Approximate Bayesian techniques for inference in stochastic dynamical systems

This thesis is concerned with approximate inference in dynamical systems, from a variational Bayesian perspective. When modelling real world dynamical systems, stochastic differential equations appear as a natural choice, mainly because of their ability to model the noise of the system by adding a variant of some stochastic process to the deterministic dynamics. Hence, inference in such processes has drawn much attention. Here two new extended frameworks are derived and presented that are based on basis function expansions and local polynomial approximations of a recently proposed variational Bayesian algorithm. It is shown that the new extensions converge to the original variational algorithm and can be used for state estimation (smoothing). However, the main focus is on estimating the (hyper-) parameters of these systems (i.e. drift parameters and diffusion coefficients). The new methods are numerically validated on a range of different systems which vary in dimensionality and non-linearity. These are the Ornstein-Uhlenbeck process, for which the exact likelihood can be computed analytically, the univariate and highly non-linear, stochastic double well and the multivariate chaotic stochastic Lorenz '63 (3-dimensional model). The algorithms are also applied to the 40 dimensional stochastic Lorenz '96 system. In this investigation these new approaches are compared with a variety of other well known methods such as the ensemble Kalman filter / smoother, a hybrid Monte Carlo sampler, the dual unscented Kalman filter (for jointly estimating the systems states and model parameters) and full weak-constraint 4D-Var. Empirical analysis of their asymptotic behaviour as a function of observation density or length of time window increases is provided.

Prevalence of subclinical thyroid dysfunction and its relation to socioeconomic deprivation in the elderly:a community-based cross-sectional survey

Context: Population-based screening has been advocated for subclinical thyroid dysfunction in the elderly because the disorder is perceived to be common, and health benefits may be accrued by detection and treatment. Objective: The objective of the study was to determine the prevalence of subclinical thyroid dysfunction and unidentified overt thyroid dysfunction in an elderly population. Design, Setting, and Participants: A cross-sectional survey of a community sample of participants aged 65 yr and older registered with 20 family practices in the United Kingdom. Exclusions: Exclusions included current therapy for thyroid disease, thyroid surgery, or treatment within 12 months. Outcome Measure: Tests of thyroid function (TSH concentration and free T 4 concentration in all, with measurement of free T3 in those with low TSH) were conducted. Explanatory Variables: These included all current medical diagnoses and drug therapies, age, gender, and socioeconomic deprivation (Index of Multiple Deprivation, 2004) Analysis: Standardized prevalence rates were analyzed. Logistic regression modeling was used to determine factors associated with the presence of subclinical thyroid dysfunction Results: A total of 5960 attended for screening. Using biochemical definitions, 94.2% [95% confidence interval (CI) 93.8-94.6%] were euthyroid. Unidentified overt hyper- and hypothyroidism were uncommon (0.3, 0.4%, respectively). Subclinical hyperthyroidism and hypothyroidism were identified with similar frequency (2.1%, 95% CI 1.8-2.3%; 2.9%, 95% CI 2.6-3.1%, respectively). Subclinical thyroid dysfunction was more common in females (P < 0.001) and with increasing age (P < 0.001). After allowing for comorbidities, concurrent drug therapies, age, and gender, an association between subclinical hyperthyroidism and a composite measure of socioeconomic deprivation remained. Conclusions: Undiagnosed overt thyroid dysfunction is uncommon. The prevalence of subclinical thyroid dysfunction is 5%. We have, for the first time, identified an independent association between the prevalence of subclinical thyroid dysfunction and deprivation that cannot be explained solely by the greater burden of chronic disease and/or consequent drug therapies in the deprived population. Copyright © 2006 by The Endocrine Society.

The challenge of real-time automatic mapping for environmental monitoring network management

The automatic interpolation of environmental monitoring network data such as air quality or radiation levels in real-time setting poses a number of practical and theoretical questions. Among the problems found are (i) dealing and communicating uncertainty of predictions, (ii) automatic (hyper)parameter estimation, (iii) monitoring network heterogeneity, (iv) dealing with outlying extremes, and (v) quality control. In this paper we discuss these issues, in light of the spatial interpolation comparison exercise held in 2004.

