The University of Western Ontario

This paper is devoted to studying the spatial dynamics of a nonlocal dispersal species modelwith annually synchronized emergence of adults. In the situation of a bounded domain,we show threshold dynamics of the adult population, and provide exact persistence criterion. In the situation of a spatially homogeneous unbounded domain, we obtain the existence and computation formula of spreading speeds, which coincide with the minimal wave speed for the traveling waves. The above results are obtained in both monotone and nonmonotone cases of maturation impulse function. Numerical simulations are carried out to demonstrate the theoretical results.

Accurate analysis of traffic flow (TF) data is crucial for the vehicular applications. Conventional deep learning models require task-specific training and are susceptible to high-frequency disturbances, degrading the feature representation capability. To overcome these limitations, this paper proposes a Token-based SelfSupervised Network (TSSN) that can learn TF features in both tokenization and task-agnostic manners. It provides a properly bootstrapped pre-training model for various downstream tasks. In support of the edge computing and vehicular cloud computing, the pooled computational resources facilitate real-time inferences of downstream models. In TSSN, TF data are segmented into tokens. A pretext task, named as Masked Token Prediction (MTP), is then developed to allow TSSN to understand the underlying correlations of TF by predicting randomly masked tokens. By utilizing MTP, TSSN is able to extract the high-level intrinsic semantics of TF, and provide general-purpose token embeddings, leading to improved overall performance and enhanced ability to adapt to different tasks. By substituting the last fully-connected layers with a group of untrained new layers and fine-tuning using small-scale task-specific data, TSSN can be utilized for a variety of downstream tasks in vehicular applications. Simulation results indicate that the TSSN enhances overall performance in comparison to state-of-the-art models.

bold xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">C loud- a ided m obile e dge n etworks (CAMENs) allow edge servers (ESs) to purchase resources from remote cloud servers (CSs), while overcoming resource shortage when handling computation-intensive tasks of mobile users (MUs). Conventional trading mechanisms (e.g., onsite trading) confront many challenges, including decision-making overhead (e.g., latency) and potential trading failures. This paper investigates a series of cross-layer matching mechanisms to achieve stable and cost-effective resource provisioning across different layers (i.e., MUs, ESs, CSs), seamlessly integrated into a novel hybrid paradigm that incorporates futures and spot trading. In futures trading, we explore an o verbooking-driven a forehand c ross- l ayer m atching (OA-CLM) mechanism, facilitating two future contract types: contract between MUs and ESs, and contract between ESs and CSs, while assessing potential risks under historical statistical analysis. In spot trading, we design two backup plans respond to current network/market conditions: determination on contractual MUs that should switch to local processing from edge/cloud services; and an o n s ite c ross- l ayer m atching (OS-CLM) mechanism that engages participants in real-time practical transactions. We next show that our matching mechanisms theoretically satisfy stability, individual rationality, competitive equilibrium, and weak Pareto optimality. Comprehensive simulations in real-world and numerical network settings confirm the corresponding efficacy, while revealing remarkable improvements in time/energy efficiency and social welfare.

Rate Splitting Multiple Access (RSMA) precoder design with the practical finite-alphabet constellations instead of Gaussian inputs has been addressed in this paper. Considering a multiuser (MU) multiple-input single-output (MISO) broadcast channel (BC) system, we derive a generalized expression of the achievable rate for each user, in a way that the derived expression is generically applicable, e.g., for both underloaded and overloaded cases. Building upon the achievable rate expression, we formulate a multi-objective problem that maximizes the weighted sum rate (WSR) of the considered system, which incorporates with the optimization of the RS precoder for both common and private symbol streams in RSMA. The emphasis here is that our derivation of the achievable rate expression, the problem formulation of the WSR and the optimization of the RSMA precoder all involve the finite alphabet constellation constraint. An iterative gradient descent algorithm with alternative optimization and line search methods is applied to solve the optimization problem. Numerical results show that RSMA can reach the maximum achievable WSR, under both underloaded and overloaded scenarios, with less transmit power compared to the traditional schemes, e.g., space division multiple access (SDMA) and power-domain non-orthogonal multiple access (NOMA). Moreover, thanks to its flexibility, RSMA subsumes both SDMA and NOMA as its subset to fit into different scenarios such as underloaded and overloaded cases with different constellation sizes.

