Game Theory Models for Multi-Robot Patrolling of Infraestructures

Abstract This work is focused on the problem of performing multi‐robot patrolling for infrastructure security applications in order to protect a known environment at critical facilities. Thus, given a set of robots and a set of points of interest, the patrolling task consists of constantly visiting these points at irregular time intervals for security purposes. Current existing solutions for these types of applications are predictable and inflexible. Moreover, most of the previous centralized and deterministic solutions and only few efforts have been made to integrate dynamic methods. Therefore, the development of new dynamic and decentralized collaborative approaches in order to solve the aforementioned problem by implementing learning models from Game Theory. The model selected in this work that includes belief‐based and reinforcement models as special cases is called Experience‐Weighted Attraction. The problem has been defined using concepts of Graph Theory to represent the environment in order to work with such Game Theory techniques. Finally, the proposed methods have been evaluated experimentally by using a patrolling simulator. The results obtained have been compared with previous available

Cooperative Multi-robot Patrolling: A study of distributed approaches based on mathematical models of game theory to protect infrastructures

El principio de Teoría de Juegos permite desarrollar modelos estocásticos de patrullaje multi-robot para proteger infraestructuras criticas. La protección de infraestructuras criticas representa un gran reto para los países al rededor del mundo, principalmente después de los ataques terroristas llevados a cabo la década pasada. En este documento el termino infraestructura hace referencia a aeropuertos, plantas nucleares u otros instalaciones. El problema de patrullaje se define como la actividad de patrullar un entorno determinado para monitorear cualquier actividad o sensar algunas variables ambientales. En esta actividad, un grupo de robots debe visitar un conjunto de puntos de interés definidos en un entorno en intervalos de tiempo irregulares con propósitos de seguridad. Los modelos de partullaje multi-robot son utilizados para resolver este problema. Hasta el momento existen trabajos que resuelven este problema utilizando diversos principios matemáticos. Los modelos de patrullaje multi-robot desarrollados en esos trabajos representan un gran avance en este campo de investigación. Sin embargo, los modelos con los mejores resultados no son viables para aplicaciones de seguridad debido a su naturaleza centralizada y determinista. Esta tesis presenta cinco modelos de patrullaje multi-robot distribuidos e impredecibles basados en modelos matemáticos de aprendizaje de Teoría de Juegos. El objetivo del desarrollo de estos modelos está en resolver los inconvenientes presentes en trabajos preliminares. Con esta finalidad, el problema de patrullaje multi-robot se formuló utilizando conceptos de Teoría de Grafos, en la cual se definieron varios juegos en cada vértice de un grafo. Los modelos de patrullaje multi-robot desarrollados en este trabajo de investigación se han validado y comparado con los mejores modelos disponibles en la literatura. Para llevar a cabo tanto la validación como la comparación se ha utilizado un simulador de patrullaje y un grupo de robots reales. Los resultados experimentales muestran que los modelos de patrullaje desarrollados en este trabajo de investigación trabajan mejor que modelos de trabajos previos en el 80% de 150 casos de estudio. Además de esto, estos modelos cuentan con varias características importantes tales como distribución, robustez, escalabilidad y dinamismo. Los avances logrados con este trabajo de investigación dan evidencia del potencial de Teoría de Juegos para desarrollar modelos de patrullaje útiles para proteger infraestructuras. ABSTRACT Game theory principle allows to developing stochastic multi-robot patrolling models to protect critical infrastructures. Critical infrastructures protection is a great concern for countries around the world, mainly due to terrorist attacks in the last decade. In this document, the term infrastructures includes airports, nuclear power plants, and many other facilities. The patrolling problem is defined as the activity of traversing a given environment to monitoring any activity or sensing some environmental variables If this activity were performed by a fleet of robots, they would have to visit some places of interest of an environment at irregular intervals of time for security purposes. This problem is solved using multi-robot patrolling models. To date, literature works have been solved this problem applying various mathematical principles.The multi-robot patrolling models developed in those works represent great advances in this field. However, the models that obtain the best results are unfeasible for security applications due to their centralized and predictable nature. This thesis presents five distributed and unpredictable multi-robot patrolling models based on mathematical learning models derived from Game Theory. These multi-robot patrolling models aim at overcoming the disadvantages of previous work. To this end, the multi-robot patrolling problem was formulated using concepts of Graph Theory to represent the environment. Several normal-form games were defined at each vertex of a graph in this formulation. The multi-robot patrolling models developed in this research work have been validated and compared with best ranked multi-robot patrolling models in the literature. Both validation and comparison were preformed by using both a patrolling simulator and real robots. Experimental results show that the multirobot patrolling models developed in this research work improve previous ones in as many as 80% of 150 cases of study. Moreover, these multi-robot patrolling models rely on several features to highlight in security applications such as distribution, robustness, scalability, and dynamism. The achievements obtained in this research work validate the potential of Game Theory to develop patrolling models to protect infrastructures.

Forest connectivity restoration through reforestation: an integrated methodology for prioritizing agricultural lands and selecting reforestation species

Forest connectivity restoration is a major goal in natural resource planning. Given the high amount of abandoned cultivated lands, setting efficient methods for the reforestation of agricultural lands offers a good opportunity to face this issue. However, reforestations must be carefully planned, which poses two main challenges. In first place, to determine those agricultural lands that, once reforested, would meet more effectively the planning goals. As a further step, in order to grant the success of the activity, it is fairly advisable to select those tree species that are more adapted to each particular environment. Here we intend to give response to both requirements by proposing a sequential and integrated methodology that has been implemented in two Spanish forest districts, which are formed by several landscape types that were previously defined and characterized. Using the software Conefor Sensinode, a powerful tool for quantifying habitat availability that is based on graph theory concepts, we determined the landscapes where forest planning should have connectivity as a major concern and, afterwards, we detected the agricultural patches that would contribute most to enhance connectivity if they were reforested. The subsequent reforestation species assessment was performed within these priority patches. Using penalized logistic regressions we fitted ecological niche models for the Spanish native tree species. The models were trained with species distribution data from the Spanish Forest Map and used climatic and lithological variables as predictors. Model predictions were used to build ordered lists of suitable species for each priority patch. The lists include dominant and non dominant tree species and allow adding biodiversity goals to the reforestation planning. The result of this combined methodology is a map of agricultural patches that would contribute most to uphold forest connectivity if they were reforested and a list of suitable tree species for each patch ordered by occurrence probability. Therefore the proposed methodology may be useful for suitable and efficient forest planning and landscape designing.

