搜索结果：Modes

For decades, target-distractor similarity has been known to induce distinct visual search modes. A highly salient target can pop out, suggesting parallel processing of all items irrespective of set size. By contrast, high similarity among items requires item-by-item assessment, a characteristic of serial search. Despite this long-standing distinction, search modes remain poorly defined due to confounding of behavioural measures by differences in local contrasts and display density just as neural correlates are confounded by distinct displays used to prompt different search modes. Here, we biased search mode by manipulating target-distractor similarity in inducer trials, while embedded test trials afforded both modes, allowing us to isolate neural signatures of serial and parallel search under visually identical displays. Behavioral results from 24 participants (21 female, 3 male) confirmed successful induction of distinct search modes. EEG decoding reliably discriminated search modes and generalized across inducer and test trials. Attentional deployment toward the target differed across search modes, revealing topographical differences in target location representations. The representation of target location was associated with response times, indicating when subjects swapped search mode from parallel to serial if target was not detected quickly. Moreover, the representation of search target diverged between search modes: A temporally stable pattern emerged during serial search, suggesting the maintenance of conjunction item in working memory, whereas the representations were dynamic in parallel search, likely reflecting the relevant feature. These findings demonstrate that search history shapes search mode, giving rise to clearly distinct neural dynamics even under visually identical stimulation.Significance statement Serial and parallel search distinction is recognized half a century ago. However, search modes remain poorly defined. This is because visual displays used to prompt distinct search modes confound behavioural measures and neural correlates of search modes. Here, we biased search modes in subsequent blocks by manipulating target-distractor similarity in a set of inducer trials. Embedded among inducer trials, test trials afforded both search modes, allowing us to isolate neural signatures of serial and parallel search under visually identical displays. We used EEG decoding to distinguish neural correlates of search modes and their influence on the modulation of target location and the representation of the target itself.

As vulnerable road users, pedestrians face a high collision risk in safety-critical traffic scenarios. When faced with approaching vehicles, pedestrians need to balance their desire to cross the road with the demand for safety, resulting in either normal walking or avoidance behaviors. Such uncertain decision affects the occurrence of collisions. Therefore, it is important to understand pedestrian decisions so that highly automated vehicles (HAVs) can better develop safe interaction strategies. To study these decisions under controlled conditions, we designed an immersive virtual reality (VR) experiment where participants encountered traffic scenarios with different spatiotemporal pressures. The experiment observed three distinct decision modes under increasing spatiotemporal pressures: a no-risk mode in which pedestrians predominantly cross normally, a game mode showing a mix of crossing and avoidance, and a life-saving mode dominated by avoidance responses. Based on these decision modes, we proposed a regression model to predict pedestrian decisions. The model achieved an average precision of 0.87 and 0.82 for predicting whether and how pedestrian avoid. Finally, real pedestrian-vehicle conflict data were used to validate the effectiveness of the experimental results. This investigation observes pedestrians' decision modes in various urgent scenarios and presents a decision model based on pedestrians' decision modes, reflecting their interaction logic with hazardous vehicles in traffic scenarios. We expect that the observed decision modes can help develop the safety algorithms of HAVs to achieve safe interactions with pedestrians on roads.

Traditional discrete humidity and hydrogen sensors suffer from large footprints and cross-interference, limiting applications in hydrogen-fueled systems and semiconductor manufacturing. Here, we demonstrate a plasmonic-photonic hybrid resonator (PPHR) for low-crosstalk dual-parameter detection via mode-division multiplexing. It integrates a polyvinyl alcohol (PVA)-guided mode resonance for humidity sensing and a Pd-Au plasmonic resonance for hydrogen sensing, yielding analyte-specific modes with strong spectral separation. Humidity is monitored from the guided-mode redshift induced by PVA swelling and refractive-index variation, whereas hydrogen is detected from the plasmonic-resonance blueshift caused by hydrogen-induced phase change in palladium (Pd). Simulations show negligible crosstalk between the two independent working modes. This platform provides a practical route to integrated real-time multiparameter optical sensing.