Recent advances in distributed machine learning and wireless network technologies are bringing new opportunities for Internet of Things (IoT) systems, where smart devices are often wirelessly connected to collaborate, jointly completing tasks known as communication-dependent computing (CDC) tasks. However, due to the dependence of computing on communication and the presence of concurrent tasks, it remains a challenge to optimize the CDC task performance and efficiency while fulfilling multi-dimensional requirements, particularly with incomplete system information and dynamic environmental impacts. To overcome these challenges, we present a concurrent CDC task framework to facilitate tasks running efficiently in resource-limited IoT systems. To fulfill task requirements, we formulate a task orchestration and resource management problem for optimizing the overall utility of CDC tasks, where each task's utility is designed as a joint metric including deviated computing result and time efficiency. We then employ auxiliary graphs to capture the topological information of tasks and resources, and update weights based on the utility in dynamic environments. Subsequently, a multi-agent reinforcement learning algorithm is leveraged to make distributed decisions with incomplete information. Experiments demonstrate that the proposed approach outperforms baselines in terms of task performance and efficiency, indicating our solution holds great potential in IoT systems.

Document key information extraction (DKIE) methods often require a large number of labeled samples, imposing substantial annotation costs in practical scenarios. Fortunately, pseudo-labeling based semi-supervised learning (PSSL) algorithms provide an effective paradigm to alleviate the reliance on labeled data by leveraging unlabeled data. However, the main challenges for PSSL in DKIE tasks: 1) context dependency of DKIE results in incorrect pseudo-labels. 2) high intra-class variance and low inter-class variation on DKIE. To this end, this paper proposes a similarity matrix Pseudo-Label Bias Rectification (PLBR) semi-supervised method for DKIE tasks, which improves the quality of pseudo-labels on DKIE benchmarks with rare labels. More specifically, the Similarity Matrix Bias Rectification (SMBR) module is proposed to improve the quality of pseudolabels, which utilizes the contextual information of DKIE data through the analysis of similarity between labeled and unlabeled data. Moreover, a dual branch adaptive alignment (DBAA) mechanism is designed to adaptively align intra-class variance and alleviate inter-class variation on DKIE benchmarks, which is composed of two adaptive alignment ways. One is the intra-class alignment branch, which is designed to adaptively align intraclass variance. The other one is the inter-class alignment branch, which is developed to adaptively alleviate inter-class variance changes on the representation level. Extensive experiment results on two benchmarks demonstrate that PLBR achieves state-ofthe-art performance and its performance surpasses the previous SOTA by 2.11% ∼ 2.53%, 2.09% ∼ 2.49% F1-score on FUNSD and CORD with rare labeled samples, respectively. Code will be open to the public

Introduction Endoscopic Combined Intrarenal Surgery (ECIRS) has emerged as a promising technique for the management of large and complex kidney stones, potentially offering advantages over traditional Percutaneous Nephrolithotomy (PCNL). This study aims to evaluate best practices, outcomes, and future perspectives associated with ECIRS. Materials and Methods A comprehensive PubMed search was conducted from 2008 to 2024, using MESH terms and the following key words: "ECIRS" and "Endoscopic Combined Intrarenal Surgery" The search yielded 157 articles, including retrospective cohort studies, two randomized controlled trials (RCTs), and four meta-analyses comparing ECIRS with PCNL. Most important findings were summarized regarding indications, patient positioning, kidney access, tract size, surgical outcomes, and complications. Results ECIRS demonstrated higher stone-free rate, lower complication rate, and a reduced need for multiple procedures compared to traditional PCNL. Additionally, ECIRS has the potential to integrate new technologies to further enhance outcomes. Conclusion ECIRS demonstrates significant advantages in the management of large kidney stones. Future research should focus on well-designed RCTs to provide robust evidence of its efficacy, safety, and cost-effectiveness, potentially establishing ECIRS as the first option treatment for complex kidney stones. Keywords: Kidney; Lithotripsy; Urinary Calculi

Accurately estimating tool-tissue interaction forces during robotics-assisted minimally invasive surgery is an important aspect of enabling haptics-based teleoperation. By collecting data regarding the state of a robot in a variety of configurations, neural networks can be trained to predict this interaction force. This paper extends existing work in this domain based on collecting one of the largest known ground truth force datasets for stationary as well as moving phantoms that replicate tissue motions found in clinical procedures. Existing methods, and a new transformer-based architecture, are evaluated to demonstrate the domain gap between stationary and moving phantom tissue data and the impact that data scaling has on each architecture's ability to generalize the force estimation task. It was found that temporal networks were more sensitive to the moving domain than singlesample Feed Forward Networks (FFNs) that were trained on stationary tissue data. However, the transformer approach results in the lowest Root Mean Square Error (RMSE) when evaluating networks trained on examples of both stationary and moving phantom tissue samples. The results demonstrate the domain gap between stationary and moving surgical environments and the effectiveness of scaling datasets for increased accuracy of interaction force prediction.

