after exoskeleton removal. Movement accuracy remained largely unchanged, task completion time showed only a non-robust tendency to decrease, and joint range-of-motion metrics showed directional but non-significant changes after correction for multiple comparisons. In contrast, more detailed indicators of motor strategy suggested short-term neuromuscular reorganization. Estimation plots indicated tendencies toward altered inter-muscle coordination, including changes in anterior deltoid and biceps contribution. The clearest statistically supported effect was observed in biceps activation, which decreased after repeated task execution in the Baseline group but increased after exoskeleton exposure. Overall, these findings suggest that short-term use of a passive shoulder exoskeleton produces limited modifications in end-effector performance and joint range of motion, while more subtle post-removal effects may emerge at the level of muscle activation and coordination strategies. These results highlight the importance of (1) assessing post-use effects of occupational exoskeletons, and (2) complementing classical biomechanical metrics with coordination and electromyographic analyses.
Master equation reduction using state projection and singular perturbation arguments
The kinetic Monte Carlo (KMC) method is becoming increasingly popular in the research of heterogeneous catalysts, yielding fundamental understanding of, e.g., their activity, selectivity, and rate of poisoning. KMC simulations can provide a link between ab initio electronic structure calculations and experiments by simulating at larger scales than is feasible for other methods, such as molecular dynamics (MD). A common problem for many systems studied using KMC, however, is the timescale separation of events considered, often leading to considerable computational effort and time spent on modeling fast events that are quasi-equilibrated and, thus, “uninteresting” from a kinetics perspective. In an attempt to make these simulations more efficient, two broad classes of schemes have been developed: those separating the state-space into “superbasins” and treating the fast events in approximate ways, and those employing absorbing Markov chain theory. In this paper, we consider the master equation underpinning KMC simulations in cases where the slow and fast events can be clearly defined, allowing us to “lump” the states into superbasins and use a perturbation expansion to derive approximations of the full solution. We thus develop a framework that unifies these two classes of schemes and enables us to derive a hierarchy of approximations that transcend the quasi-equilibration approximation. The validity of these approximations is then examined by testing them on four model systems, two of them relevant to on-lattice reaction kinetics.
Fuzzy rule-based network: a fuzzy logic-based interpretable and modular machine learning model
Abstract This paper introduces the Fuzzy Rule-Based Network (FRBN or FN), a novel machine learning architecture designed to bridge the gap between high-performance modeling and human-understandable decision-making. By combining the layered, hierarchical structure of artificial neural networks (NN) with the logical transparency of fuzzy systems, the FN operates as an interpretable “gray-box” model. Unlike traditional “black-box” neural networks that obscure their internal logic, the FN explicitly encodes its learned knowledge using fuzzy “if-then” rules at every layer, providing structural modularity and insight into how decisions are made. To optimize both FN and NN models, various strategies were comprehensively evaluated, including first-order gradient methods such as Adam, the quasi-second-order Levenberg–Marquardt algorithm, the gradient-free Bacterial Evolutionary Algorithm (BEA), and a hybrid Bacterial Memetic Algorithm (BMA) implemented for both neural and fuzzy network models. The predictive accuracy, structural modularity, and interpretability of the proposed FN model were assessed on several synthetic regression benchmarks, including the sinc function (where the FN model achieved a best-fold validation MSE of $$6.2 \times 10^{-8}$$ ), a multidimensional trigonometric dataset and validated on an industrial Micro-Electromechanical Systems (MEMS) sensor dataset. These experimental results demonstrate that the FN model provides a solution competitive with the NN model in terms of accuracy and training efficiency. The experiments demonstrate how the structural modularity of the FN model enables the reduction of parameters through the post-hoc extraction of localized sub-models. These specific fuzzy rule-sets maintain the exact predictive accuracy of the full model within their designated sub-intervals. The study further illustrates how the significance score identifies which specific rules govern particular subdomains, while the minimum antecedent coverage ratio evaluates the spatial scope of rules to distinguish between global and local influences. Additionally, the model’s interpretability is validated on the MEMS dataset, where linguistic interpretations are associated with the fuzzy rules. The FN can be initialized using domain knowledge, which stabilizes and accelerates training convergence. Finally, the trained rules can be translated back into human-understandable if-then statements, representing a step towards transparent decision-making.