Detecting periodicity with horizontal visibility graphs

The horizontal visibility algorithm was recently introduced as a mapping between time series and networks. The challenge lies in characterizing the structure of time series (and the processes that generated those series) using the powerful tools of graph theory. Recent works have shown that the visibility graphs inherit several degrees of correlations from their associated series, and therefore such graph theoretical characterization is in principle possible. However, both the mathematical grounding of this promising theory and its applications are in its infancy. Following this line, here we address the question of detecting hidden periodicity in series polluted with a certain amount of noise. We first put forward some generic properties of horizontal visibility graphs which allow us to define a (graph theoretical) noise reduction filter. Accordingly, we evaluate its performance for the task of calculating the period of noisy periodic signals, and compare our results with standard time domain (autocorrelation) methods. Finally, potentials, limitations and applications are discussed.

Contributions to Bayesian network learning with applications to neuroscience

Neuronal morphology is a key feature in the study of brain circuits, as it is highly related to information processing and functional identification. Neuronal morphology affects the process of integration of inputs from other neurons and determines the neurons which receive the output of the neurons. Different parts of the neurons can operate semi-independently according to the spatial location of the synaptic connections. As a result, there is considerable interest in the analysis of the microanatomy of nervous cells since it constitutes an excellent tool for better understanding cortical function. However, the morphologies, molecular features and electrophysiological properties of neuronal cells are extremely variable. Except for some special cases, this variability makes it hard to find a set of features that unambiguously define a neuronal type. In addition, there are distinct types of neurons in particular regions of the brain. This morphological variability makes the analysis and modeling of neuronal morphology a challenge. Uncertainty is a key feature in many complex real-world problems. Probability theory provides a framework for modeling and reasoning with uncertainty. Probabilistic graphical models combine statistical theory and graph theory to provide a tool for managing domains with uncertainty. In particular, we focus on Bayesian networks, the most commonly used probabilistic graphical model. In this dissertation, we design new methods for learning Bayesian networks and apply them to the problem of modeling and analyzing morphological data from neurons. The morphology of a neuron can be quantified using a number of measurements, e.g., the length of the dendrites and the axon, the number of bifurcations, the direction of the dendrites and the axon, etc. These measurements can be modeled as discrete or continuous data. The continuous data can be linear (e.g., the length or the width of a dendrite) or directional (e.g., the direction of the axon). These data may follow complex probability distributions and may not fit any known parametric distribution. Modeling this kind of problems using hybrid Bayesian networks with discrete, linear and directional variables poses a number of challenges regarding learning from data, inference, etc. In this dissertation, we propose a method for modeling and simulating basal dendritic trees from pyramidal neurons using Bayesian networks to capture the interactions between the variables in the problem domain. A complete set of variables is measured from the dendrites, and a learning algorithm is applied to find the structure and estimate the parameters of the probability distributions included in the Bayesian networks. Then, a simulation algorithm is used to build the virtual dendrites by sampling values from the Bayesian networks, and a thorough evaluation is performed to show the model’s ability to generate realistic dendrites. In this first approach, the variables are discretized so that discrete Bayesian networks can be learned and simulated. Then, we address the problem of learning hybrid Bayesian networks with different kinds of variables. Mixtures of polynomials have been proposed as a way of representing probability densities in hybrid Bayesian networks. We present a method for learning mixtures of polynomials approximations of one-dimensional, multidimensional and conditional probability densities from data. The method is based on basis spline interpolation, where a density is approximated as a linear combination of basis splines. The proposed algorithms are evaluated using artificial datasets. We also use the proposed methods as a non-parametric density estimation technique in Bayesian network classifiers. Next, we address the problem of including directional data in Bayesian networks. These data have some special properties that rule out the use of classical statistics. Therefore, different distributions and statistics, such as the univariate von Mises and the multivariate von Mises–Fisher distributions, should be used to deal with this kind of information. In particular, we extend the naive Bayes classifier to the case where the conditional probability distributions of the predictive variables given the class follow either of these distributions. We consider the simple scenario, where only directional predictive variables are used, and the hybrid case, where discrete, Gaussian and directional distributions are mixed. The classifier decision functions and their decision surfaces are studied at length. Artificial examples are used to illustrate the behavior of the classifiers. The proposed classifiers are empirically evaluated over real datasets. We also study the problem of interneuron classification. An extensive group of experts is asked to classify a set of neurons according to their most prominent anatomical features. A web application is developed to retrieve the experts’ classifications. We compute agreement measures to analyze the consensus between the experts when classifying the neurons. Using Bayesian networks and clustering algorithms on the resulting data, we investigate the suitability of the anatomical terms and neuron types commonly used in the literature. Additionally, we apply supervised learning approaches to automatically classify interneurons using the values of their morphological measurements. Then, a methodology for building a model which captures the opinions of all the experts is presented. First, one Bayesian network is learned for each expert, and we propose an algorithm for clustering Bayesian networks corresponding to experts with similar behaviors. Then, a Bayesian network which represents the opinions of each group of experts is induced. Finally, a consensus Bayesian multinet which models the opinions of the whole group of experts is built. A thorough analysis of the consensus model identifies different behaviors between the experts when classifying the interneurons in the experiment. A set of characterizing morphological traits for the neuronal types can be defined by performing inference in the Bayesian multinet. These findings are used to validate the model and to gain some insights into neuron morphology. Finally, we study a classification problem where the true class label of the training instances is not known. Instead, a set of class labels is available for each instance. This is inspired by the neuron classification problem, where a group of experts is asked to individually provide a class label for each instance. We propose a novel approach for learning Bayesian networks using count vectors which represent the