Sporicidal activity of two disinfectants against Clostridium difficile spores

The sporicidal activity of an odour-free peracetic acid-based disinfectant (Wofasteril®) and a widely-used dichloroisocyanurate preparation (Chlor-clean®) was assessed against spores of the hyper-virulent strain of Clostridium difficile (ribotype 027), in the presence and absence of organic matter. In environmentally clean conditions, dichloroisocyanurate achieved a >3 log10 reduction in 3 minutes, but a minimum contact time of 9 minutes was required to reduce the viable spore load to below detection levels. Peracetic acid achieved a >3 log10 reduction in 30 minutes and was overall significantly less effective (P<0.05). However, in the presence of organic matter - which reflects the true clinical environment - there was no significant difference between the sporicidal activity of dichloroisocyanurate and peracetic acid over a 60-minute period (P=0.188). Given the greater occupational health hazards generally associated with chlorine-releasing agents, odour-free peracetic acid-based disinfectants may offer a suitable alternative for environmental disinfection.

Dissolution rate enhancement, in vitro evaluation and investigation of drug release kinetics of chloramphenicol and sulphamethoxazole solid dispersions

Formulation of solid dispersions is one of the effective methods to increase the rate of solubilization and dissolution of poorly soluble drugs. Solid dispersions of chloramphenicol (CP) and sulphamethoxazole (SX) as model drugs were prepared by melt fusion method using polyethylene glycol 8000 (PEG 8000) as an inert carrier. The dissolution rate of CP and SX were rapid from solid dispersions with low drug and high polymer content. Characterization was performed using fourier transform infrared spectroscopy (FTIR), differential scanning calorimetry (DSC) and scanning electron microscopy (SEM). FTIR analysis for the solid dispersions of CP and SX showed that there was no interaction between PEG 8000 and the drugs. Hyper-DSC studies revealed that CP and SX were converted into an amorphous form when formulated as solid dispersion in PEG 8000. Mathematical analysis of the release kinetics demonstrated that drug release from the various formulations followed different mechanisms. Permeability studies demonstrated that both CP and SX when formulated as solid dispersions showed enhanced permeability across Caco-2 cells and CP can be classified as well-absorbed compound when formulated as solid dispersions. © 2013 Informa Healthcare USA, Inc.

Estimating parameters in stochastic systems:a variational Bayesian approach

This work is concerned with approximate inference in dynamical systems, from a variational Bayesian perspective. When modelling real world dynamical systems, stochastic differential equations appear as a natural choice, mainly because of their ability to model the noise of the system by adding a variation of some stochastic process to the deterministic dynamics. Hence, inference in such processes has drawn much attention. Here a new extended framework is derived that is based on a local polynomial approximation of a recently proposed variational Bayesian algorithm. The paper begins by showing that the new extension of this variational algorithm can be used for state estimation (smoothing) and converges to the original algorithm. However, the main focus is on estimating the (hyper-) parameters of these systems (i.e. drift parameters and diffusion coefficients). The new approach is validated on a range of different systems which vary in dimensionality and non-linearity. These are the Ornstein–Uhlenbeck process, the exact likelihood of which can be computed analytically, the univariate and highly non-linear, stochastic double well and the multivariate chaotic stochastic Lorenz ’63 (3D model). As a special case the algorithm is also applied to the 40 dimensional stochastic Lorenz ’96 system. In our investigation we compare this new approach with a variety of other well known methods, such as the hybrid Monte Carlo, dual unscented Kalman filter, full weak-constraint 4D-Var algorithm and analyse empirically their asymptotic behaviour as a function of observation density or length of time window increases. In particular we show that we are able to estimate parameters in both the drift (deterministic) and the diffusion (stochastic) part of the model evolution equations using our new methods.

A new variational radial basis function approximation for inference in multivariate diffusions

In this paper we present a radial basis function based extension to a recently proposed variational algorithm for approximate inference for diffusion processes. Inference, for state and in particular (hyper-) parameters, in diffusion processes is a challenging and crucial task. We show that the new radial basis function approximation based algorithm converges to the original algorithm and has beneficial characteristics when estimating (hyper-)parameters. We validate our new approach on a nonlinear double well potential dynamical system.