Coulomb interaction unlocks Majorana-mediated electron teleportation between quantum dots
We investigate quantum transport through a hybrid system consisting of two quantum dots (QDs) coupled via a pair of spatially separated Majorana zero modes (MZMs) with negligible coupling energy. The transport properties, especially the nonlocal correlation mediated by the MZMs, are studied with a focus on the role of the Coulomb interaction U between the QDs and the Majorana wire. Using the numerically exact fermionic dissipation equation of motion method, we calculate both the transient current and the current–current cross correlation noise spectrum. Our results demonstrate that in the non-interacting case (U = 0), destructive interference between the normal tunneling and anomalous tunneling channels suppresses electron teleportation between the dots. Introducing a finite Coulomb interaction U lifts this channel degeneracy, thereby establishing strong nonlocal correlations and enabling inter-dot electron teleportation. This effect manifests as a robust signal in the cross correlation noise spectrum, which is significantly stronger than that induced by a finite Majorana coupling energy ɛM. Our work proposes Coulomb interaction as an efficient and experimentally accessible control parameter for generating and detecting Majorana-mediated nonlocal transport in the topologically relevant long-wire limit (ɛM → 0).
Correction: Buried soils from the Holocene Humid Period in Wadi Shuwayhi, Al-Khashbah (Oman)
Designing the ground state is not enough: Lessons from the self-assembly of Archimedean shells
Designing particle interactions such that a target structure is the thermodynamic ground state is a central paradigm in self-assembly. However, ensuring that the target is lowest in energy and that obvious competitors are energetically penalized does not, by itself, guarantee successful assembly at finite temperature, since even under these conditions competing structures can reappear as minima of the free-energy landscape. Focusing on the colloidal Archimedean snub-cube, we compute the full free-energy landscape of all competing aggregates using a cluster-based thermodynamic approach. While the target structure is uniquely selected at the level of potential energy, we find that competing clusters can become thermodynamically favored due to entropic contributions, particularly when bond directionality is high. In this regime, incomplete structures, such as icosahedra, are stabilized despite their higher energy per particle, leading to a dramatic suppression of the target yield. Our results provide quantitative guidelines for inverse design strategies that explicitly account for entropy, bond flexibility, and experimental conditions.
Telephone-based assessment of functional outcome scores after proximal femur fracture: a valid alternative to in-person evaluation
Self-averaging parameter estimation for coarse-grained particle models
We introduce a parameter estimation method that utilizes microscopic data, specifically averages and correlations of selected microscopic observables, to determine the parameters of a stochastic differential equation governing coarse-grained degrees of freedom. The method is not only limited to static parameters found in the reversible part of the coarse-grained dynamics, such as those in the free energy function or potential of mean force but also extends to dynamic parameters, including friction coefficients. The method couples the stochastic differential equation with free parameters to dynamic equations for the parameters. The coupled system self-averages, according to Anosov–Kifer’s theorem, in such a way that the final state of the parameters gives coincidence between the microscopic and mesoscopic averages and correlations of selected observables. The method is validated in two examples: a Brownian particle in a harmonic potential, and a set of Brownian particles interacting hydrodynamically with the Rotne–Prager–Yamakawa mobility tensor. This latter case illustrates how the method can be used not only to determine coefficients but also state dependent transport properties—in this case, the position dependent form of the mobility tensor. The parameter estimation for these two models yields excellent results. Subsequently, we use the methodology to study a bimodal-mass Lennard-Jones fluid for which we infer both the potential of mean force between the heavy particles and its hydrodynamic mobility tensor.