number of experts who selected each class label for each instance. These Bayesian networks are evaluated using artificial datasets from supervised learning problems. Resumen La morfología neuronal es una característica clave en el estudio de los circuitos cerebrales, ya que está altamente relacionada con el procesado de información y con los roles funcionales. La morfología neuronal afecta al proceso de integración de las señales de entrada y determina las neuronas que reciben las salidas de otras neuronas. Las diferentes partes de la neurona pueden operar de forma semi-independiente de acuerdo a la localización espacial de las conexiones sinápticas. Por tanto, existe un interés considerable en el análisis de la microanatomía de las células nerviosas, ya que constituye una excelente herramienta para comprender mejor el funcionamiento de la corteza cerebral. Sin embargo, las propiedades morfológicas, moleculares y electrofisiológicas de las células neuronales son extremadamente variables. Excepto en algunos casos especiales, esta variabilidad morfológica dificulta la definición de un conjunto de características que distingan claramente un tipo neuronal. Además, existen diferentes tipos de neuronas en regiones particulares del cerebro. La variabilidad neuronal hace que el análisis y el modelado de la morfología neuronal sean un importante reto científico. La incertidumbre es una propiedad clave en muchos problemas reales. La teoría de la probabilidad proporciona un marco para modelar y razonar bajo incertidumbre. Los modelos gráficos probabilísticos combinan la teoría estadística y la teoría de grafos con el objetivo de proporcionar una herramienta con la que trabajar bajo incertidumbre. En particular, nos centraremos en las redes bayesianas, el modelo más utilizado dentro de los modelos gráficos probabilísticos. En esta tesis hemos diseñado nuevos métodos para aprender redes bayesianas, inspirados por y aplicados al problema del modelado y análisis de datos morfológicos de neuronas. La morfología de una neurona puede ser cuantificada usando una serie de medidas, por ejemplo, la longitud de las dendritas y el axón, el número de bifurcaciones, la dirección de las dendritas y el axón, etc. Estas medidas pueden ser modeladas como datos continuos o discretos. A su vez, los datos continuos pueden ser lineales (por ejemplo, la longitud o la anchura de una dendrita) o direccionales (por ejemplo, la dirección del axón). Estos datos pueden llegar a seguir distribuciones de probabilidad muy complejas y pueden no ajustarse a ninguna distribución paramétrica conocida. El modelado de este tipo de problemas con redes bayesianas híbridas incluyendo variables discretas, lineales y direccionales presenta una serie de retos en relación al aprendizaje a partir de datos, la inferencia, etc. En esta tesis se propone un método para modelar y simular árboles dendríticos basales de neuronas piramidales usando redes bayesianas para capturar las interacciones entre las variables del problema. Para ello, se mide un amplio conjunto de variables de las dendritas y se aplica un algoritmo de aprendizaje con el que se aprende la estructura y se estiman los parámetros de las distribuciones de probabilidad que constituyen las redes bayesianas. Después, se usa un algoritmo de simulación para construir dendritas virtuales mediante el muestreo de valores de las redes bayesianas. Finalmente, se lleva a cabo una profunda evaluaci ón para verificar la capacidad del modelo a la hora de generar dendritas realistas. En esta primera aproximación, las variables fueron discretizadas para poder aprender y muestrear las redes bayesianas. A continuación, se aborda el problema del aprendizaje de redes bayesianas con diferentes tipos de variables. Las mixturas de polinomios constituyen un método para representar densidades de probabilidad en redes bayesianas híbridas. Presentamos un método para aprender aproximaciones de densidades unidimensionales, multidimensionales y condicionales a partir de datos utilizando mixturas de polinomios. El método se basa en interpolación con splines, que aproxima una densidad como una combinación lineal de splines. Los algoritmos propuestos se evalúan utilizando bases de datos artificiales. Además, las mixturas de polinomios son utilizadas como un método no paramétrico de estimación de densidades para clasificadores basados en redes bayesianas. Después, se estudia el problema de incluir información direccional en redes bayesianas. Este tipo de datos presenta una serie de características especiales que impiden el uso de las técnicas estadísticas clásicas. Por ello, para manejar este tipo de información se deben usar estadísticos y distribuciones de probabilidad específicos, como la distribución univariante von Mises y la distribución multivariante von Mises–Fisher. En concreto, en esta tesis extendemos el clasificador naive Bayes al caso en el que las distribuciones de probabilidad condicionada de las variables predictoras dada la clase siguen alguna de estas distribuciones. Se estudia el caso base, en el que sólo se utilizan variables direccionales, y el caso híbrido, en el que variables discretas, lineales y direccionales aparecen mezcladas. También se estudian los clasificadores desde un punto de vista teórico, derivando sus funciones de decisión y las superficies de decisión asociadas. El comportamiento de los clasificadores se ilustra utilizando bases de datos artificiales. Además, los clasificadores son evaluados empíricamente utilizando bases de datos reales. También se estudia el problema de la clasificación de interneuronas. Desarrollamos una aplicación web que permite a un grupo de expertos clasificar un conjunto de neuronas de acuerdo a sus características morfológicas más destacadas. Se utilizan medidas de concordancia para analizar el consenso entre los expertos a la hora de clasificar las neuronas. Se investiga la idoneidad de los términos anatómicos y de los tipos neuronales utilizados frecuentemente en la literatura a través del análisis de redes bayesianas y la aplicación de algoritmos de clustering. Además, se aplican técnicas de aprendizaje supervisado con el objetivo de clasificar de forma automática las interneuronas a partir de sus valores morfológicos. A continuación, se presenta una metodología para construir un modelo que captura las opiniones de todos los expertos. Primero, se genera una red bayesiana para cada experto y se propone un algoritmo para agrupar las redes bayesianas que se corresponden con expertos con comportamientos similares. Después, se induce una red bayesiana que modela la opinión de cada grupo de expertos. Por último, se construye una multired bayesiana que modela las opiniones del conjunto completo de expertos. El análisis del modelo consensuado permite identificar diferentes comportamientos entre los expertos a la hora de clasificar las neuronas. Además, permite extraer un conjunto de características morfológicas relevantes para cada uno de los tipos neuronales mediante inferencia con la multired bayesiana. Estos descubrimientos se utilizan para validar el modelo y constituyen información relevante acerca de la morfología neuronal. Por último, se estudia un problema de clasificación en el que la etiqueta de clase de los datos de entrenamiento es incierta. En cambio, disponemos de un conjunto de etiquetas para cada instancia. Este problema está inspirado en el problema de la clasificación de neuronas, en el que un grupo de expertos proporciona una etiqueta de clase para cada instancia de manera individual. Se propone un método para aprender redes bayesianas utilizando vectores de cuentas, que representan el número de expertos que seleccionan cada etiqueta de clase para cada instancia. Estas redes bayesianas se evalúan utilizando bases de datos artificiales de problemas de aprendizaje supervisado.

