No / Synthetic Biology aspires to design, compose and engineer biological systems that implement specified behaviour. When designing such systems, hypothesis testing via computational modelling and simulation is vital in order to reduce the need of costly wet lab experiments. As a case study, we discuss the use of computational modelling and stochastic simulation for engineered genetic circuits that implement Boolean AND and OR gates that have been reported in the literature. We present performance analysis results for nine different state-of-the-art stochastic simulation algorithms and analyse the dynamic behaviour of the proposed gates. Stochastic simulations verify the desired functioning of the proposed gate designs.

No / The ability to predict how far a drug will penetrate into the tumour microenvironment within its pharmacokinetic (PK) lifespan would provide valuable information about therapeutic response. As the PK profile is directly related to the route and schedule of drug administration, an in silico tool that can predict the drug administration schedule that results in optimal drug delivery to tumours would streamline clinical trial design. This paper investigates the application of mathematical and computational modelling techniques to help improve our understanding of the fundamental mechanisms underlying drug delivery, and compares the performance of a simple model with more complex approaches. Three models of drug transport are developed, all based on the same drug binding model and parametrized by bespoke in vitro experiments. Their predictions, compared for a ‘tumour cord’ geometry, are qualitatively and quantitatively similar. We assess the effect of varying the PK profile of the supplied drug, and the binding affinity of the drug to tumour cells, on the concentration of drug reaching cells and the accumulated exposure of cells to drug at arbitrary distances from a supplying blood vessel. This is a contribution towards developing a useful drug transport modelling tool for informing strategies for the treatment of tumour cells which are ‘pharmacokinetically resistant’ to chemotherapeutic strategies.

Yes / The tumour vasculature and microenvironment is complex and heterogeneous, contributing to reduced delivery of cancer drugs to the tumour. We have developed an in silico model of drug transport in a tumour cord to explore the effect of different drug regimes over a 72 h period and how changes in pharmacokinetic parameters affect tumour exposure to the cytotoxic drug doxorubicin. We used the model to describe the radial and axial distribution of drug in the tumour cord as a function of changes in the transport rate across the cell membrane, blood vessel and intercellular permeability, flow rate, and the binding and unbinding ratio of drug within the cancer cells. We explored how changes in these parameters may affect cellular exposure to drug. The model demonstrates the extent to which distance from the supplying vessel influences drug levels and the effect of dosing schedule in relation to saturation of drug-binding sites. It also shows the likely impact on drug distribution of the aberrant vasculature seen within tumours. The model can be adapted for other drugs and extended to include other parameters. The analysis confirms that computational models can play a role in understanding novel cancer therapies to optimize drug administration and delivery.

Online services such as e-marketplaces, social networking sites, online gaming environments etc have grown in popularity in the recent years. These services represent situation where participants do not get to negotiate face to face before interaction and most of the time parties to transaction remain anonymous. It is thus necessary to have a system that rightly assesses trustworthiness of the other party in order to maintain quality assurance in such systems. Recent works on Trust and Reputation in online communities have focused on identifying probable defaulters, but less effort has been put to come up with system that make cooperation attractive over defection in order to achieve cooperation without enforcement. Our work in this regard concerns design and investigation of trust assessment systems that not only filter defaulters but also promote evolution of cooperativeness in player society. Based on the concept of game theory and prisoner's dilemma, we model business games and design incentive method, compensation method, acquaintance based assessment method and decision theoretic assessment method as mechanisms to assure trustworthiness in online business environments. Effectiveness of each of these methods in promoting the evolution of cooperation in player society has been investigated. Our results show that these methods contribute positively in promoting cooperative evolution. We have further extended our trust assessment model to suit the needs of a mobile ad-hoc network setting. The effectiveness of this model has been tested against its capability to reduce packet drop rate and energy conservation. In both of these the results show promise.