Generative AI tools such as ChatGPT now perform core cognitive operations-reasoning, synthesis, evaluation, and creative generation-on users' behalf, raising urgent questions for educational psychology about how AI use relates to cognitive development. Yet research on cognitive offloading has largely treated AI use as a unidimensional phenomenon, obscuring a theoretically consequential distinction: whether AI substitutes for the user's own thinking or scaffolds it. Drawing on the autonomous-dependent help typology and self-determination theory, the present study introduces and examines a distinction between dependent cognitive offloading (delegating core thinking to AI) and autonomous cognitive offloading (using AI as a scaffold while retaining cognitive agency). In a three-wave time-lagged survey study (N = 589 university students and early-career knowledge workers), we tested a dual-pathway model linking the two offloading modes to four perceived downstream cognitive outcomes-autonomous capability, creativity, deep processing, and independent judgment-through cognitive agency transfer and intrinsic motivation. Dependent offloading was positively associated with cognitive agency transfer and negatively associated with intrinsic motivation, which in turn were linked to poorer perceived outcomes. Autonomous offloading was positively associated with intrinsic motivation and more favorable perceived outcomes. Metacognitive monitoring attenuated the link between dependent offloading and cognitive agency transfer but did not buffer the negative motivational association. Notably, both offloading modes yielded comparable immediate benefits despite divergent downstream correlates, suggesting that potentially maladaptive AI engagement may be difficult for users to detect from immediate experience. These findings highlight the manner of AI engagement-not merely its frequency-as a key factor in understanding its associations with perceived cognitive functioning, and point to practical strategies for educators, learners, and AI tool designers seeking to harness AI without undermining cognitive autonomy.

Gyrokinetic simulations of a turbulence-reduced Wendelstein 7-X discharge-characterized by a steep density gradient, moderate temperature gradients, and low plasma beta-show that microtearing mode (MTM) turbulence dominates transport. The simulated heat and particle fluxes agree with experimental measurements, leaving MTMs as the only mode consistent with the data. These conditions, the quasi-isodynamic and nearly max-J magnetic configuration of the Wendelstein 7-X stabilize ion temperature gradient modes and density-gradient-driven trapped-electron modes, while moderate collisionality and low magnetic shear further enable MTM growth. Further nonlinear scans of the density gradient reveal a significant reduction in turbulent transport at the experimentally observed threshold, which we identify as an ion temperature gradient to MTM-dominated turbulence transition. These findings provide a robust explanation for the turbulence suppression and deepen our understanding of low turbulent transport regimes in optimized stellarators.

Evaluating machine learning in scientific domains requires separating correct predictions from correct reasons under realistic distribution shifts. We introduce PertReason, a knowledge-grounded benchmark and framework suite for cell-state--conditioned reasoning about perturbation effects. At its core, PertReasonQA is a benchmark that tests whether models can generate mechanistically faithful explanations while remaining robust to complex shifts, such as new cells and unseen perturbations. PertReasonQA combines single-cell genetic and chemical perturbation data across multiple cellular contexts with knowledge graphs, and dynamically conditions pathways on cell-specific basal states to avoid generic memorization. Evaluations on state-of-the-art models reveal systematic gaps between predictive accuracy and mechanistic reasoning. Specifically, these models exhibit failure modes largely invisible to standard benchmarks, such as deriving correct answers through flawed logic, ignoring cellular context, and generating directionally inconsistent mechanisms. As a reference probe of the benchmark, we present PertReasonLM, a large language model trained to align outcome predictions with context-specific mechanistic reasoning. Our model targets the identified failure modes by grounding rationales in context-specific pathways and tightening agreement between outcomes and mechanisms. Together, we provide a diagnostic framework for exposing and mitigating failures in faithful reasoning in data-rich scientific systems.