Sensitization profiles and cross-reactivity among plant allergens. Molecular mechanisms of food allergy development and the role of enhancers

La prevalencia de las alergias está aumentando desde mediados del siglo XX, y se estima que actualmente afectan a alrededor del 2-8 % de la población, pero las causas de este aumento aún no están claras. Encontrar el origen del mecanismo por el cual una proteína inofensiva se convierte en capaz de inducir una respuesta alérgica es de vital importancia para prevenir y tratar estas enfermedades. Aunque la caracterización de alérgenos relevantes ha ayudado a mejorar el manejo clínico y a aclarar los mecanismos básicos de las reacciones alérgicas, todavía queda un largo camino para establecer el origen de la alergenicidad y reactividad cruzada. El objetivo de esta tesis ha sido caracterizar las bases moleculares de la alergenicidad tomando como modelo dos familias de panalergenos (proteínas de transferencia de lípidos –LTPs- y taumatinas –TLPs-) y estudiando los mecanismos que median la sensibilización y la reactividad cruzada para mejorar tanto el diagnóstico como el tratamiento de la alergia. Para ello, se llevaron a cabo dos estrategias: estudiar la reactividad cruzada de miembros de familias de panalérgenos; y estudiar moléculas-co-adyuvantes que pudieran favorecer la capacidad alergénica de dichas proteínas. Para estudiar la reactividad cruzada entre miembros de la misma familia de proteínas, se seleccionaron LTPs y TLPs, descritas como alergenos, tomando como modelo la alergia a frutas. Por otra parte, se estudiaron los perfiles de sensibilización a alérgenos de trigo relacionados con el asma del panadero, la enfermedad ocupacional más relevante de origen alérgico. Estos estudios se llevaron a cabo estandarizando ensayos tipo microarrays con alérgenos y analizando los resultados por la teoría de grafos. En relación al estudiar moléculas-co-adyuvantes que pudieran favorecer la capacidad alergénica de dichas proteínas, se llevaron a cabo estudios sobre la interacción de los alérgenos alimentarios con células del sistema inmune humano y murino y el epitelio de las mucosas, analizando la importancia de moléculas co-transportadas con los alérgenos en el desarrollo de una respuesta Th2. Para ello, Pru p 3(LTP y alérgeno principal del melocotón) se selección como modelo para llevarlo a cabo. Por otra parte, se analizó el papel de moléculas activadoras del sistema inmune producidas por patógenos en la inducción de alergias alimentarias seleccionando el modelo kiwi-alternaria, y el papel de Alt a 1, alérgeno mayor de dicho hongo, en la sensibilización a Act d 2, alérgeno mayor de kiwi. En resumen, el presente trabajo presenta una investigación innovadora aportando resultados de gran utilidad tanto para la mejora del diagnóstico como para nuevas investigaciones sobre la alergia y el esclarecimiento final de los mecanismos que caracterizan esta enfermedad. ABSTRACT Allergies are increasing their prevalence from mid twentieth century, and they are currently estimated to affect around 2-8% of the population but the underlying causes of this increase remain still elusive. The understanding of the mechanism by which a harmless protein becomes capable of inducing an allergic response provides us the basis to prevent and treat these diseases. Although the characterization of relevant allergens has led to improved clinical management and has helped to clarify the basic mechanisms of allergic reactions, it seems justified in aspiring to molecularly dissecting these allergens to establish the structural basis of their allergenicity and cross-reactivity. The aim of this thesis was to characterize the molecular basis of the allergenicity of model proteins belonging to different families (Lipid Transfer Proteins –LTPs-, and Thaumatin-like Proteins –TLPs-) in order to identify mechanisms that mediate sensitization and cross reactivity for developing new strategies in the management of allergy, both diagnosis and treatment, in the near future. With this purpose, two strategies have been conducted: studies of cross-reactivity among panallergen families and molecular studies of the contribution of cofactors in the induction of the allergic response by these panallergens. Following the first strategy, we studied the cross-reactivity among members of two plant panallergens (LTPs , Lipid Transfer Proteins , and TLPs , Thaumatin-like Proteins) using the peach allergy as a model. Similarly, we characterized the sensitization profiles to wheat allergens in baker's asthma development, the most relevant occupational disease. These studies were performed using allergen microarrays and the graph theory for analyzing the results. Regarding the second approach, we analyzed the interaction of plant allergens with immune and epithelial cells. To perform these studies , we examined the importance of ligands and co-transported molecules of plant allergens in the development of Th2 responses. To this end, Pru p 3, nsLTP (non-specific Lipid Transfer Protein) and peach major allergen, was selected as a model to investigate its interaction with cells of the human and murine immune systems as well as with the intestinal epithelium and the contribution of its ligand in inducing an allergic response was studied. Moreover, we analyzed the role of pathogen associated molecules in the induction of food allergy. For that, we selected the kiwi- alternaria system as a model and the role of Alt a 1 , major allergen of the fungus, in the development of Act d 2-sensitization was studied. In summary, this work presents an innovative research providing useful results for improving diagnosis and leading to further research on allergy and the final clarification of the mechanisms that characterize this disease.

A framework for the analysis and optimization of delay in multiview video coding schemes