Dans le cerveau, il est difficile de mettre au point des expériences avec un niveau de contrôle approprié à cause du haut niveau de connectivité. Pour traiter ce problème, les modèles mathématiques sont utilisés pour représenter le cerveau d’une façon plus compréhensible. En effet, les modèles mathématiques peuvent être plus pratiques que les expériences pour tester des hypothèses et chercher à extraire l’essence même du principe étudié, en le simplifiant. De plus, la modélisation computationnelle forme une branche spécifique de la modélisation mathématique, permettant de résoudre de gros calculs numériques. Dans cette thèse, j’ai utilisé la modélisation computationnelle à travers différentes approches pour étudier certaines régions cérébrales. Nous avons collaboré avec des neurobiologistes en appliquant nos modèles à des données expérimentales pour contribuer à mieux comprendre l’action de la dopamine, un neuromodulateur. J’ai étudié la diversité de l’action de la dopamine à trois échelles: la région cérébrale, le niveau cellulaire et le niveau moléculaire. La dopamine a un gros impact sur le cerveau et elle est principalement connue pour son implication dans le système de récompense. En effet, c’est une molécule associée à la prédiction de récompense et de punition. Peu de régions produisent de la dopamine et ces régions sont altérées par la maladie de Parkinson ou perturbées par la dépression. Pour la maladie de Parkinson, j’ai conçu un modèle de type taux de décharge pour reproduire l’activité neuronale des ganglions de la base. Ce modèle montre des réponses neuronales significativement différentes, entre la condition témoin et la condition parkinsonienne. Par ailleurs, avec un modèle de type Hodgin-Huxley prenant en compte la dynamique de l’ion potassium, j’ai pu appuyer l’hypothèse que la région cérébrale appelée l’habenula, lorsqu’elle est hyperactive, induirait la dépression. Cette dépression serait due à un déséquilibre de la concentration en potassium à cause d’une dysfonction de l’astrocyte (surexpression des canaux Kir4.1). Enfin, la dopamine est aussi impliquée dans la plasticité synaptique, un phénomène à la base de la mémoire. Je l’ai étudié avec un troisième modèle, prenant en compte plusieurs résultats expérimentaux relatifs à la plasticité en fonction du timing des potentiels d’action et de sa modulation. / In the brain, the high connectivity level makes it difficult to set up experiments with an appropriate level of control. To address that issue, mathematical models are used to represent the brain in a more comprehensive way. Easier than experiments to test hypotheses, mathematical models can extend them closer to reality and aim to extract the studied principle essence, by simplifying it. Computational modelling is a specific branch of mathematical modelling allowing to solve large numerical calculations. In this thesis, I used computational modelling to study brain parts through different approaches, all in collaboration with neurobiologists and applied to experimental data. A common framework is given by the goal of contributing to a picture of the action of the neuromodulator dopamine. I studied the diversity of dopamine's action at three different scales: the brain region, the cellular level and the molecular level. Dopamine has a large impact on the brain and it is mainly known for its rewarding dimension, it is, indeed, the molecule associated with reward prediction and punishment. Few regions in the brain produce dopamine and these regions are impaired in Parkinson's disease or disrupted in major depressive disorders. Concerning Parkinson's disease, I designed a firing-rate model to fit experimental basal ganglia neural activity, which disclosed significant changes of the neural response between control and Parkinsonian condition. Furthermore, with a Hodgkin-Huxley model accounting for the dynamics of the potassium ion, I could support the hypothesis that the brain region called lateral habenula hyper-activates and induces major depressive disorders because of unbalanced potassium concentration due to astrocyte dysfunction (Kir4.1 channels overexpression). Dopamine is also involved in synaptic plasticity, a phenomenon at the basis of memory that I explored with a third model accounting for several experimental results pertaining to spike-timing-dependent plasticity and its modulation.