Developing effective, sustainable strategies that promote social inclusion, reduce isolation, and support older adults’ wellbeing continues to be important to aging communities in Canada. One strategy that targets community-living older adults involves identifying naturally occurring retirement communities (NORCs) and supporting them through supportive service programs (NORC-SSPs). This qualitative descriptive study utilized semi-structured interviews to explore how older adults living in a NORC supported by an SSP, sought to build, and maintain, a sense of community during the COVID-19 pandemic. Analysis revealed how changes in context prompted changes in the program and community, and how despite lack of in-person opportunities participants continued to be together and do occupations together in creative ways that supported their sense of community. NORC-SSPs, like Oasis, play an important role in supporting older adults’ capacity to build strong, resilient communities that support wellbeing, during a global pandemic and in non-pandemic times.

Background Apathy in patients with Alzheimer's disease (AD) is associated with significant morbidity and is often one of the first neuropsychiatric symptoms to present in mild cognitive impairment (MCI). Apathy is associated with accelerated cognitive decline and atrophy in fronto-striatal regions of the brain. Previous work has shown a link between apathy and the APOE gene in the context of AD, as the APOE ε4 allele is already known to be associated with the onset of AD. However, other genetic associations with apathy are largely unexplored. Objective To examine whether interactions between genetic variants related to neurotransmitter systems and regional brain atrophy are associated with apathy in patients with MCI and AD. Methods In a sample of individuals with AD (n = 266), MCI (n = 518), and cognitively normal controls (n = 378), a partial least squares correspondence analysis modeled interactions between single nucleotide polymorphisms, structural whole-brain imaging variables, and apathy. Results An interaction was found between apathy, the possession of an APOE ε4 allele combined with minor homozygosity for the DAT1 (dopamine transporter 1) gene, and regional brain atrophy. This interaction was closely linked to the MCI and AD groups. Conclusions The results point to an association of a dopaminergic genetic marker and apathy in the AD continuum and may inform future design of clinical trials of apathy, as well as new treatment targets.

Healthcare systems are continuously evolving to respond to new geodemographic demands, among other challenges. At the forefront of this exercise of malleability, Emergency Departments (EDs) are often put to test as the default access point, while the rest of the system takes time to adapt. Once highly adaptable, years of cumulative strain have stressed the limits of the current organization of Emergency Departments (ED) within the healthcare system worldwide. The consequences are many, most notably for Emergency Physicians (EPs), who now face the highest rate of burnout among all medical specialties, with career resilience at an all-time low and diminished interest in the profession. Understanding how EDs are structured within their respective healthcare system provides a unique lens through which areas of improvement can be assessed. This paper discusses solutions to improve the overall structure of the healthcare system to help improve responsiveness, reduce relegation of tasks to the ED, and help improve working conditions and wellbeing for EPs.

The competing growth of two-dimensional (2D) and three-dimensional (3D) crystals of layered transition metal dichalcogenides (TMDCs) has been reproducibly observed in a large variety of chemical vapor deposition (CVD) reactors and demands a comprehensive understanding in terms of involved energetics. 2D and 3D growth is fundamentally different due to the large difference in the in-plane and out-of-plane binding energies in TMDC materials. Here, an analytical model describing TMDC growth via CVD is developed. The two most common TMDC structures produced via CVD growth (2D triangular flakes and 3D tetrahedra) are considered, and their formation energies are determined as a function of their growth parameters. By calculating the associated energies of 2D triangular or 3D tetrahedral flakes, we predict the minimum sizes of the critical nuclei of 2D triangular and 3D morphologies, and thereby determine the minimum realizable dimensions of TMDC, in the form of quantum dots. Analysis of growth rates shows that CVD favors 2D growth of MoS2 between 820 K and 900 K and 3D growth over 900 K. Our model also suggests that the flow rates of TMDC precursors (metal oxide and sulfur) in a long, cylindrical CVD reactor are important parameters for attaining uniform growth. Our model provides a compressive analysis of TMDC growth via CVD. Therefore, it is a critical tool for helping to achieve reproducible growth of 2D and 3D TMDCs for a variety of applications.