Esta tesis presenta un novedoso marco de referencia para el análisis y optimización del retardo de codificación y descodificación para vídeo multivista. El objetivo de este marco de referencia es proporcionar una metodología sistemática para el análisis del retardo en codificadores y descodificadores multivista y herramientas útiles en el diseño de codificadores/descodificadores para aplicaciones con requisitos de bajo retardo. El marco de referencia propuesto caracteriza primero los elementos que tienen influencia en el comportamiento del retardo: i) la estructura de predicción multivista, ii) el modelo hardware del codificador/descodificador y iii) los tiempos de proceso de cuadro. En segundo lugar, proporciona algoritmos para el cálculo del retardo de codificación/ descodificación de cualquier estructura arbitraria de predicción multivista. El núcleo de este marco de referencia consiste en una metodología para el análisis del retardo de codificación/descodificación multivista que es independiente de la arquitectura hardware del codificador/descodificador, completada con un conjunto de modelos que particularizan este análisis del retardo con las características de la arquitectura hardware del codificador/descodificador. Entre estos modelos, aquellos basados en teoría de grafos adquieren especial relevancia debido a su capacidad de desacoplar la influencia de los diferentes elementos en el comportamiento del retardo en el codificador/ descodificador, mediante una abstracción de su capacidad de proceso. Para revelar las posibles aplicaciones de este marco de referencia, esta tesis presenta algunos ejemplos de su utilización en problemas de diseño que afectan a codificadores y descodificadores multivista. Este escenario de aplicación cubre los siguientes casos: estrategias para el diseño de estructuras de predicción que tengan en consideración requisitos de retardo además del comportamiento tasa-distorsión; diseño del número de procesadores y análisis de los requisitos de velocidad de proceso en codificadores/ descodificadores multivista dado un retardo objetivo; y el análisis comparativo del comportamiento del retardo en codificadores multivista con diferentes capacidades de proceso e implementaciones hardware. ABSTRACT This thesis presents a novel framework for the analysis and optimization of the encoding and decoding delay for multiview video. The objective of this framework is to provide a systematic methodology for the analysis of the delay in multiview encoders and decoders and useful tools in the design of multiview encoders/decoders for applications with low delay requirements. The proposed framework characterizes firstly the elements that have an influence in the delay performance: i) the multiview prediction structure ii) the hardware model of the encoder/decoder and iii) frame processing times. Secondly, it provides algorithms for the computation of the encoding/decoding delay of any arbitrary multiview prediction structure. The core of this framework consists in a methodology for the analysis of the multiview encoding/decoding delay that is independent of the hardware architecture of the encoder/decoder, which is completed with a set of models that particularize this delay analysis with the characteristics of the hardware architecture of the encoder/decoder. Among these models, the ones based in graph theory acquire special relevance due to their capacity to detach the influence of the different elements in the delay performance of the encoder/decoder, by means of an abstraction of its processing capacity. To reveal possible applications of this framework, this thesis presents some examples of its utilization in design problems that affect multiview encoders and decoders. This application scenario covers the following cases: strategies for the design of prediction structures that take into consideration delay requirements in addition to the rate-distortion performance; design of number of processors and analysis of processor speed requirements in multiview encoders/decoders given a target delay; and comparative analysis of the encoding delay performance of multiview encoders with different processing capabilities and hardware implementations.

Horizontal visibility graphs generated by type-I intermittency

The type-I intermittency route to (or out of) chaos is investigated within the horizontal visibility (HV) graph theory. For that purpose, we address the trajectories generated by unimodal maps close to an inverse tangent bifurcation and construct their associatedHVgraphs.We showhowthe alternation of laminar episodes and chaotic bursts imprints a fingerprint in the resulting graph structure. Accordingly, we derive a phenomenological theory that predicts quantitative values for several network parameters. In particular, we predict that the characteristic power-law scaling of the mean length of laminar trend sizes is fully inherited by the variance of the graph degree distribution, in good agreement with the numerics. We also report numerical evidence on how the characteristic power-law scaling of the Lyapunov exponent as a function of the distance to the tangent bifurcation is inherited in the graph by an analogous scaling of block entropy functionals defined on the graph. Furthermore, we are able to recast the full set of HV graphs generated by intermittent dynamics into a renormalization-group framework, where the fixed points of its graph-theoretical renormalization-group flow account for the different types of dynamics.We also establish that the nontrivial fixed point of this flow coincides with the tangency condition and that the corresponding invariant graph exhibits extremal entropic properties.

White matter damage disorganizes brain functional networks in amnestic mild cognitive impairment

Although progressive functional brain network disruption has been one of the hallmarks of Alzheimer?s Dis- ease, little is known about the origin of this functional impairment that underlies cognitive symptoms. We in- vestigated how the loss of white matter (WM) integrity disrupts the organization of the functional networks at different frequency bands. The analyses were performed in a sample of healthy elders and mild cognitive im- pairment (MCI) subjects. Spontaneous brain magnetic activity (measured with magnetoencephalography) was characterized with phase synchronization analysis, and graph theory was applied to the functional networks. We identified WM areas (using diffusion weighted magnetic resonance imaging) that showed a statistical de- pendence between the fractional anisotropy and the graph metrics. These regions are part of an episodic mem- ory network and were also related to cognitive functions. Our data support the hypothesis that disruption of the anatomical networks influences the organization at the functional level resulting in the prodromal dementia syndrome of MCI.

Disparate connectivity for structural and functional networks is revealed when physical location of the connected nodes is considered

Macroscopic brain networks have been widely described with the manifold of metrics available using graph theory. However, most analyses do not incorporate information about the physical position of network nodes. Here, we provide a multimodal macroscopic network characterization while considering the physical positions of nodes. To do so, we examined anatomical and functional macroscopic brain networks in a sample of twenty healthy subjects. Anatomical networks are obtained with a graph based tractography algorithm from diffusion-weighted magnetic resonance images (DW-MRI). Anatomical con- nections identified via DW-MRI provided probabilistic constraints for determining the connectedness of 90 dif- ferent brain areas. Functional networks are derived from temporal linear correlations between blood-oxygenation level-dependent signals derived from the same brain areas. Rentian Scaling analysis, a technique adapted from very- large-scale integration circuits analyses, shows that func- tional networks are more random and less optimized than the anatomical networks. We also provide a new metric that allows quantifying the global connectivity arrange- ments for both structural and functional networks. While the functional networks show a higher contribution of inter-hemispheric connections, the anatomical networks highest connections are identified in a dorsal?ventral arrangement. These results indicate that anatomical and functional networks present different connectivity organi- zations that can only be identified when the physical locations of the nodes are included in the analysis.