Purpose Mixed-reality (MR) surgical navigation requires an intraoperatively measured organ surface for deformable image-to-physical registration, yet commercial headsets often rely on manual stylus digitization or external sensors. This work evaluates whether Magic Leap 2 (ML2) onboard sensors can provide this surface without external hardware, and whether its depth sensor or an RGB learning-based reconstruction is more accurate. Methods Two paradigms were developed from identical ML2 input in a common SLAM-tracked frame. The first used the onboard short-range indirect time-of-flight (iToF) stream. The second recovered geometry from an RGB sweep using Depth Anything 3, with metric scale anchored by SLAM-tracked camera centers. Both were evaluated against optically tracked stylus ground truth on opaque plaster and translucent silicone liver phantoms, a surface-treated breast phantom, and an ex vivo porcine liver, with five runs per specimen. Sensor-based feasibility was also demonstrated in vivo on an anaesthetized pig. Results ML2 world-origin drift remained below 2 mm median across four perturbations, and the iToF sensor achieved approximately 2 mm absolute accuracy over 0.30-0.70 m. Accuracy depended on surface optics. The sensor-based method was more accurate on opaque plaster livers (RMSE 2.6 and 3.1 mm versus 4.7 and 5.0 mm), whereas the learning-based method was more accurate on translucent silicone livers (4.3 and 4.6 mm versus 4.8 and 5.3 mm). The sensor-based method also performed better on the sunscreen-coated breast phantom (2.9 versus 3.9 mm) and retained a modest advantage on the ex vivo porcine liver (4.2 versus 4.8 mm). In vivo, the sensor cloud showed a mean residual of 3.4 mm to its fitted surface. Conclusion A commercial MR headset can acquire intraoperative organ surfaces using onboard sensors alone. Sensor-based reconstruction is preferred when active depth returns are reliable, whereas RGB learning-based reconstruction is more robust to optical failure modes but requires adequate multi-view coverage.

Bright and broadband integrated photon sources at telecom wavelengths are essential for quantum communication and information processing. Materials with high nonlinear refractive index (n2), such as aluminum gallium arsenide (AlGaAs), enable efficient nonlinear interactions at low excitation powers in integrated cavities. Here, we demonstrate a broadband quantum frequency comb (QFC) in a low free spectral range (FSR) (45 GHz) aluminum gallium arsenide on insulator (AlGaAsOI) resonator, with up to 40 pairwise frequency-correlated modes across a 1.97 THz bandwidth in the C-band. The dense mode spacing enables scalable frequency multiplexing and frequency-bin encoding within a compact device operating at a low excitation power of 6.33 µW. The measured joint-spectral intensity (JSI) is well reproduced by a theoretical model, enabling predictive control of spectral correlations across the generated QFC. The moderate Q-factor provides a tradeoff between nonlinear enhancement, broadband phase-matching, and number of accessible modes, supporting dense multimode operation over a large spectral range. These results highlight the potential of AlGaAsOI resonators for scalable low-power QFC sources.

Scaling laws describe how model performance improves as the amount of training data increases, and recent theories such as the zeta law suggest that scaling behavior is influenced by the eigenspectrum of the model's latent representation. Here, we evaluated whether the distribution of discriminative signals across spectral modes predicts the future scaling behavior, for MRI transformers trained for disease classification. We trained three supervised 3D vision transformers (ViT3D, MINiT, and NIT) for Alzheimer's disease classification using 2,822 training scans from the Alzheimer's Disease Neuroimaging Initiative (ADNI); we compared their encoder spectra with that of a frozen self-supervised DINO ViT-B/16 encoder adapted to 3D MRI. The supervised models learned highly concentrated representations, with 90-96% of CLS-token variance captured by a single principal component, whereas DINO distributed signal across many latent directions. Via spectral expansion of the Mahalanobis signal, we found that supervised training concentrated disease information into a single dominant mode, while self-supervised training produced a richer spectral geometry with higher effective rank and discoverability. This led to different scaling behavior: supervised models exhibited flatter AUC( N ) curves, yet DINO continued to improve as sample size increased, gaining 11.0 percentage points from N=50 to N=2,822. Overall, the spectral distribution of the discriminative signal, for these different encoder types, influenced how much performance remained discoverable as sample size increased. Distributed representations may retain signal across many latent modes and continue to improve with additional data, whereas concentrated representations tend to exhaust most of the discoverable signal at much lower sample sizes.

