Thesis (M.Sc. (Physics))--North-West University, Potchefstroom Campus, 2008.

Thesis (M.Sc. (Physics))--North-West University, Potchefstroom Campus, 2008.

Movement is central for life, and all animals depend on accurate regulation of movement for purposeful behavior. There is great diversity of movements, ranging between simple and vital breathing movements to minute and subtle movements of the face used to communicate emotions. Consequently, motor neurons, which are the only route of central nervous system output, are essential for all motor behaviors. To control the many motor behaviors expressed by an animal, motor neurons are exposed to a large number and variety of modulating synaptic inputs and have evolved into subtypes with specific functions. In this thesis, motor neuron subtypes and the synaptic input to motor neurons from Renshaw cells and Ia afferents have been studied. Novel molecular markers that identify subtypes of motor neurons are described. Three markers, Chodl, Calca and ERRβ, have been used to study the degeneration of subtypes of motor neurons in a mouse model of the neurodegenerative disease amyotrophic lateral sclerosis (ALS). Another marker, 5-ht1d, has been used to record the electrophysiological character of gamma motor neurons. In mice that lack 5-ht1d, motor neurons develop with reduced proprioceptive input. Remarkably, these mice had fewer foot faults than control animals when challenged to cross a narrow beam suggesting that the amplitude of monosynaptic proprioceptive input to motor neurons is not essential for motor coordination. In a final set of experiments, genetic removal of vesicular transport of neurotransmitter from Renshaw cells suggest that Renshaw cells are not integral for motor circuit function or motor behaviors. However, they are involved in the development of motor circuits in the spinal cord. Together, this thesis provides novel molecular tools for studies of motor neuron subtypes and novel data regarding the development and function of spinal motor circuits.

<p>n this dissertation a comparison of two evolutionary methods for evolving ANNs for robot control is made. The methods compared are SANE with enforced sub-population and delta-coding, and marker-based encoding. In an attempt to speed up evolution, marker-based encoding is extended with delta-coding. The task selected for comparison is the hunter-prey task. This task requires the robot controller to posess some form of memory as the prey can move out of sensor range. Incremental evolution is used to evolve the complex behaviour that is required to successfully handle this task. The comparison is based on computational power needed for evolution, and complexity, robustness, and generalisation of the resulting ANNs. The results show that marker-based encoding is the most efficient method tested and does not need delta-coding to increase the speed of evolution process. Additionally the results indicate that delta-coding does not increase the speed of evolution with marker-based encoding.</p>

We present a theoretical approach to the statistical analysis of the dependence of the gap time length between consecutive recurrent events, on a set of explanatory random variables and in the presence of right censoring. The dependence is expressed through regression-like and overdispersion parameters, estimated via estimating functions and equations. The mean and variance of the length of each gap time, conditioned on the observed history of prior events and other covariates, are known functions of parameters and covariates, and are part of the estimating functions. Under certain conditions on censoring, we construct normalized estimating functions that are asymptotically unbiased and contain only observed data. We then use modern mathematical techniques to prove the existence, consistency and asymptotic normality of a sequence of estimators of the parameters. Simulations support our theoretical results.