Multimodal integration of functional and structural brain connectivity. An application to the characterization of Mild Cognitive Impairment

La investigación para el conocimiento del cerebro es una ciencia joven, su inicio se remonta a Santiago Ramón y Cajal en 1888. Desde esta fecha a nuestro tiempo la neurociencia ha avanzado mucho en el desarrollo de técnicas que permiten su estudio. Desde la neurociencia cognitiva hoy se explican muchos modelos que nos permiten acercar a nuestro entendimiento a capacidades cognitivas complejas. Aun así hablamos de una ciencia casi en pañales que tiene un lago recorrido por delante. Una de las claves del éxito en los estudios de la función cerebral ha sido convertirse en una disciplina que combina conocimientos de diversas áreas: de la física, de las matemáticas, de la estadística y de la psicología. Esta es la razón por la que a lo largo de este trabajo se entremezclan conceptos de diferentes campos con el objetivo de avanzar en el conocimiento de un tema tan complejo como el que nos ocupa: el entendimiento de la mente humana. Concretamente, esta tesis ha estado dirigida a la integración multimodal de la magnetoencefalografía (MEG) y la resonancia magnética ponderada en difusión (dMRI). Estas técnicas son sensibles, respectivamente, a los campos magnéticos emitidos por las corrientes neuronales, y a la microestructura de la materia blanca cerebral. A lo largo de este trabajo hemos visto que la combinación de estas técnicas permiten descubrir sinergias estructurofuncionales en el procesamiento de la información en el cerebro sano y en el curso de patologías neurológicas. Más específicamente en este trabajo se ha estudiado la relación entre la conectividad funcional y estructural y en cómo fusionarlas. Para ello, se ha cuantificado la conectividad funcional mediante el estudio de la sincronización de fase o la correlación de amplitudes entre series temporales, de esta forma se ha conseguido un índice que mide la similitud entre grupos neuronales o regiones cerebrales. Adicionalmente, la cuantificación de la conectividad estructural a partir de imágenes de resonancia magnética ponderadas en difusión, ha permitido hallar índices de la integridad de materia blanca o de la fuerza de las conexiones estructurales entre regiones. Estas medidas fueron combinadas en los capítulos 3, 4 y 5 de este trabajo siguiendo tres aproximaciones que iban desde el nivel más bajo al más alto de integración. Finalmente se utilizó la información fusionada de MEG y dMRI para la caracterización de grupos de sujetos con deterioro cognitivo leve, la detección de esta patología resulta relevante en la identificación precoz de la enfermedad de Alzheimer. Esta tesis está dividida en seis capítulos. En el capítulos 1 se establece un contexto para la introducción de la connectómica dentro de los campos de la neuroimagen y la neurociencia. Posteriormente en este capítulo se describen los objetivos de la tesis, y los objetivos específicos de cada una de las publicaciones científicas que resultaron de este trabajo. En el capítulo 2 se describen los métodos para cada técnica que fue empleada: conectividad estructural, conectividad funcional en resting state, redes cerebrales complejas y teoría de grafos y finalmente se describe la condición de deterioro cognitivo leve y el estado actual en la búsqueda de nuevos biomarcadores diagnósticos. En los capítulos 3, 4 y 5 se han incluido los artículos científicos que fueron producidos a lo largo de esta tesis. Estos han sido incluidos en el formato de la revista en que fueron publicados, estando divididos en introducción, materiales y métodos, resultados y discusión. Todos los métodos que fueron empleados en los artículos están descritos en el capítulo 2 de la tesis. Finalmente, en el capítulo 6 se concluyen los resultados generales de la tesis y se discuten de forma específica los resultados de cada artículo. ABSTRACT In this thesis I apply concepts from mathematics, physics and statistics to the neurosciences. This field benefits from the collaborative work of multidisciplinary teams where physicians, psychologists, engineers and other specialists fight for a common well: the understanding of the brain. Research on this field is still in its early years, being its birth attributed to the neuronal theory of Santiago Ramo´n y Cajal in 1888. In more than one hundred years only a very little percentage of the brain functioning has been discovered, and still much more needs to be explored. Isolated techniques aim at unraveling the system that supports our cognition, nevertheless in order to provide solid evidence in such a field multimodal techniques have arisen, with them we will be able to improve current knowledge about human cognition. Here we focus on the multimodal integration of magnetoencephalography (MEG) and diffusion weighted magnetic resonance imaging. These techniques are sensitive to the magnetic fields emitted by the neuronal currents and to the white matter microstructure, respectively. The combination of such techniques could bring up evidences about structural-functional synergies in the brain information processing and which part of this synergy fails in specific neurological pathologies. In particular, we are interested in the relationship between functional and structural connectivity, and how two integrate this information. We quantify the functional connectivity by studying the phase synchronization or the amplitude correlation between time series obtained by MEG, and so we get an index indicating similarity between neuronal entities, i.e. brain regions. In addition we quantify structural connectivity by performing diffusion tensor estimation from the diffusion weighted images, thus obtaining an indicator of the integrity of the white matter or, if preferred, the strength of the structural connections between regions. These quantifications are then combined following three different approaches, from the lowest to the highest level of integration, in chapters 3, 4 and 5. We finally apply the fused information to the characterization or prediction of mild cognitive impairment, a clinical entity which is considered as an early step in the continuum pathological process of dementia. The dissertation is divided in six chapters. In chapter 1 I introduce connectomics within the fields of neuroimaging and neuroscience. Later in this chapter we describe the objectives of this thesis, and the specific objectives of each of the scientific publications that were produced as result of this work. In chapter 2 I describe the methods for each of the techniques that were employed, namely structural connectivity, resting state functional connectivity, complex brain networks and graph theory, and finally, I describe the clinical condition of mild cognitive impairment and the current state of the art in the search for early biomarkers. In chapters 3, 4 and 5 I have included the scientific publications that were generated along this work. They have been included in in their original format and they contain introduction, materials and methods, results and discussion. All methods that were employed in these papers have been described in chapter 2. Finally, in chapter 6 I summarize all the results from this thesis, both locally for each of the scientific publications and globally for the whole work.

An optimal yet fast pruning algorithm to reduce latency in multiview prediction structures

We propose a new algorithm for the design of prediction structures with low delay and limited penalty in the rate-distortion performance for multiview video coding schemes. This algorithm constitutes one of the elements of a framework for the analysis and optimization of delay in multiview coding schemes that is based in graph theory. The objective of the algorithm is to find the best combination of prediction dependencies to prune from a multiview prediction structure, given a number of cuts. Taking into account the properties of the graph-based analysis of the encoding delay, the algorithm is able to find the best prediction dependencies to eliminate from an original prediction structure, while limiting the number of cut combinations to evaluate. We show that this algorithm obtains optimum results in the reduction of the encoding latency with a lower computational complexity than exhaustive search alternatives.