Collagen scaffolds are porous structures which are used in bioreactors and in a wide range of tissue engineering applications. In these contexts, the scaffolds may be subjected to conditions in which fluid is forced through the structure and the scaffold is simultaneously compressed. It is clear that fluid transport within collagen scaffolds, and the inter-relationships between permeability, scaffold structure, fluid pressure and scaffold deformation are of key importance. However, these relationships remain poorly understood. In this thesis, a series of isotropic collagen structures were produced using a freeze-drying technique from aqueous slurry concentrations 0.5, 0.75 and 1 wt%, and fully characterised using X-ray micro-tomography and compression testing. It was found that collagen wt% influenced structural parameters such as pore size, porosity, relative density and mechanical properties. Percolation theory was used to investigate the pore interconnectivity of each scaffold. Structures with lower collagen fraction resulted in larger percolation diameters, but lower mechanical stiffness. Aligned collagen scaffolds were also produced by altering the freeze-drying protocol and using different types of mould materials and designs. It was found that a polycarbonate mould with stainless base resulted in vertically aligned structures with low angular variation. When compared with isotropic scaffolds from slurry of the same concentration, aligned scaffolds had a larger percolation diameter. Tortuosity was used as a mathematical tool to characterise the interconnected pathways within each porous structure. The effect of the size of the region of interest (ROI) chosen and the size of the virtual probe particle used in the analysis on the values of tortuosity calculated were determined and an optimised calculation methodology developed. Increasing the collagen fraction within isotropic scaffolds increased the tortuosity, and aligned structures had smaller tortuosity values than their isotropic counterparts. Permeability studies were conducted using two complementary experimental rigs designed to cover a range of pressure regimes and the results were compared with predictions from mathematical models and computational simulations. At low pressures, it was found that the lower collagen fraction structures, which had more open morphologies, had higher permeabilities. Alignment of the structure also enhanced permeability. The scaffolds all experienced deformation at high pressures resulting in a restriction of fluid flow. The lower collagen fraction scaffolds experienced a sharper decrease in permeability with increased pressure and aligned structures were more responsive to deformation than their isotropic counterparts. The inter-relationships between permeability, scaffold structure, fluid pressure and deformation of collagen scaffolds were explored. For isotropic samples, permeability followed a broad $(1- \epsilon)^2$ behaviour with strain as predicted by a tetrakaidecahedral structural model, with the constant of proportionality changing with collagen fraction. In contrast, the aligned structures did not follow this behaviour with the permeability dropping much more sharply in the early stages of compression. Open-cell polyurethane (PU) foams, sometimes used as dressings in wound healing applications, are often compared with collagen scaffolds in permeability models and were used in this thesis as a comparison structure. The foam had a higher permeability than the scaffolds due to its larger pore sizes and higher interconnectivity. In the light of the effects of compression on permeability, the changes in porous structure with compression were explored in isotropic and aligned 0.75 wt% scaffolds. Unlike the fluid flow experiments, these experiments were carried out in the dry state. Deformation in simple linear compression and in step-wise compression was studied, and the stress relaxation behaviour of the scaffolds characterised. A methodology was developed to characterise the structural changes accompanying compression using X-ray micro-tomography with an in situ compression stage. The methodology accounted for the need for samples to remain unchanged during the scan collection period for stable image reconstruction. The scaffolds were studied in uniaxial compression and biaxial compression and it was found that pore size and percolation diameter decreased with increasing compressive strain, while the tortuosity increased. The aligned structure was less affected than the isotropic at low compressions, in contrast to the results from the permeability study in which the aligned structure was more responsive to strain. This suggests that the degree of hydration may affect the structural changes observed. The insights gained in this study of the inter-relationships between microstructure, fluid dynamics and deformation in collagen scaffolds are of relevance to the informed design of porous structures for medical applications.