Made available in DSpace on 2015-07-22T20:49:31Z (GMT). No. of bitstreams: 0 Previous issue date: 2010-10-27 / Objetivo: Comparar a sobrevida de pacientes com recidiva de câncer de ovário que realizaram estudo com 18F FDG PET-CT não dedicado, tomografia computadorizada e marcador CA-125 com as enfermas que se submeteram ao tratamento e seguimento usual com tomografia computadorizada e o marcador CA-125. Metodologia: Estudaram-se pacientes com neoplasia ovariana epitelial maligna no Hospital do Câncer AC Camargo- São Paulo, no período compreendido entre janeiro de 1996 à dezembro de 2003, as quais apresentaram recidiva, comprovada pela cirurgia. Avaliaram-se 44 pacientes (grupo A) que efetuaram exame 18F-FDG em PET-CT não dedicado e 31 doentes (grupo B) que efetuaram o seguimento habitual. Excluíram-se as pacientes em vigência de quimioterapia (menos de 48 horas) da data da realização do exame 18F-FDG em PET-CT não dedicado. Resultados: Pacientes com exames de PET-CT não dedicado (grupo A) apresentaram significante aumento da sobrevida em comparação com o grupo B (p=0.0200). Outrossim, 11 pacientes com estudo anátomopatológico “borderline” apresentaram apenas um exame falso negativo (implantes peritoneais), 10 verdadeiro negativos e 18 verdadeiro positivos. Conclusão: O emprego do FDG-PET-CT não dedicado na avaliação de pacientes com neoplasia ovariana maligna mostrou ser procedimento que resultou em benefícios, interferindo na atitude dos cirurgiões, mormente na recidiva subclínica. Verificou-se maior sobrevida das enfermas após o tratamento da recidiva quando foi identificado pelo FDG-PET-CT não dedicado seja por sítios ou locais insuspeitos ou mesmo pela completa ausência de sintomas e sinais clínicos. / Purpose The aim of this study was to compare the impact of the survival in patients with ovarian cancer recurrence undergoing non-dedicated [18F]FDG PET-CT, CT and serum CA-125 (group A) and patients that were submitted to conventional follow-up with CT and serum CA-125. Methods Patients diagnosed with epithelial ovarian cancer at the AC Camargo Cancer Hospital from January 1996 to December 2003 who presented recurrence (checked by surgery), and performed [18F]FDG PET-CT on a total of 44 patients- group A, and 31 patients -group B undergoing conventional follow up. Patients undergoing chemotherapy less than 48 hours before non-dedicated PET-CT were excluded from the research. Results Patients of Group A presented an increase on their survival rate of 5.94 in comparison with group B (p=0.0200). The 11 borderline patients in group A performed 29 [18F]FDG PET-CT studies with 1 false negative (peritoneal implants), 10 true negatives and 18 true positives. Conclusion Oncologists are used to test several markers in order to perform diagnosis, staging and prognosis, but their real benefit in the treatment is still uncertain. The use of [18F]FDG PET-CT in the evaluation of subclinical ovarian cancer recurrence has been shown to be accurate and presented a rewarding cost-benefit relationship, interfering in the surgeons conduct facing a subclinical recurrence. / TEDE / BV UNIFESP: Teses e dissertações