Systematic analysis of the decoding delay in multiview video

We present a framework for the analysis of the decoding delay in multiview video coding (MVC). We show that in real-time applications, an accurate estimation of the decoding delay is essential to achieve a minimum communication latency. As opposed to single-view codecs, the complexity of the multiview prediction structure and the parallel decoding of several views requires a systematic analysis of this decoding delay, which we solve using graph theory and a model of the decoder hardware architecture. Our framework assumes a decoder implementation in general purpose multi-core processors with multi-threading capabilities. For this hardware model, we show that frame processing times depend on the computational load of the decoder and we provide an iterative algorithm to compute jointly frame processing times and decoding delay. Finally, we show that decoding delay analysis can be applied to design decoders with the objective of minimizing the communication latency of the MVC system.

Image processing methods for human brain connectivity analysis from in-vivo diffusion MRI

The structural connectivity of the brain is considered to encode species-wise and subject-wise patterns that will unlock large areas of understanding of the human brain. Currently, diffusion MRI of the living brain enables to map the microstructure of tissue, allowing to track the pathways of fiber bundles connecting the cortical regions across the brain. These bundles are summarized in a network representation called connectome that is analyzed using graph theory. The extraction of the connectome from diffusion MRI requires a large processing flow including image enhancement, reconstruction, segmentation, registration, diffusion tracking, etc. Although a concerted effort has been devoted to the definition of standard pipelines for the connectome extraction, it is still crucial to define quality assessment protocols of these workflows. The definition of quality control protocols is hindered by the complexity of the pipelines under test and the absolute lack of gold-standards for diffusion MRI data. Here we characterize the impact on structural connectivity workflows of the geometrical deformation typically shown by diffusion MRI data due to the inhomogeneity of magnetic susceptibility across the imaged object. We propose an evaluation framework to compare the existing methodologies to correct for these artifacts including whole-brain realistic phantoms. Additionally, we design and implement an image segmentation and registration method to avoid performing the correction task and to enable processing in the native space of diffusion data. We release PySDCev, an evaluation framework for the quality control of connectivity pipelines, specialized in the study of susceptibility-derived distortions. In this context, we propose Diffantom, a whole-brain phantom that provides a solution to the lack of gold-standard data. The three correction methodologies under comparison performed reasonably, and it is difficult to determine which method is more advisable. We demonstrate that susceptibility-derived correction is necessary to increase the sensitivity of connectivity pipelines, at the cost of specificity. Finally, with the registration and segmentation tool called regseg we demonstrate how the problem of susceptibility-derived distortion can be overcome allowing data to be used in their original coordinates. This is crucial to increase the sensitivity of the whole pipeline without any loss in specificity.

Análisis de la actividad neuronal y cerebral mediante técnicas de redes complejas