Artificial Intelligence (AI) is rapidly transitioning from experimental research to daily medical practice, yet the medical community's understanding of these tools remains largely confined to visible 'front-end' applications with which the clinician can directly interact (such as decision support systems and conversational agents). This perspective overlooks the proliferation of 'back-end' AI, the algorithms that are embedded within devices, systems, and hospital infrastructure that silently reconstruct data while remaining invisible to clinicians. This paper presents a clinician-oriented framework that distinguishes between these two categories and proposes a clinically oriented guide for their evaluation. We performed a narrative review of AI applications in minimally invasive therapy and medical imaging to categorize systems into 'front-end' (interactive/visible) and 'back-end' (embedded/invisible) modalities. We synthesized evaluation metrics from computer science and engineering literature, selecting those relevant for clinical safety and decision-making to create a practical literacy guide. Front-end and back-end systems require distinct validation strategies to ensure safety. while front-end evaluation must prioritize decision quality, spatial precision, and human-computer interaction to mitigate risks like automation bias, back-end evaluation requires rigorous technical benchmarking of signal fidelity and temporal latency to ensure that algorithmic reconstruction does not distort clinical reality. To facilitate this, we developed a structured inquiry framework to guide clinicians in auditing these systems for data provenance, transparency, and failure modes. Crucially, we emphasize that mathematical optimization does not guarantee clinical efficacy; technical metrics must always be paired with specific clinical contexts to ensure they align with patient-centered outcomes. Clinical safety in the AI era demands 'algorithmic literacy'. By applying this front-end/back-end framework and understanding key technical metrics, medical professionals can better identify failure modes, ensure data integrity, and maintain clear lines of clinical accountability, shifting from passive consumers to active evaluators of medical technology.

We report controllable two-dimensional (2D) Hermite-Gaussian (HG) lasing in a compact plano-concave cavity by exploiting the intrinsic birefringence of an a-cut Nd:YVO4 crystal. The birefringence-induced astigmatism provides modal discrimination against tilted one-dimensional (1D) competitors, allowing axis-aligned HGm,n modes to oscillate under 2D off-axis pumping. By increasing the birefringent crystal length, the output evolves from tilted 1D modes to axis-aligned 2D HG modes. We further show that cavity-length design can enhance the same modal discrimination, enabling 2D HG oscillation in a shortened-cavity configuration. These results verify a simple and compact route to controllable two-index HG lasing without pump shaping or intracavity loss elements.

A multi-mode immunofiltration detection assay is introduced, which uses Prussian blue-gold nanocomposites (PB@Au NPs) as nanoprobes and integrates four signal output modes: colorimetric (CM), enzymatic catalysis (CL), photothermal (PT), and surface-enhanced Raman scattering (SERS). Rapid vertical flow technology (RVFT) serves as the detection platform, achieving highly sensitive, qualitative, and diversified quantitative detection of brucellosis antibodies. First, PB@Au NPs were successfully synthesized using the wet chemical method, exhibiting excellent catalytic activity, a PT conversion efficiency of 36.2%, and strong Raman signal enhancement. These nanoprobes were then used to construct PB@Au-RVFT for brucellosis detection. In CM mode, the detection limit was 10 IU·mL⁻¹, which decreased to 4 IU·mL⁻¹ with enzyme-mediated CM amplification. PT and SERS modes further improved sensitivity, reaching detection limits of 0.167 and 0.144 IU·mL⁻¹, respectively, which are significantly better than traditional colloidal gold test strips. Additionally, the recovery rates for spiked clinical serum samples ranged from 94.59 to 105.31%, confirming the reliability of the method for practical applications. The proposed method enables multi-signal collaborative detection, enhancing anti-interference capacity and accuracy in complex samples. Its adaptability various detection environments give this technology the potential to conduct early screening and precise control of brucellosis in resource-poor areas. Meanwhile, it also provides a flexible and highly sensitive diagnostic strategy for other infectious diseases.