Contemporary materials for second language (L2) learning feature exercises on collocations (i.e., word partnerships such as catch fire), many of which require learners to select correct word combinations from two or more candidates. A few studies of the effectiveness of these selected-response exercises, which are essentially multiple-choice exercises, have found that learners later reproduce wrong collocations that they were exposed to in the exercise. However, it is unclear if this is a side effect of the exercises or if the re-emergence of wrong candidate responses is just accidental. The present study examines if wrong candidate responses that learners see in collocation exercises interfere with learners’ recall of the correct responses by having learners of L2 English tackle multiple-choice items on verb-noun collocations and verbally report two weeks later in a post-test what they remember about them. The verbal reports revealed that the learners recalled reading and responding to most of the exercise items, but for only one- third of them did they also recall which candidate response had turned out to be correct according to the feedback they received. Furthermore, for close to one- fifth of collocations that learners said they already knew at the start of the exercise, these participants mistook a wrong candidate response for the correct one when they revisited the exercise two weeks later. The findings call for a cautious approach to the implementation of selected-response exercises for collocation learning.

Objective assessments of shoulder motion are paramount for effective rehabilitation and evaluation of surgical outcomes. Inertial Measurement Units (IMU) have demonstrated promise in providing unbiased movement data. This study is dedicated to evaluating the concurrent construct validity and accuracy of a wearable IMU-based sensor system, called "Motion Shirt", for the assessment of shoulder motion arcs in patients awaiting shoulder replacement surgery. This evaluation was conducted by comparing Motion Shirt data with the Dartfish Motion Analyzer software during the Functional Impairment Test-Hand and Neck/Shoulder/Arm (FIT-HaNSA) test. Thirteen patients (age>50), who were awaiting shoulder replacement surgery, were recruited. The Motion Shirt was employed to measure angular shoulder movements in two planes during the FIT-HaNSA test. Simultaneously, two cameras recorded the participants' movements to provide reference data. Bland-Altman plots were generated to visualize agreement between the Motion Shirt and the reference data obtained from the Dartfish Motion Analyzer software. The data analysis on Bland-Altman plots revealed a substantial level of agreement between the Motion Shirt and Dartfish analysis in measuring shoulder motion. In Task-1, no significant systematic errors were exhibited, with only 3.27% and 2.18% of points exceeding the limits of agreement (LOA) in both elevation and the Plane of Elevation (POE), signifying a high level of concordance. In Task-2, a high level of agreement was also observed in Elevation, with only 3.8% of points exceeding the LOA. However, 5.98% of points exceeded LOA in POE for Task-2. In Task-3, focused on sustained overhead activity, the Motion Shirt showed strong agreement with Dartfish in Elevation (2.44% points exceeded LOA), but in POE, 7.32% points exceeded LOA. The Motion Shirt demonstrated a robust concordance with Dartfish Motion Analyzer system in assessing shoulder motion during the FIT-HaNSA test. These results affirm the Motion Shirt's suitability for objective motion analysis in patients awaiting shoulder replacement surgery.

Pre-clinical animal models of human brain tumors have been invaluable tools for studying cancer pathogenesis and exploring novel treatment modalities. Such models recapitulate important aspects of the human disease such as the stem-progenitor-differentiated cell hierarchy. Although powerful, we argue that animal models are inherently limited in their ability to phenocopy certain important aspects of human brain tumor biology. We specifically highlight the inability of mouse models to generate certain forms aggressive pediatric medulloblastoma likely owing to cellular, anatomic, and genetic differences between the human and mouse brains. Additionally, we review some limitations of human brain tumor derived cell lines and outline why they are a sub-optimal system for purposes of pre-clinical modeling. Below, we present the case for human stem cell-based models of brain tumors, focusing mainly on glioblastoma and medulloblastoma. Drawing on several recently published studies, we review the exciting progress that has been made towards modeling human brain tumors using two-dimensional adherent stem cell cultures and three-dimensional organoids. We identify the important advances arrived at using these human stem cell-based models and suggest opportunities for future work in this direction. In this review article, we aim to highlight the utility and promises of human stem cell-based models of brain tumors as a complementary system to traditional transgenic animal and cell line systems.