El cerebro humano es probablemente uno de los sistemas más complejos a los que nos enfrentamos en la actualidad, si bien es también uno de los más fascinantes. Sin embargo, la compresión de cómo el cerebro organiza su actividad para llevar a cabo tareas complejas es un problema plagado de restos y obstáculos. En sus inicios la neuroimagen y la electrofisiología tenían como objetivo la identificación de regiones asociadas a activaciones relacionadas con tareas especificas, o con patrones locales que variaban en el tiempo dada cierta actividad. Sin embargo, actualmente existe un consenso acerca de que la actividad cerebral tiene un carácter temporal multiescala y espacialmente extendido, lo que lleva a considerar el cerebro como una gran red de áreas cerebrales coordinadas, cuyas conexiones funcionales son continuamente creadas y destruidas. Hasta hace poco, el énfasis de los estudios de la actividad cerebral funcional se han centrado en la identidad de los nodos particulares que forman estas redes, y en la caracterización de métricas de conectividad entre ellos: la hipótesis subyacente es que cada nodo, que es una representación mas bien aproximada de una región cerebral dada, ofrece a una única contribución al total de la red. Por tanto, la neuroimagen funcional integra los dos ingredientes básicos de la neuropsicología: la localización de la función cognitiva en módulos cerebrales especializados y el rol de las fibras de conexión en la integración de dichos módulos. Sin embargo, recientemente, la estructura y la función cerebral han empezado a ser investigadas mediante la Ciencia de la Redes, una interpretación mecánico-estadística de una antigua rama de las matemáticas: La teoría de grafos. La Ciencia de las Redes permite dotar a las redes funcionales de una gran cantidad de propiedades cuantitativas (robustez, centralidad, eficiencia, ...), y así enriquecer el conjunto de elementos que describen objetivamente la estructura y la función cerebral a disposición de los neurocientíficos. La conexión entre la Ciencia de las Redes y la Neurociencia ha aportado nuevos puntos de vista en la comprensión de la intrincada anatomía del cerebro, y de cómo las patrones de actividad cerebral se pueden sincronizar para generar las denominadas redes funcionales cerebrales, el principal objeto de estudio de esta Tesis Doctoral. Dentro de este contexto, la complejidad emerge como el puente entre las propiedades topológicas y dinámicas de los sistemas biológicos y, específicamente, en la relación entre la organización y la dinámica de las redes funcionales cerebrales. Esta Tesis Doctoral es, en términos generales, un estudio de cómo la actividad cerebral puede ser entendida como el resultado de una red de un sistema dinámico íntimamente relacionado con los procesos que ocurren en el cerebro. Con este fin, he realizado cinco estudios que tienen en cuenta ambos aspectos de dichas redes funcionales: el topológico y el dinámico. De esta manera, la Tesis está dividida en tres grandes partes: Introducción, Resultados y Discusión. En la primera parte, que comprende los Capítulos 1, 2 y 3, se hace un resumen de los conceptos más importantes de la Ciencia de las Redes relacionados al análisis de imágenes cerebrales. Concretamente, el Capitulo 1 está dedicado a introducir al lector en el mundo de la complejidad, en especial, a la complejidad topológica y dinámica de sistemas acoplados en red. El Capítulo 2 tiene como objetivo desarrollar los fundamentos biológicos, estructurales y funcionales del cerebro, cuando éste es interpretado como una red compleja. En el Capítulo 3, se resumen los objetivos esenciales y tareas que serán desarrolladas a lo largo de la segunda parte de la Tesis. La segunda parte es el núcleo de la Tesis, ya que contiene los resultados obtenidos a lo largo de los últimos cuatro años. Esta parte está dividida en cinco Capítulos, que contienen una versión detallada de las publicaciones llevadas a cabo durante esta Tesis. El Capítulo 4 está relacionado con la topología de las redes funcionales y, específicamente, con la detección y cuantificación de los nodos mas importantes: aquellos denominados “hubs” de la red. En el Capítulo 5 se muestra como las redes funcionales cerebrales pueden ser vistas no como una única red, sino más bien como una red-de-redes donde sus componentes tienen que coexistir en una situación de balance funcional. De esta forma, se investiga cómo los hemisferios cerebrales compiten para adquirir centralidad en la red-de-redes, y cómo esta interacción se mantiene (o no) cuando se introducen fallos deliberadamente en la red funcional. El Capítulo 6 va un paso mas allá al considerar las redes funcionales como sistemas vivos. En este Capítulo se muestra cómo al analizar la evolución de la topología de las redes, en vez de tratarlas como si estas fueran un sistema estático, podemos caracterizar mejor su estructura. Este hecho es especialmente relevante cuando se quiere tratar de encontrar diferencias entre grupos que desempeñan una tarea de memoria, en la que las redes funcionales tienen fuertes fluctuaciones. En el Capítulo 7 defino cómo crear redes parenclíticas a partir de bases de datos de actividad cerebral. Este nuevo tipo de redes, recientemente introducido para estudiar las anormalidades entre grupos de control y grupos anómalos, no ha sido implementado nunca en datos cerebrales y, en este Capítulo explico cómo hacerlo cuando se quiere evaluar la consistencia de la dinámica cerebral. Para concluir esta parte de la Tesis, el Capítulo 8 se centra en la relación entre las propiedades topológicas de los nodos dentro de una red y sus características dinámicas. Como mostraré más adelante, existe una relación entre ellas que revela que la posición de un nodo dentro una red está íntimamente correlacionada con sus propiedades dinámicas. Finalmente, la última parte de esta Tesis Doctoral está compuesta únicamente por el Capítulo 9, el cual contiene las conclusiones y perspectivas futuras que pueden surgir de los trabajos expuestos. En vista de todo lo anterior, espero que esta Tesis aporte una perspectiva complementaria sobre uno de los más extraordinarios sistemas complejos frente a los que nos encontramos: El cerebro humano. ABSTRACT The human brain is probably one of the most complex systems we are facing, thus being a timely and fascinating object of study. Characterizing how the brain organizes its activity to carry out complex tasks is highly non-trivial. While early neuroimaging and electrophysiological studies typically aimed at identifying patches of task-specific activations or local time-varying patterns of activity, there has now been consensus that task-related brain activity has a temporally multiscale, spatially extended character, as networks of coordinated brain areas are continuously formed and destroyed. Up until recently, though, the emphasis of functional brain activity studies has been on the identity of the particular nodes forming these networks, and on the characterization of connectivity metrics between them, the underlying covert hypothesis being that each node, constituting a coarse-grained representation of a given brain region, provides a unique contribution to the whole. Thus, functional neuroimaging initially integrated the two basic ingredients of early neuropsychology: localization of cognitive function into specialized brain modules and the role of connection fibres in the integration of various modules. Lately, brain structure and function have started being investigated using Network Science, a statistical mechanics understanding of an old branch of pure mathematics: graph theory. Network Science allows endowing networks with a great number of quantitative properties, thus vastly enriching the set of objective descriptors of brain structure and function at neuroscientists’ disposal. The link between Network Science and Neuroscience has shed light about how the entangled anatomy of the brain is, and how cortical activations may synchronize to generate the so-called functional brain networks, the principal object under study along this PhD Thesis. Within this context, complexity appears to be the bridge between the topological and dynamical properties of biological systems and, more specifically, the interplay between the organization and dynamics of functional brain networks. This PhD Thesis is, in general terms, a study of how cortical activations can be understood as the output of a network of dynamical systems that are intimately related with the processes occurring in the brain. In order to do that, I performed five studies that encompass both the topological and the dynamical aspects of such functional brain networks. In this way, the Thesis is divided into three major parts: Introduction, Results and Discussion. In the first part, comprising Chapters 1, 2 and 3, I make an overview of the main concepts of Network Science related to the analysis of brain imaging. More specifically, Chapter 1 is devoted to introducing the reader to the world of complexity, specially to the topological and dynamical complexity of networked systems. Chapter 2 aims to develop the biological, topological and functional fundamentals of the brain when it is seen as a complex network. Next, Chapter 3 summarizes the main objectives and tasks that will be developed along the forthcoming Chapters. The second part of the Thesis is, in turn, its core, since it contains the results obtained along these last four years. This part is divided into five Chapters, containing a detailed version of the publications carried out during the Thesis. Chapter 4 is related to the topology of functional networks and, more specifically, to the detection and quantification of the leading nodes of the network: the hubs. In Chapter 5 I will show that functional brain networks can be viewed not as a single network, but as a network-of-networks, where its components have to co-exist in a trade-off situation. In this way, I investigate how the brain hemispheres compete for acquiring the centrality of the network-of-networks and how this interplay is maintained (or not) when failures are introduced in the functional network. Chapter 6 goes one step beyond by considering functional networks as living systems. In this Chapter I show how analyzing the evolution of the network topology instead of treating it as a static system allows to better characterize functional networks. This fact is especially relevant when trying to find differences between groups performing certain memory tasks, where functional networks have strong fluctuations. In Chapter 7 I define how to create parenclitic networks from brain imaging datasets. This new kind of networks, recently introduced to study abnormalities between control and anomalous groups, have not been implemented with brain datasets and I explain in this Chapter how to do it when evaluating the consistency of brain dynamics. To conclude with this part of the Thesis, Chapter 8 is devoted to the interplay between the topological properties of the nodes within a network and their dynamical features. As I will show, there is an interplay between them which reveals that the position of a node in a network is intimately related with its dynamical properties. Finally, the last part of this PhD Thesis is composed only by Chapter 9, which contains the conclusions and future perspectives that may arise from the exposed results. In view of all, I hope that reading this Thesis will give a complementary perspective of one of the most extraordinary complex systems: The human brain.