Visual processing is the earliest core process required to support a normal reading system. However, little attention has been given to its role in any of the existing cognitive/computational models of reading. The ultimate goal of this thesis is to create a large-scale model of reading, which can generate phonology and semantics from print. Building such a model will allow for the exploration of a number of theoretically important cognitive phenomena in both normal and impaired reading including: font and size invariance; letter confusability; length effects; and pure alexic reading patterns. To achieve this goal, there are a number of important sub-goals that need to be achieved: (1) to develop a visual processing component which is capable of recognising letters in different fonts and sizes; (2) to produce a model that can develop useful intermediate (orthographic) representations as a consequence of learning; (3) to develop a set of semantic representations compact enough to allow efficient learning but that can still capture realistic semantic similarity relationships; (4) to integrate all the components together into a large-scale recurrent reading model; and (5) to extend the model to support picture naming, and to explore whether damage to the visual system can produce symptoms similar to those found in PA patients. Chapter 2 started by developing a simple feedforward network for letter recognition. The model was trained with letters in various transformations, which allowed the model to learn to deal with size and shape invariance problems as well as accounting for letter confusability effects and generalising to previously unseen letters. The model achieved this by extracting key features from visual input which could be used to support accurate letter recognition. Chapter 3 incorporated the letter recognition component developed in Chapter 2 into a word reading model. The reading model was trained on the mappings between print and phonology, with the orthographic representations which learn to emerge over training. The model could support accurate nonword naming and simulated the length by lexicality interaction observed in normal reading. A system of semantic representations was developed in Chapter 4 by using co-occurrence statistics to generate semantic codes that preserved realistic similarity relationships. Chapter 5 integrated all the components developed in the previous chapters together into a large-scale recurrent reading model. Finally, Chapter 6 extended the reading model to perform object recognition along with the reading task. When the model's visual system was damaged it was able to simulate the abnormal length effect typically seen in PA patients. The damaged model also showed impaired reaction times in object naming and preserved sensitivity to lexical/semantic variables in reading. The picture naming performance was modulated by visual complexity. In summary, the results highlight the importance of incorporating visual information into computational models of single word reading, and suggest that doing so will enable the exploration of a wide range of effects that were previously inaccessible to these types of connectionist models.

This dissertation concerns the computational modelling of early life development of music perception and cognition. Experimental psychology and neuroscience show results that suggest that the development of musical representations in infancy, whether concerning pitch or rhythm features, depend on exposure both to music and language. Early musical and linguistic skills seem to be, therefore, tangled in ways we are yet to characterize. In parallel, computational modelling has produced powerful frameworks for the study of learning and development. The use of these models for studying the development of music information perception and cognition, connecting music and language still remains to be explored. This way, we propose to produce computational solutions suitable for studying factors that contribute to shape our cognitive structure, building our predispositions that allow us to enjoy and make sense of music. We will also adopt a comparative approach to the study of early development of musical predispositions that involves both music and language, searching for possible interactions and correlations. We first address pitch representation (absolute vs relative) and its relations with development. Simulations have allowed us to observe a parallel between learning and the type of pitch information being used, where the type of encoding influenced the ability of the model to perform a discrimination task correctly. Next, we have performed a prosodic characterization of infant-directed speech and singing by comparing rhythmic and melodic patterning in two Portuguese (European and Brazilian) variants. In the computational experiments, rhythm related descriptors exhibited a strong predictive ability for both speech and singing language variants' discrimination tasks, presenting different rhythmic patterning for each variant. This reveals that the prosody of the surrounding sonic environment of an infant is a source of rich information and rhythm as a key element for characterizing the prosody from language and songs from each culture. Finally, we built a computational model based on temporal information processing and representation for exploring how the temporal prosodic patterns of a specific culture influence the development of rhythmic representations and predispositions. The simulations show that exposure to the surrounding sound environment influences the development of temporal representations and that the structure of the exposure environment, specifically the lack of maternal songs, has an impact on how the model organizes its internal representations. We conclude that there is a reciprocal influence between music and language. The exposure to the structure of the sonic background influences the shaping of our cognitive structure, which supports our understanding of musical experience. Among the sonic background, language's structure has a predominant role in biasing the building of musical predispositions and representations. / Esta tesis aborda la modelización computacional de algunos