BASE TEÓRICA: O Nevo recorrente (NR) é uma lesão melanocítica que surge na cicatriz de tratamentos cirúrgicos de um nevo melanocítico benigno prévio, apresentando uma alta prevalência na prática dermatológica diária. A recorrência ocorre principalmente após exéreses por shaving, método cirúrgico muito usado na prática clínica devido ao seu bom resultado estético e simplicidade técnica. Estas lesões podem ser um desafio para o dermatologista devido ao fato de algumas vezes simularem tanto clínica quanto histologicamente o melanoma. OBJETIVOS: Avaliar as características dermatoscópicas da lesão primária, potencialmente relacionadas com a sua recorrência após exérese por shaving. Avaliar também as características clínicas, histológicas e a dermatoscopia do reverso da peça cirúrgica preditoras de recorrência. MÉTODOS: Foi realizada a avaliação clínica, histológica e dermatoscópica de 224 nevos melanocíticos benignos de 61 pacientes antes da sua remoção por shaving. A remoção foi realizada de maneira padronizada e o reverso da peça cirúrgica avaliado também dermatoscopicamente após o procedimento. Após seis meses, os pacientes foram reavaliados quanto à presença de recorrência local. RESULTADOS: Recorrência de nevo foi observada em 59 (30,3%) lesões. No modelo de regressão univariável a presença de nevo recorrente mostrou associação inversa com a idade, sendo que em cada ano de aumento na mesma, há uma redução de 3% na prevalência das recorrências (p<0,001). Houve também associação com o fototipo, sendo os mais altos mais prevalentes (p=0,043). O tipo histológico da lesão primária mais associado com a recorrência foi o composto (p=0,005). Quanto às características dermatoscópicas, a presença de pigmento dermatoscópico no reverso da lesão (p<0,001), a presença de coloração marrom-escura (p=0,025), presença de pontos (p=0,007), presença de pêlo terminal (p=0,040) e presença de hiperpigmentação anular (p<0,001) se correlacionaram com os nevos recorrentes. A presença de um maior número de cores à dermatoscopia, também aumenta a prevalência da recorrência (p=0,002). Por outro lado, a presença de vasos (p=0,008) e presença de vasos lineares (p=0,014) se correlacionaram com a ausência de recorrência. Na análise multivariável a presença de hiperpigmentação anular [RP=3,659 (IC: 2,544 – 5,264) p<0,001] e a idade [PR=0,971 (CI: 0,958 – 0,984) p<0,001] destacam-se como os preditores mais significativos para recorrência. CONCLUSÃO Em cada ano de aumento da idade, a prevalência dos NRs é reduzida em 3%. Descrevemos um novo padrão de pigmentação, onde o pigmento é visto distribuído de forma anular, muitas vezes circundando as estruturas anexiais, que se relaciona a uma maior prevalência de recorrência. As lesões que apresentam este padrão, têm uma prevalência de recorrência 3,66 vezes maior, corrigida para a idade. / BACKGROUND: Recurrent Nevus (RN) is a melanocytic lesion that arises in a scar after removal of a previous benign melanocytic nevus. Shaving excisions, that are commonly used due to its optimal cosmetic outcomes and technical simplicity, are mostly related with recurrent nevi. Recurrences might be a challenge for the dermatologist due its clinical and histological aspects that might simulate melanomas. OBJECTIVES: This study aimed to evaluate the dermoscopic features observed in primary lesion of recurrent nevi after shaving excision. Additionally, we analyzed clinical and histological features and surgical specimen reverse of these lesions. METHODS: Clinical, histological and dermoscopic evaluation of 224 benign melanocytic nevi from 61 patients, prior to their removal by shaving, was performed. The removal was performed by standard method. Additionally we performed ex-vivo dermoscopy on the reverse of the surgical specimen after the procedure. After six months, patients were reassessed for the presence of local recurrence. RESULTS: Nevus recurrence was observed in 59 (30.3%) lesions. In the univariate regression model the presence of recurrent nevus was inversely associated with age, and in each year of increase in the same, there is a 3% reduction in the prevalence of recurrence (p <0.001). Was also found an association with skin type. Higher skin types are more associated with recurrent nevus (p = 0.043). Regarding histologyc type of the primary lesion, the compound nevus presents de higher prevalence of recurrence (p=0.005). Regarding dermoscopic characteristics, the presence of pigment in the reverse of surgical specimen (p <0.001), the presence of dark brown color (p = 0.025), presence of dots (p = 0.007), the presence of terminal hair (p = 0.040) and the presence of annular hyperpigmentation (p <0.001) correlated with recurrent nevi. The presence of a larger number of colors in dermatoscopy also increases the prevalence of recurrence (p = 0.002). On the other hand, the presence of blood vessels (p = 0.008) and the presence of linear vessels (p = 0.014) correlated with the absence of recurrence. In multivariable analysis, the presence of annular hyperpigmentation [PR=3.659 (CI: 2.544 – 5.264) p<0.001] and age [PR=0.971 (CI: 0.958 – 0.984) p<0.001] remained significant. CONCLUSION: The prevalence of RN is reduced by 3% after each year of increasing age. We describe annular hyperpigmentation, especially around adnexial structures as a novel dermoscopic structure as a predictor of nevi recurrency. The lesions with annular hyperpigmentation have a 3.66 times higher prevalence of RN.

Text classification has always been a concern in area of natural language processing, especially nowadays the data are getting massive due to the development of internet. Recurrent neural network (RNN) is one of the most popular method for natural language processing due to its recurrent architecture which give it ability to process serialized information. In the meanwhile, Convolutional neural network (CNN) has shown its ability to extract features from visual imagery. This paper combine the advantages of RNN and CNN and proposed a model called BLSTM-C for Chinese text classification. BLSTM-C begins with a Bidirectional long short-term memory (BLSTM) layer which is an special kind of RNN to get a sequence output based on the past context and the future context. Then it feed this sequence to CNN layer which is utilized to extract features from the previous sequence. We evaluate BLSTM-C model on several tasks such as sentiment classification and category classification and the result shows our model’s remarkable performance on these text tasks.