The influence of tool geometry on the performance of drilling tools

The main objective of the work presented in this thesis is to investigate the two sides of the flute, the face and the heel of a twist drill. The flute face was designed to yield straight diametral lips which could be extended to eliminate the chisel edge, and consequently a single cutting edge will be obtained. Since drill rigidity and space for chip conveyance have to be a compromise a theoretical expression is deduced which enables optimum chip disposal capacity to be described in terms of drill parameters. This expression is used to describe the flute heel side. Another main objective is to study the effect on drill performance of changing the conventional drill flute. Drills were manufactured according to the new flute design. Tests were run in order to compare the performance of a conventional flute drill and non conventional design put forward. The results showed that 50% reduction in thrust force and approximately 18% reduction in torque were attained for the new design. The flank wear was measured at the outer corner and found to be less for the new design drill than for the conventional one in the majority of cases. Hole quality, roundness, size and roughness were also considered as a further aspect of drill performance. Improvement in hole quality is shown to arise under certain cutting conditions. Accordingly it might be possible to use a hole which is produced in one pass of the new drill which previously would have required a drilled and reamed hole. A subsidiary objective is to design the form milling cutter that should be employed for milling the foregoing special flute from drill blank allowing for the interference effect. A mathematical analysis in conjunction with computing technique and computers is used. To control the grinding parameter, a prototype drill grinder was designed and built upon the framework of an existing cincinnati cutter grinder. The design and build of the new grinder is based on a computer aided drill point geometry analysis. In addition to the conical grinding concept, the new grinder is also used to produce spherical point utilizing a computer aided drill point geometry analysis.

The influence of cutting tool geometry upon aspects of chip flow and tool wear:a theoretical, three dimensional examination of the cutting geometry and the shape of the twist drill

This work is undertaken in the attempt to understand the processes at work at the cutting edge of the twist drill. Extensive drill life testing performed by the University has reinforced a survey of previously published information. This work demonstrated that there are two specific aspects of drilling which have not previously been explained comprehensively. The first concerns the interrelating of process data between differing drilling situations, There is no method currently available which allows the cutting geometry of drilling to be defined numerically so that such comparisons, where made, are purely subjective. Section one examines this problem by taking as an example a 4.5mm drill suitable for use with aluminium. This drill is examined using a prototype solid modelling program to explore how the required numerical information may be generated. The second aspect is the analysis of drill stiffness. What aspects of drill stiffness provide the very great difference in performance between short flute length, medium flute length and long flute length drills? These differences exist between drills of identical point geometry and the practical superiority of short drills has been known to shop floor drilling operatives since drilling was first introduced. This problem has been dismissed repeatedly as over complicated but section two provides a first approximation and shows that at least for smaller drills of 4. 5mm the effects are highly significant. Once the cutting action of the twist drill is defined geometrically there is a huge body of machinability data that becomes applicable to the drilling process. Work remains to interpret the very high inclination angles of the drill cutting process in terms of cutting forces and tool wear but aspects of drill design may already be looked at in new ways with the prospect of a more analytical approach rather than the present mix of experience and trial and error. Other problems are specific to the twist drill, such as the behaviour of the chips in the flute. It is now possible to predict the initial direction of chip flow leaving the drill cutting edge. For the future the parameters of further chip behaviour may also be explored within this geometric model.

Mechanisms contributing to the enhanced respiratory burst of neutrophils observed in periodontitis

The aim of this thesis is to investigate possible mechanisms that may contribute to neutrophil hyperactivity and hyper-reactivity. One possibility is the presence of a neutrophil priming factors within the peripheral circulation of periodontitis patients. To examine this possibility differentiated HL-60 cells and primary neutrophils were studied in the presence and absence of plasma from periodontitis patients. In independent experiments, plasma was depleted of IL-8, GM-CSF, interferon-a, immunoglobulins and albumin. This work demonstrated that plasma factors such as IL-8, GM-CSF, and interferon-a present during periodontitis may contribute towards the reported hyperactive neutrophil phenotype. Furthermore, this work demonstrated that products from Pg may regulate neutrophil accumulation at infected periodontal sites by promoting gingipain-dependent modification of IL-8-77 into a more biologically active chemokine. To elucidate whether the oxidatively stressed environment that neutrophils are exposed to in periodontitis could influence hyperactivity and hyper-reactivity, neutrophils were depleted of glutathione. This work showed that during oxidative stress, where cellular redox-levels have been altered, neutrophils exhibit an increased respiratory burst. In conclusion, this work highlights the multiple mechanisms that may contribute to neutrophil hyperactivity and hyperreactivity including gingipain-modulated activity of IL-8 variants, the effect of host factors such as IL-8, GM-CSF, interferon-a on neutrophils priming and activation, and the shift of neutrophil GSH:GSSG ratio in favour of a more oxidised environment as observed in periodontitis.

