Modelling to infer the role of animals in gambiense human African trypanosomiasis transmission and elimination in DRC

Gambiense human African trypanosomiasis (gHAT) has been targeted for elimination of transmission (EoT) to humans by 2030. Whilst this ambitious goal is rapidly approaching, there remain fundamental questions about the presence of non-human animal transmission cycles and their potential role in slowing progress towards, or even preventing, EoT. In this study we focus on the country with the most gHAT disease burden, the Democratic Republic of Congo (DRC), and use mathematical modelling to assess whether animals may contribute to transmission in specific regions, and if so, how their presence could impact the likelihood and timing of EoT. By fitting two model variants – one with, and one without animal transmission – to the human case data from 2000–2016 we estimate model parameters for 158 endemic health zones of DRC. We evaluate the statistical support for each model variant in each health zone and infer the contribution of animals to overall transmission and how this could impact predicted time to EoT. We conclude that there are 24/158 health zones where there is moderate or high statistical support for some animal transmission. However, – even in these regions – we estimate that animals would be extremely unlikely to maintain transmission on their own. Animal transmission could hamper progress towards EoT in some settings, with projections under continuing interventions indicating that the number of health zones expected to achieve EoT by 2030 reduces from 68 to 61 if animals are included in the model. With supplementary vector control (at a modest 60% tsetse reduction) added to medical screening and treatment interventions, the predicted number of health zones meeting the goal increases to 147/158 for the model including animals. This is due to the impact of vector reduction on transmission to and from all hosts.

Comparison of methodologies for modeling directional deep brain stimulation electrodes

Deep brain stimulation (DBS) is an established clinical therapy, and directional DBS electrode designs are now commonly used in clinical practice. Directional DBS leads have the ability to increase the therapeutic window of stimulation, but they also increase the complexity of clinical programming. Therefore, computational models of DBS have become available in clinical software tools that are designed to assist in the identification of therapeutic settings. However, the details of how the DBS model is implemented can influence the predictions of the software. The goal of this study was to compare different methods for representing directional DBS electrodes within finite element volume conductor (VC) models. We evaluated 15 different DBS VC model variants and quantified how their differences influenced estimates on the spatial extent of axonal activation from DBS. Each DBS VC model included the same representation of the brain and head, but the details of the current source and electrode contact were different for each model variant. The more complex VC models explicitly represented the DBS electrode contacts, while the more simple VC models used boundary condition approximations. The more complex VC models required 2–3 times longer to mesh, build, and solve for the DBS voltage distribution than the more simple VC models. Differences in individual axonal activation thresholds across the VC model variants were substantial (-24% to +47%). However, when comparing total activation of an axon population, or estimates of an activation volume, the differences between model variants decreased (-7% to +8%). Nonetheless, the technical details of how the electrode contact and current source are represented in the DBS VC model can directly affect estimates of the voltage distribution and electric field in the brain tissue.

Requirements for the containment of COVID-19 disease outbreaks through periodic testing, isolation, and quarantine

We employ individual-based Monte Carlo computer simulations of a stochastic SEIR model variant on a two-dimensional Newman-Watts small-world network to investigate the control of epidemic outbreaks through periodic testing and isolation of infectious individuals, and subsequent quarantine of their immediate contacts. Using disease parameters informed by the COVID-19 pandemic, we investigate the effects of various crucial mitigation features on the epidemic spreading: fraction of the infectious population that is identifiable through the tests; testing frequency; time delay between testing and isolation of positively tested individuals; and the further time delay until quarantining their contacts as well as the quarantine duration. We thus determine the required ranges for these intervention parameters to yield effective control of the disease through both considerable delaying the epidemic peak and massively reducing the total number of sustained infections.