Although nonlinear model predictive control (NMPC) might be the best choice for a nonlinear plant, it is still not widely used. This is mainly due to the computational burden associated with solving online a set of nonlinear differential equations and a nonlinear dynamic optimization problem in real time. This thesis is concerned with strategies aimed at reducing the computational burden involved in different stages of the NMPC such as optimization problem, state estimation, and nonlinear model identification. A major part of the computational burden comes from function and derivative evaluations required in different parts of the NMPC algorithm. In this work, the problem is tackled using a recently introduced efficient tool, the automatic differentiation (AD). Using the AD tool, a function is evaluated together with all its partial derivative from the code defining the function with machine accuracy. A new NMPC algorithm based on nonlinear least square optimization is proposed. In a first–order method, the sensitivity equations are integrated using a linear formula while the AD tool is applied to get their values accurately. For higher order approximations, more terms of the Taylor expansion are used in the integration for which the AD is effectively used. As a result, the gradient of the cost function against control moves is accurately obtained so that the online nonlinear optimization can be efficiently solved. In many real control cases, the states are not measured and have to be estimated for each instance when a solution of the model equations is needed. A nonlinear extended version of the Kalman filter (EKF) is added to the NMPC algorithm for this purpose. The AD tool is used to calculate the required derivatives in the local linearization step of the filter automatically and accurately. Offset is another problem faced in NMPC. A new nonlinear integration is devised for this case to eliminate the offset from the output response. In this method, an integrated disturbance model is added to the process model input or output to correct the plant/model mismatch. The time response of the controller is also improved as a by–product. The proposed NMPC algorithm has been applied to an evaporation process and a two continuous stirred tank reactor (two–CSTR) process with satisfactory results to cope with large setpoint changes, unmeasured severe disturbances, and process/model mismatches. When the process equations are not known (black–box) or when these are too complicated to be used in the controller, modelling is needed to create an internal model for the controller. In this thesis, a continuous time recurrent neural network (CTRNN) in a state–space form is developed to be used in NMPC context. An efficient training algorithm for the proposed network is developed using AD tool. By automatically generating Taylor coefficients, the algorithm not only solves the differentiation equations of the network but also produces the sensitivity for the training problem. The same approach is also used to solve online the optimization problem of the NMPC. The proposed CTRNN and the predictive controller were tested on an evaporator and two–CSTR case studies. A comparison with other approaches shows that the new algorithm can considerably reduce network training time and improve solution accuracy. For a third case study, the ALSTOM gasifier, a NMPC via linearization algorithm is implemented to control the system. In this work a nonlinear state–space class Wiener model is used to identify the black–box model of the gasifier. A linear model of the plant at zero–load is adopted as a base model for prediction. Then, a feedforward neural network is created as the static gain for a particular output channel, fuel gas pressure, to compensate its strong nonlinear behavior observed in open–loop simulations. By linearizing the neural network at each sampling time, the static nonlinear gain provides certain adaptation to the linear base model. The AD tool is used here to linearize the neural network efficiently. Noticeable performance improvement is observed when compared with pure linear MPC. The controller was able to pass all tests specified in the benchmark problem at all load conditions.

In this thesis, we explore the interface between symbolic and dynamical system computation, with particular regard to dynamical system models of neuronal networks. In doing so, we adhere to a definition of computation as the physical realization of a formal system, where we say that a dynamical system performs a computation if a correspondence can be found between its dynamics on a vectorial space and the formal system’s dynamics on a symbolic space. Guided by this definition, we characterize computation in a range of neuronal network models. We first present a constructive mapping between a range of formal systems and Recurrent Neural Networks (RNNs), through the introduction of a Versatile Shift and a modular network architecture supporting its real-time simulation. We then move on to more detailed models of neural dynamics, characterizing the computation performed by networks of delay-pulse-coupled oscillators supporting the emergence of heteroclinic dynamics. We show that a correspondence can be found between these networks and Finite-State Transducers, and use the derived abstraction to investigate how noise affects computation in this class of systems, unveiling a surprising facilitatory effect on information transmission. Finally, we present a new dynamical framework for computation in neuronal networks based on the slow-fast dynamics paradigm, and discuss the consequences of our results for future work, specifically for what concerns the fields of interactive computation and Artificial Intelligence.