Activation of the neutrophil respiratory burst by plasma from periodontitis patients is mediated by pro-inflammatory cytokines

Aim: To determine the effect of periodontitis patients' plasma on the neutrophil oxidative burst and the role of albumin, immunoglobulins (Igs) and cytokines. Materials and Methods: Plasma was collected from chronic periodontitis patients (n=11) and periodontally healthy controls (n=11) and used with/without depletion of albumin and Ig or antibody neutralization of IL-8, GM-CSF or IFN-a to prime/stimulate peripheral blood neutrophils, isolated from healthy volunteers. The respiratory burst was measured by lucigenin-dependent chemiluminescence. Plasma cytokine levels were determined by ELISA. Results: Plasmas from patients were significantly more effective in both directly stimulating neutrophil superoxide production and priming for subsequent formyl-met-leu-phe (fMLP)-stimulated superoxide production than plasmas from healthy controls (p<0.05). This difference was maintained after depletion of albumin and Ig. Plasma from patients contained higher mean levels of IL-8, GM-CSF and IFN-a. Individual neutralizing antibodies against IL-8, GM-CSF or IFN-a inhibited the direct stimulatory effect of patients' plasma, whereas the ability to prime for fMLP-stimulated superoxide production was only inhibited by neutralization of IFN-a. The stimulating and priming effects of control plasma were unaffected by antibody neutralization. Conclusions: This study demonstrates that plasma cytokines may have a role in inducing the hyperactive (IL-8, GM-CSF, IFN-a) and hyper-reactive (IFN-a) neutrophil phenotype seen in periodontitis patients.

A solar powered liquid-desiccant cooling system for greenhouses

Liquid desiccant systems are of potential interest as a means of cooling greenhouses to temperatures below those achieved by conventional means. However, only very little work has been done on this technology with previous workers focussing on the cooling of human dwellings using expensive desiccants such as lithium salts. In this study we are designing a system for greenhouse cooling based on magnesium chloride desiccant which is an abundant and non-toxic substance. Magnesium chloride is found in seawater, for example, and is a by-product from solar salt works. We have carried out a detailed experimental study of the relevant properties of magnesium rich solutions. In addition we have constructed a test rig that includes the main components of the cooling system, namely a dehumidifier and solar regenerator. The dehumidifier is a cross-flow device that consists of a structured packing made of corrugated cellulose paper sheets with different flute angles and embedded cooling tubes. The regenerator is of the open type with insulated backing and fabric covering to spread the flow of desiccant solution. Alongside these experiments we are developing a mathematical model in gPROMS® that combines and simulates the heat and mass transfer processes in these components. The model can be applied to various geographical locations. Here we report predictions for Havana (Cuba) and Manila (Philippines), where we find that average wet-bulb temperatures can be lowered by 2.2 and 3°C, respectively, during the month of May.

Sulforaphane restores cellular glutathione levels and reduces chronic periodontitis neutrophil hyperactivity in vitro