fenómenos de la percepción y cognición de la música durante el período de desarrollo en la primera infancia. La Psicología experimental y la Neurociencia muestran resultados que sugieren que el desarrollo de las representaciones del ritmo o de la altura musicales durante la infancia son dependientes de la exposición tanto a la música como al lenguaje de las culturas en las que se nace y crece. La capacidad musical y lingüística, durante los primeros años de desarrollo, están inter-relacionadas de formas que aún no ha sido posible caracterizar. En paralelo, las herramientas computacionales proporcionan un marco teórico y empírico eficaz para el estudio del aprendizaje y el desarrollo. El uso de los modelos computacionales para estudiar el desarrollo de la percepción y la cognición de información musical, conectando la música y el lenguaje, todavía queda por explorar. Así, nos proponemos producir soluciones computacionales adecuadas para el estudio de los factores que contribuyen a dar forma a nuestra estructura cognitiva y a la construcción de las predisposiciones que nos permiten disfrutar y dar sentido a la música. También adoptamos una perspectiva comparativa para la investigación que, englobando la música y el lenguaje, busca sus posibles interacciones y correlaciones. Primeramente, hemos abordado la representación de la altura tonal (absoluta vs. relativa) y sus relaciones con el desarrollo. Las simulaciones computacionales han permitido observar que el tipo de codificación utilizada ha influido en la capacidad del modelo para efectuar correctamente una tarea de discriminación, lo cual sugiere una relación entre el aprendizaje y el tipo de información de altura que se utiliza. Seguidamente, se ha realizado una caracterización prosódica del habla y del canto dirigidos al bebé, mediante la comparación de patrones rítmicos y melódicos en dos variantes de Portugués (Europeo y Brasileño). En los experimentos computacionales, los descriptores relacionados con el ritmo han exhibido una fuerte capacidad predictiva para el habla y canto, en tareas de discriminación de variante de lenguaje, siendo observados diferentes patrones rítmicos para cada variante. Se revela que la prosodia del entorno sonoro de un bebé es una fuente rica de información y que el ritmo es un elemento fundamental para la caracterización de la prosodia del lenguaje y las canciones de una cultura. Por último, se construyó un modelo computacional basado en el procesamiento y representación de información temporal para explorar cómo los patrones prosódicos temporales del habla de una cultura específica influyen en el desarrollo de las representaciones y predisposiciones rítmicas. Las simulaciones muestran que la exposición al ambiente sonoro circundante influye en el desarrollo de las representaciones temporales y que la estructura del entorno a que se esta expuesto, específicamente, la falta de canciones maternales, tiene un impacto sobre la forma como el modelo organiza sus representaciones rítmicas internas. Se concluye que existe una influencia recíproca entre la música y el lenguaje. La exposición a la estructura del entorno sonoro influye en la formación de la estructura cognitiva, que sustenta la comprensión de la experiencia musical. De entre todos los “inputs” del entorno sonoro, la estructura del lenguaje tiene una influencia predominante en la construcción de predisposiciones y representaciones musicales.

We develop in this thesis a computational model of music expectation, which may be one of the most important aspects in music listening. Many phenomenons related to music listening such as preference, surprise or emo- tions are linked to the anticipatory behaviour of listeners. In this thesis, we concentrate on a statistical account to music expectation, by modelling the processes of learning and predicting spectro-temporal regularities in a causal fashion. The principle of statistical modelling of expectation can be applied to several music representations, from symbolic notation to audio signals. We first show that computational learning architectures can be used and evaluated to account behavioral data concerning auditory perception and learning. We then propose a what/when representation of musical events which enables to sequentially describe and learn the structure of acoustic units in musical audio signals. The proposed representation is applied to describe and anticipate timbre features and musical rhythms. We suggest ways to exploit the properties of the expectation model in music analysis tasks such as structural segmentation. We finally explore the implications of our model for interactive music applications in the context of real-time transcription, concatenative synthesis, and visualization. / Esta tesis presenta un modelo computacional de expectativa musical, que es un aspecto muy importante de como procesamos la música que oímos. Muchos fenómenos relacionados con el procesamiento de la música están vinculados a una capacidad para anticipar la continuación de una pieza de música. Nos enfocaremos en un acercamiento estadístico de la expectativa musical, modelando los procesos de aprendizaje y de predicción de las regularidades espectro-temporales de forma causal. El principio de modelado estadístico de la expectativa se puede aplicar a varias representaciones de estructuras musicales, desde las notaciones simbólicas a la señales de audio. Primero demostramos que ciertos algoritmos de aprendizaje de secuencias se pueden usar y evaluar en el contexto de la percepción y el aprendizaje de secuencias auditivas. Luego, proponemos una representación, denominada qué/cuándo, para representar eventos musicales de una forma que permite describir y aprender la estructura secuencial de unidades acústicas en señales de audio musical. Aplicamos esta representación para describir y anticipar características tímbricas y ritmos. Sugerimos que se pueden explotar las propiedades del modelo de expectativa para resolver tareas de análisis como la segmentación estructural de piezas musicales. Finalmente, exploramos las implicaciones de nuestro modelo a la hora de definir nuevas aplicaciones en el contexto de la transcripción en tiempo real, la síntesis concatenativa y la visualización.