Pupils returning to primary schools in England during 2020: rapid estimations of punctual COVID-19 infection rates

Drawing on risk methods from volcano crises, we developed a rapid COVID-19 infection model for the partial return of pupils to primary schools in England in June and July 2020, and a full return in September 2020. The model handles uncertainties in key parameters, using a stochastic re-sampling technique, allowing us to evaluate infection levels as a function of COVID-19 prevalence and projected pupil and staff headcounts. Assuming average national adult prevalence, for the first scenario (as at 1 June 2020) we found that between 178 and 924 [90% CI] schools would have at least one infected individual, out of 16 769 primary schools in total. For the second return (July), our estimate ranged between 336 (2%) and 1873 (11%) infected schools. For a full return in September 2020, our projected range was 661 (4%) to 3310 (20%) infected schools, assuming the same prevalence as for 5 June. If national prevalence fell to one-quarter of that, the projected September range would decrease to between 381 (2%) and 900 (5%) schools but would increase to between 2131 (13%) and 9743 (58%) schools if prevalence increased to 4× June level. When regional variations in prevalence and school size distribution were included in the model, a slight decrease in the projected number of infected schools was indicated, but uncertainty on estimates increased markedly. The latter model variant indicated that 82% of infected schools would be in areas where prevalence exceeded the national average and the probability of multiple infected persons in a school would be higher in such areas. Post hoc , our model projections for 1 September 2020 were seen to have been realistic and reasonable (in terms of related uncertainties) when data on schools' infections were released by official agencies following the start of the 2020/2021 academic year.

Enhancing Ocean Biogeochemical Models With Phytoplankton Variable Composition

Chlorophyll (Chl) is widely taken as a proxy for phytoplankton biomass, despite well-known variations in Chl:C:biomass ratios as an acclimative response to changing environmental conditions. For the sake of simplicity and computational efficiency, many large scale biogeochemical models ignore this flexibility, compromising their ability to capture phytoplankton dynamics. Here we evaluate modelling approaches of differing complexity for phytoplankton growth response: fixed stoichiometry, fixed stoichiometry with photoacclimation, classical variable-composition with photoacclimation, and Instantaneous Acclimation with optimal resource allocation. Model performance is evaluated against biogeochemical observations from time-series sites BATS and ALOHA, where phytoplankton composition varies substantially. We analyse the sensitivity of each model variant to the affinity parameters for light and nutrient, respectively. Models with fixed stoichiometry are more sensitive to parameter perturbations, but the inclusion of photoacclimation in the fixed-stoichiometry model generally captures Chl observations better than other variants when individually tuned for each site and when using similar parameter sets for both sites. Compared to the fixed stoichiometry model including photoacclimation, models with variable C:N ratio perform better in cross-validation experiments using model-specific parameter sets tuned for the other site; i.e., they are more portable. Compared to typical variable composition approaches, instantaneous acclimation, which requires fewer state variables, generally yields better performance but somewhat lower portability than the fully dynamic variant. Further assessments using objective optimisation and more contrasting stations are suggested.

Automated Inter-Ictal Epileptiform Discharge Detection from Routine EEG

Epilepsy is the most common neurological disorder. The diagnosis commonly requires manual visual electroencephalogram (EEG) analysis which is time-consuming. Deep learning has shown promising performance in detecting interictal epileptiform discharges (IED) and may improve the quality of epilepsy monitoring. However, most of the datasets in the literature are small (n≤100) and collected from single clinical centre, limiting the generalization across different devices and settings. To better automate IED detection, we cross-evaluated a Resnet architecture on 2 sets of routine EEG recordings from patients with idiopathic generalized epilepsy collected at the Alfred Health Hospital and Royal Melbourne Hospital (RMH). We split these EEG recordings into 2s windows with or without IED and evaluated different model variants in terms of how well they classified these windows. The results from our experiment showed that the architecture generalized well across different datasets with an AUC score of 0.894 (95% CI, 0.881–0.907) when trained on Alfred’s dataset and tested on RMH’s dataset, and 0.857 (95% CI, 0.847–0.867) vice versa. In addition, we compared our best model variant with Persyst and observed that the model was comparable.