The presence of Li2CO3 has been identified as the cause of poor lithophilicity in garnet‐type Li7La3Zr2O12 (LLZO) solid‐state batteries. A Li2CO3‐free garnet is expected to enhance the Li/LLZO interface contact. However, permanently eradicating regenerative Li2CO3 from the LLZO surface is extremely challenging and the influence of regenerated Li2CO3 is often ignored. Herein, it is found that glossy Li2CO3 pellets can also be perfectly wetted by molten Li, contradicting the common belief that Li2CO3 is lithiophobic. Therefore, reducing the surface roughness of LLZO allows it to be directly wetted by lithium metal, regardless of the presence of Li2CO3. Additionally, smooth LLZO exhibits better air stability due to its reduced active area. The symmetric cell with a smooth LLZO pellet shows a low interfacial impedance of 2 Ω cm² and a high critical current density of 1.4 mA cm⁻² at 25 °C. This work highlights the surface physics of garnet which significantly influences its interface properties, apart from surface chemistry.

Background Endovascular thrombectomy (EVT) is the standard of care for patients with acute ischemic stroke (AIS) and intracranial vessel occlusion. Tandem occlusions (TO) comprise 20% of all anterior circulation AIS and are related to a poorer prognosis. The optimal EVT treatment strategy remains controversial. Our main objective was to determine if simultaneous endovascular treatment of intracranial and extracranial occlusions in patients with TO results in faster recanalization times, with similar efficacy and safety, compared with the sequential approach. Methods Single center, retrospective analysis of patients with TO undergoing EVT using the simultaneous or sequential technical approach. The primary outcome was puncture-to-final recanalization time. Secondary outcomes included modified Rankin scale (mRS) score at 3 months, 30 day mortality, and hemorrhagic transformation. Results We included 111 patients with TO (35 treated with the simultaneous approach and 76 treated with the sequential approach). Successful recanalization was achieved in 91.9% of cases, and the first pass effect was 50.5%, with no differences between groups. The simultaneous technique resulted in shorter puncture-to-final recanalization time (33.0 min (IQR 25.0–55.0) vs 52.0 (30.0–73.0), P=0.018), adjusting for number of passes, first pass effect, thrombolysis, age, and previous stroke (adjusted β −0.21 (95% CI −29.47 to −2.79); P=0.018). No significant differences were found in 30 day functional outcome, mortality, or rate of hemorrhagic transformation when comparing simultaneous and sequential techniques. Conclusion The simultaneous approach was effective, safe, and faster than the classic sequential approach in patients with TO. This result may obviate the debate over which occlusion should be addressed first during EVT.

Aims: When administered in early type 2 diabetes (T2DM), the strategy of 'induction' with short-term intensive insulin therapy (IIT) followed by 'maintenance' with metformin thereafter can yield outstanding glycaemic control, with some patients achieving A1c in the normal range of its assay. We thus sought to identify determinants of sustained normalisation of A1c in response to this treatment strategy. Materials and methods: In this study, adults with T2DM of mean duration 1.7 ± 1.4 years received induction IIT (glargine, lispro) for 3 weeks, followed by metformin maintenance either with or without periodic 2-week courses of IIT every 3 months for 2 years. Sustained glycaemic normalisation was defined by A1c <6.0% at 2 years. Results: Of 101 participants, 26 achieved A1c <6.0% at 2 years. At baseline, these individuals had lower A1c and fasting glucose than the other participants, along with better beta-cell function. During maintenance therapy from 3 weeks to 2 years, they had greater reduction of adiposity (body mass index: p = 0.02; waist circumference: p = 0.02), hepatic insulin resistance (HOMA-IR: p = 0.02) and ALT (p = 0.005), coupled with relative stabilisation of beta-cell function and glycaemia. On logistic regression analyses, significant independent predictors of normalisation of A1c at 2 years were baseline A1c (adjusted odds ratio [aOR] = 0.01 [95% CI 0.001-0.16], p = 0.001) and the changes in waist circumference (aOR = 0.77 [0.63-0.94], p = 0.012) and ALT (aOR = 0.90 [0.82-0.98], p = 0.019) during maintenance therapy from 3 weeks to 2 years. Conclusions: While lower baseline A1c and greater reduction in central adiposity predicted A1c <6.0% at 2 years as anticipated, the emergence of greater reduction in ALT as a concomitant determinant highlights the role of the liver in the achievement of sustained glycaemic normalisation.