Although pancake-bonded organic diradicals are promising systems for molecular spintronics, their magnetic exchange couplings, 2J values, are sensitive to thermal fluctuations due to their highly delocalized, multicenter bonds. Here, we present a systematic evaluation of 2J values and the corresponding mode-specific exchange-vibronic couplings, d(2J)/dQi, across a set of covalently tethered bi-Blatter diradicals. Specifically, we varied the peri-fused bridging scaffold that connects the monomer radicals to 1,8-naphthalene, 5,6-acenaphthylene, 1,8-anthracene and 4,5-phenanthrene, allowing us to investigate how strain tunes magnetic interactions. To do this, we first resolve the methodological challenge of calculating accurate singlet-triplet gaps, by demonstrating that spin-flip time-dependent DFT, SF-TDDFT, and multireference single-point calculations performed atop broken-symmetry DFT, BS-DFT, geometries yield excellent quantitative agreement with experimental 2J benchmarks. The bridge topology dictates the static ground state, switching the coupling from strongly antiferromagnetic in the phenanthrene and naphthalene scaffolds to weak coupling in the anthracene analogue. We probe the magnetic exchange response by displacing the geometries along low-frequency normal modes. For these dynamic properties, BS-DFT provides overly-large and unphysical exchange-vibronic gradients as well as thermally-induced 2J fluctuations. By contrast, SF-TDDFT accurately tracks the multireference methods, both in sign and magnitude. We find that the exchange-vibronic coupling is tied to only a few low-frequency lateral shearing and inter-deck compression modes. By connecting the d(2J)/dQi gradients with harmonic thermal displacement models, we provide insights for suppressing thermally driven fluctuations in organic spintronic architectures.

We investigate a novel concept of a switchable diffractor in which the slit spacing can be switched by controlling carbon black (CB) nanoparticles. The slit width varies depending on the driving mode, where the nanoparticle positions are controlled by vertical electric fields generated between the planar electrode on the top substrate and the interdigitated electrodes on the bottom substrate. The diffraction angle of the proposed diffractor can be adjusted by controlling the slit spacing, operating in either wide-slit or narrow-slit modes. The diffractor also exhibits bistability, maintaining both slit modes without an applied field when the solution is optimized with a thickening agent. Furthermore, the device demonstrates a wide operating temperature range from -50°C to at least 100°C. In addition, the CB diffractor operates independently of the polarization state of incident light. Accordingly, this tunable diffractor utilizing CB nanoparticles exhibits outstanding potential for application in various advanced photonic devices.

Halide-substituted argyrodite materials have attracted increasing attention for energy applications since compositional tuning provides an effective strategy to modulate their structure and transport characteristics. While Li+-based halide argyrodites have been extensively studied, a unified composition-resolved understanding of Cu+-based halide argyrodites that integrates phase evolution, local structure, lattice dynamics, and electronic and ionic transport remain limited. In this work, we investigate Cu6PS5X (X = Cl, Br, I, Cl0.5Br0.5, Cl0.5I0.5, and Br0.5I0.5) within a combined experimental and computational framework. All compositions adopt an average cubic F4̅3m structure at room temperature, while local structural analysis reveals deviations from cubic symmetry consistent with a monoclinic Cc model involving PS43- tetrahedral tilting. 31P MAS NMR spectroscopy corroborates this local symmetry breaking through multiple distinct phosphorus environments arising from relative tetrahedral orientation rather than S2-/X- site disorder. Halide substitution modifies the Cu+ conductivity through changes in the activation energy and the Arrhenius pre-exponential factor, following the Meyer-Neldel behavior, with additional contributions from variations in jump distances and migration pathways. Direction-projected phonon density of states analysis identifies low-frequency Cu+ vibrational components along the crystallographic migration pathways. Analysis of the Meyer-Neldel slope further suggests phonon assisted ion hopping involving multiphonon excitation of low-frequency Cu+ vibrational modes. Together, these findings offer insight into structure-property relationships in Cu6PS5X and suggest that, alongside the migration energy landscape, the vibrational energy scale, thermal population, and directionality of mobile ion modes should be considered when interpreting ion transport, thereby providing a vibrational perspective for the design of solid-state ion conductors.

Ferroptosis has long been regarded as a cell-autonomous form of regulated cell death driven by iron-dependent lipid peroxidation. Recent work, however, suggests that ferroptotic commitment can extend beyond the initiating cell, spreading to neighboring cells and, in some contexts, across tissue-scale distances. Multiple, non-mutually exclusive modes of contagion have now been described, including reactive oxygen species-triggered death waves, direct membrane contact-dependent transfer, and extracellular vesicle-mediated paracrine signaling. These findings redefine ferroptosis from a single-cell execution pathway into a spatially coordinated, multicellular process. In this perspective, we integrate mechanistic insights with experimental evidence on ferroptotic contagion and propose a unifying, multiscale framework in which distinct modes of transmission operate over different spatial scales, from short-range membrane transfer to longer-range oxidative and extracellular relay mechanisms. We discuss how nonlinear redox amplification, membrane biophysics, and tissue architecture together determine the dynamics, limits, and patterning of ferroptotic injury in vivo. This emerging framework has important implications for developmental tissue remodeling, the progression of organ injury, and ferroptosis-based cancer therapy. More broadly, defining how ferroptotic contagion is initiated, constrained, and manipulated therapeutically may help establish ferroptosis as a fundamental organizing principle for understanding tissue-level regulation and pathological escalation.