Stability and Instability of the MaxWeight Policy

Consider a switched queueing network with general routing among its queues. The MaxWeight policy assigns available service by maximizing the objective function [Formula: see text] among the different feasible service options, where [Formula: see text] denotes queue size and [Formula: see text] denotes the amount of service to be executed at queue [Formula: see text]. MaxWeight is a greedy policy that does not depend on knowledge of arrival rates and is straightforward to implement. These properties and its simple formulation suggest MaxWeight as a serious candidate for implementation in the setting of switched queueing networks; MaxWeight has been extensively studied in the context of communication networks. However, a fluid model variant of MaxWeight was previously shown not to be maximally stable. Here, we prove that MaxWeight itself is not in general maximally stable. We also prove MaxWeight is maximally stable in a much more restrictive setting, and that a weighted version of MaxWeight, where the weighting depends on the traffic intensity, is always stable.

Critical speed estimated by statistically appropriate fitting procedures

Purpose Intensity domains are recommended when prescribing exercise. The distinction between heavy and severe domains is made by the critical speed (CS), therefore requiring a mathematically accurate estimation of CS. The different model variants (distance versus time, running speed versus time, time versus running speed, and distance versus running speed) are mathematically equivalent. Nevertheless, error minimization along the correct axis is important to estimate CS and the distance that can be run above CS (d′). We hypothesized that comparing statistically appropriate fitting procedures, which minimize the error along the axis corresponding to the properly identified dependent variable, should provide similar estimations of CS and d′ but that different estimations should be obtained when comparing statistically appropriate and inappropriate fitting procedure. Methods Sixteen male runners performed a maximal incremental aerobic test and four exhaustive runs at 90, 100, 110, and 120% of their peak speed on a treadmill. Several fitting procedures (a combination of a two-parameter model variant and regression analysis: weighted least square) were used to estimate CS and d′. Results Systematic biases (P < 0.001) were observed between each pair of fitting procedures for CS and d′, even when comparing two statistically appropriate fitting procedures, though negligible, thus corroborating the hypothesis. Conclusion The differences suggest that a statistically appropriate fitting procedure should be chosen beforehand by the researcher. This is also important for coaches that need to prescribe training sessions to their athletes based on exercise intensity, and their choice should be maintained over the running seasons.

Second-wave Dynamics of COVID-19: Impact of Behavioral Changes, Immunity Loss, New Strains, and Vaccination

An epidemiological model for COVID-19 developed earlier was extended to determine the effects of behavioral changes, immunity loss, and vaccination on second and subsequent wave dynamics of the pandemic. A model variant that distinguishes four demographic groups with different infection rates and mortality rates was developed to test the hypothesis that behavioral divergence between groups can explain both the larger incidence and lower mortality rate of COVID-19’s second wave. A model version that incorporates immunity loss was developed to test the hypothesis that immunity loss can explain the second wave. Simulations indicate that of the two hypotheses, only the former is consistent with observed trends. Nevertheless, loss of immunity can significantly increase total number of deaths in the long run, particularly in cases where vaccine distribution is barely sufficient to reach herd immunity. The observed trends are illustrated with detailed simulations of the progression of COVID-19 in the United Kingdom, including the appearance of new strains. The U.K. case study indicates the extent to which NPI can be relaxed during the distribution of the vaccine.

The Central Bank as Lender of Last Resort

In this chapter we review the function of the central bank as lender of last resort (LOLR), starting from the understanding of financial crises developed in the previous chapter. We recall long-established LOLR principles: proactive lending, inertia of the central bank risk control framework, and risk endogeneity. Because of its systemic role, a central bank should not tighten its collateral framework in a crisis, as restrictive policies are likely to not only increase the overall damage done by a crisis to society, but to even increase central bank losses. We explain in more detail the main reasons why a central bank should act as LOLR: prevent negative externalities from fire sales; its unique status as institution with unlimited liquidity; its status as a risk-free counterparty making others accept to deliver collateral to it even at high haircuts; and its mandate to preserve price stability. We distinguish three different forms of LOLR: elements built into the regular operational framework; readiness to relax parameters in a crisis; and provision of emergency liquidity assistance to individual firms. We then discuss what could be the optimal propensity of a central bank to engage in LOLR activities and outline possible trade-offs. Last but not least, we develop a bank-run model which highlights the role of asset liquidity and central bank eligible collateral. We calculate through a model variant with binary asset liquidity and uniform central bank collateral haircut, but then also introduce a model variant with continuous asset liquidity and haircuts.