The production of high levels of reactive oxygen species by neutrophils is associated with the local and systemic destructive phenotype found in the chronic inflammatory disease periodontitis. In the present study, we investigated the ability of sulforaphane (SFN) to restore cellular glutathione levels and reduce the hyperactivity of circulating neutrophils associated with chronic periodontitis. Using differentiated HL60 cells as a neutrophil model, here we show that generation of extracellular O2 . - by the nicotinamide adenine dinucleotide (NADPH) oxidase complex is increased by intracellular glutathione depletion. This may be attributed to the upregulation of thiol regulated acid sphingomyelinase driven lipid raft formation. Intracellular glutathione was also lower in primary neutrophils from periodontitis patients and, consistent with our previous findings, patients neutrophils were hyper-reactive to stimuli. The activity of nuclear factor erythroid-2-related factor 2 (Nrf2), a master regulator of the antioxidant response, is impaired in circulating neutrophils from chronic periodontitis patients. Although patients' neutrophils exhibit a low reduced glutathione (GSH)/oxidised glutathione (GSSG) ratio and a higher total Nrf2 level, the DNA-binding activity of nuclear Nrf2 remained unchanged relative to healthy controls and had reduced expression of glutamate cysteine ligase catalytic (GCLC), and modifier (GCLM) subunit mRNAs, compared to periodontally healthy subjects neutrophils. Pre-treatment with SFN increased expression of GCLC and GCM, improved intracellular GSH/GSSG ratios and reduced agonist-activated extracellular O2 . - production in both dHL60 and primary neutrophils from patients with periodontitis and controls. These findings suggest that a deficiency in Nrf2-dependent pathways may underpin susceptibility to hyper-reactivity in circulating primary neutrophils during chronic periodontitis. © 2013 Dias et al.

On the interpretation of synchronization in EEG hyperscanning studies:a cautionary note

EEG Hyperscanning is a method for studying two or more individuals simultaneously with the objective of elucidating how co-variations in their neural activity (i.e., hyperconnectivity) are influenced by their behavioral and social interactions. The aim of this study was to compare the performance of different hyper-connectivity measures using (i) simulated data, where the degree of coupling could be systematically manipulated, and (ii) individually recorded human EEG combined into pseudo-pairs of participants where no hyper-connections could exist. With simulated data we found that each of the most widely used measures of hyperconnectivity were biased and detected hyper-connections where none existed. With pseudo-pairs of human data we found spurious hyper-connections that arose because there were genuine similarities between the EEG recorded from different people independently but under the same experimental conditions. Specifically, there were systematic differences between experimental conditions in terms of the rhythmicity of the EEG that were common across participants. As any imbalance between experimental conditions in terms of stimulus presentation or movement may affect the rhythmicity of the EEG, this problem could apply in many hyperscanning contexts. Furthermore, as these spurious hyper-connections reflected real similarities between the EEGs, they were not Type-1 errors that could be overcome by some appropriate statistical control. However, some measures that have not previously been used in hyperconnectivity studies, notably the circular correlation co-efficient (CCorr), were less susceptible to detecting spurious hyper-connections of this type. The reason for this advantage in performance is discussed and the use of the CCorr as an alternative measure of hyperconnectivity is advocated. © 2013 Burgess.

Antenatal corticosteroids impact the inflammatory rather than the antiangiogenic profile of women with preeclampsia

Circulating antiangiogenic factors and proinflammatory cytokines are implicated in the pathogenesis of preeclampsia. This study was performed to test the hypothesis that steroids modify the balance of inflammatory and proangiogenic and antiangiogenic factors that potentially contribute to the patient’s evolving clinical state. Seventy singleton women, admitted for antenatal corticosteroid treatment, were enrolled prospectively. The study group consisted of 45 hypertensive women: chronic hypertension (n=6), severe preeclampsia (n=32), and superimposed preeclampsia (n=7). Normotensive women with shortened cervix (<2.5 cm) served as controls (n=25). Maternal blood samples of preeclampsia cases were obtained before steroids and then serially up until delivery. A clinical severity score was designed to clinically monitor disease progression. Serum levels of angiogenic factors (soluble fms-like tyrosine kinase-1 [sFlt-1], placental growth factor [PlGF], soluble endoglin [sEng]), endothelin-1 (ET-1), and proinflammatory markers (IL-6, C-reactive protein [CRP]) were assessed before and after steroids. Soluble IL-2 receptor (sIL-2R) and total immunoglobulins (IgG) were measured as markers of T- and B-cell activation, respectively. Steroid treatment coincided with a transient improvement in clinical manifestations of preeclampsia. A significant decrease in IL-6 and CRP was observed although levels of sIL-2R and IgG remained unchanged. Antenatal corticosteroids did not influence the levels of angiogenic factors but ET-1 levels registered a short-lived increase poststeroids. Although a reduction in specific inflammatory mediators in response to antenatal steroids may account for the transient improvement in clinical signs of preeclampsia, inflammation is unlikely to be the major contributor to severe preeclampsia or useful for therapeutic targeting.