Motivated by the recent discovery of anomalously large magnetic response of chiral phonons in dipolar magnets, we introduce the concept of pseudochiral phonons which are shown to emerge in multipolar magnets. We consider Raman active quantum phonons, such as doublet E_{g} phonons (d_{x^{2}-y^{2}},d_{3z^{2}-r^{2}}) in cubic crystals, which feature a symmetry-allowed linear coupling to local quadrupolar moments. We show that distinct multipolar orders can imprint distinct patterns of degeneracy breaking for phonons, with quadrupolar orders favoring nonchiral eigenmodes and time-reversal breaking octupolar orders favoring unconventional pseudochiral phonons. We compute the temperature dependent phonon matrix Green's function using a path integral approach where "fast" phonon modes sense "slow" pseudospin fluctuations over a thermal background sampled using Monte Carlo simulations. We propose helicity-resolved Raman spectroscopy of these pseudochiral phonons as a probe of hidden octupolar order in quantum materials such as Ba_{2}CaOsO_{6} and PrV_{2}Al_{20}.

Herein, the interaction mechanisms between pepsin (PEP) and two p-phenylenediamine derivatives, DPPD and CPPD, were investigated via multi-spectroscopic, computational, and enzymatic methods. UV-vis absorption spectroscopy, fluorescence spectroscopy, and time-resolved fluorescence confirmed the formation of ground-state complexes with a static quenching mechanism. Binding constant analysis revealed a stronger, more endothermic, and entropy-driven binding of CPPD compared to DPPD, primarily driven by hydrophobic forces. Both compounds induced secondary structural changes in PEP, evidenced by an increase in β-sheet and a decrease in α-helix. Crucially, DPPD and CPPD exerted distinct effects on PEP conformation. Molecular dynamics simulations showed DPPD binding stabilized PEP's structure while CPPD increased its flexibility. Molecular surface electrostatic potential (MESP) analysis rationalized these differences: CPPD's more polarized and extensive electrostatic field, stemming from its asymmetric cyclohexyl substituent, fostered stronger interactions but greater structural perturbation. These structural changes provided a direct rationale for the observed inhibition of PEP's enzymatic activity, where DPPD, despite weaker overall binding, induced specific active-site changes causing more pronounced inhibition. The antioxidant activities of both ligands decreased upon complexation. This study elucidates how subtle ligand structural differences dictate binding modes and functional outcomes in protein-ligand systems.

Biosensors are pivotal for detecting foodborne and waterborne hazards due to their portability, low cost, and rapid response. However, performance often degrades in real samples, where complex matrices reduce sensitivity and specificity and increase false positives/negatives. This systematic review synthesizes recent advances in biosensor platforms for monitoring contaminants in food and water, emphasizing how matrix properties govern analytical reliability and field usability. We present a matrix-first benchmarking perspective that compares biosensor performance across low-biomass waters, high-organic wastewater, and complex food extracts (high fat/protein, high particulate load, acidic, or high-salt), and summarizes dominant interference modes (fouling, nonspecific binding, ionic-strength shifts, and optical turbidity) alongside practical mitigation workflows (dilution/filtration, cleanup extraction, antifouling coatings, and microfluidic preconcentration). Beyond bacteria and viruses, this revision integrates pesticides as a third hazard class, covering enzyme-inhibition, aptamer, immuno-, and molecularly imprinted polymer sensing strategies, with representative case studies including glyphosate/AMPA, paraquat/diquat, chlorpyrifos, and atrazine. Overall, the matrix-first framework highlights design and workflow choices most likely to translate biosensors from proof-of-concept to deployable, multi-hazard monitoring tools for food and water safety.