Computational modelling has been gaining an increasing amount of support from the neuroscience community as a tool to assay cognition and computational processes in the brain. Lately, scientists have started to apply computational methods from neuroscience to the study of psychiatry to gain further insight into the mechanisms leading to mental disorders. In fact, only recently has psychiatry started to move away from categorising illnesses using behavioural symptoms in an attempt for a more biologically driven diagnosis. To date, several neurobiological anomalies have been found in schizophrenia and led to a multitude of conceptual framework attempting to link the biology to the patients’ symptoms. Computational modelling can be applied to formalise these conceptual frameworks in an effort to test the validity or likelihood of each hypothesis. Recently, a novel conceptual model has been proposed to describe how positive symptoms (delusions, hallucinations and thought disorder) and cognitive symptoms (poor decision-making, i.e. “executive functioning”) might arise in schizophrenia. This framework however, has not been tested experimentally or against computational models. The focus of this thesis was to use a combination of behavioural experiments and computational models to independently assess the validity of each component that make up this framework. The first study of this thesis focused on the computational analysis of a disrupted prediction-error signalling and its implications for decision-making performances in complex tasks. Briefly, we used a reinforcement-learning model of a gambling task in rodents and disrupted the prediction-error signal known to be critical for learning. We found that this disruption can account for poor performances in decision-making due to an incorrect acquisition of the model of the world. This study illustrates how disruptions in prediction-error signalling (known to be present in schizophrenia) can lead to the acquisition of an incorrect world model which can lead to poor executive functioning or false beliefs (delusions) as seen in patients. The second study presented in this thesis addressed spatial working memory performances in chronic schizophrenia, bipolar disorder, first episode psychosis and family relatives of DISC1 translocation carriers. We build a probabilistic inference model to solve the working memory task optimally and then implemented various alterations of this model to test commonly debated hypotheses of cognitive deficiency in schizophrenia. Our goal was to find which of these hypotheses accounts best for the poor performance observed in patients. We found that while the performance at the task was significantly different for most patients groups in comparison to controls, this effect disappeared after controlling for IQ in one group. The models were nonetheless fitted to the experimental data and suggest that working memory maintenance is most likely to account for the poor performances observed in patients. We propose that the maintenance of information in working memory might have indirect implications for measures of general cognitive performance, as these rely on a correct filtering of information against distractions and cortical noise. Finally the third study presented in this thesis assessed the performance of medicated chronic schizophrenia patients in a statistical learning task of visual stimuli and measured how the acquired statistics influenced their perception. We find that patient with chronic schizophrenia appear to be unimpaired at statistical learning of visual stimuli. The acquired statistics however appear to induce less expectation-driven ‘hallucinations’ of the stimuli in the patients group than in controls. We find that this is in line with previous literature showing that patients are less susceptible to expectation-driven illusions than controls. This study highlights however the idea that perceptual processes during sensory integration diverge from this of healthy controls. In conclusion, this thesis suggests that impairments in reinforcement learning and Bayesian inference appear to be able to account for the positive and cognitive symptoms observed in schizophrenia, but that further work is required to merge these findings. Specifically, while our studies addressed individual components such as associative learning, working memory, implicit learning & perceptual inference, we cannot conclude that deficits of reinforcement learning and Bayesian inference can collectively account for symptoms in schizophrenia. We argue however that the studies presented in this thesis provided evidence that impairments of reinforcement learning and Bayesian inference are compatible with the emergence of positive and cognitive symptoms in schizophrenia.