Humans are unique in their ability to flexibly and rapidly adapt their behaviour and select courses of action that lead to future reward. Several ‘component processes’ must be implemented by the human brain in order to facilitate this behaviour. This thesis examines two such components; (i) the neural substrates supporting action selection during value- guided choice using magnetoencephalography (MEG), and (ii) learning the value of environmental stimuli and other people’s actions using functional magnetic resonance imaging (fMRI). In both situations, it is helpful to formally model the underlying component process, as this generates predictions of trial-to-trial variability in the signal from a brain region involved in its implementation. In the case of value-guided action selection, a biophysically realistic implementation of a drift diffusion model is used. Using this model, it is predicted that there are specific times and frequency bands at which correlates of value are seen. Firstly, there are correlates of the overall value of the two presented options, and secondly the difference in value between the options. Both correlates should be observed in the local field potential, which is closely related to the signal measured using MEG. Importantly, the content of these predictions is quite distinct from the function of the model circuit, which is to transform inputs relating to the value of each option into a categorical decision. In the case of social learning, the same reinforcement learning model is used to track both the value of two stimuli that the subject can choose between, and the advice of a confederate who is playing alongside them. As the confederate advice is actually delivered by a computer, it is possible to keep prediction error and learning rate terms for stimuli and advice orthogonal to one another, and so look for neural correlates of both social and non-social learning in the same fMRI data. Correlates of intentional inference are found in a network of brain regions previously implicated in social cognition, notably the dorsomedial prefrontal cortex, the right temporoparietal junction, and the anterior cingulate gyrus.

Com os avanços em eletrônica analógica e digital dos últimos 50 anos, a neurociência ganhou grande momentum e nasceu uma de suas áreas que atualmente mais recebe financiamento: neurociência computacional. Estudos nessa área, ainda considerada recente, vão desde estudos moleculares de trocas iônicas por canais iônicos (escala nanométrica), até influências de populações neurais no comportamento de grandes mamíferos (escala de até metros). O coração da neurociência computacional compreende técnicas inter- e multidisciplinares, envolvendo biologia de sistemas, bioquímica, modelagem matemática, estatística, termodinâmica, física estatística, etc. O impacto em áreas de grande interesse, como o desenvolvimento de fármacos e dispositivos militares, é a grande força motriz desta área. Especificamente para este último, a compreensão do código neural e como informação sensorial é trabalhada por populações de neurônios é essencial. E ainda estamos num estágio muito inicial de desvendar todo o funcionamento de muitos dos sistemas sensoriais mais complexos. Um exemplo é de um dos sentidos que parece existir desde as formas mais primitivas de vida: o olfato. Em mamíferos, o número de estudos parece sempre crescer com os anos. Ainda estamos, no entanto, longe de um consenso sobre o funcionamento de muitos dos mecanismos básicos do olfato. A literatura é extensa em termos bioquímicos e comportamental, mas reunir tudo em um único modelo é talvez o grande desafio atual. Nesta tese discuto, em duas partes, sistemas sensoriais seguindo uma linha bastante ligada ao sistema olfativo. Na primeira parte, um modelo formal que lembra o bulbo olfativo (de mamíferos) é considerado para investigar a relação entre a performance da codificação neural e a existência de uma dinâmica crítica. Em especial, discuto sobre últimos experimentos baseados em observações de leis de potência como evidências da existência de criticalidade e ótima performance em populações neurais. Mostro que, apesar de a performance das redes estar, sim, ligada ao ponto crítico do sistema, a existência de leis de potência não está ligada nem com tal ponto crítico, nem com a ótima performance. Experimentos recentes confirmam estas observações. Na segunda parte, discuto e proponho uma modelagem inicial para o órgão central do sentido olfativo em insetos: o Corpo Cogumelar. A novidade deste modelo está na integração temporal, além de conseguir tanto fazer reconhecimento de padrões (qual odor) e estimativa de concentrações de odores. Com este modelo, proponho uma explicação para uma recente observação de antecipação neural no Corpo Cogumelar, em que sua última camada paradoxalmente parece antecipar a primeira camada. Proponho a existência de um balanço entre agilidade do código neural contra acurácia no reconhecimento de padrões. Este balanço pode ser empiricamente testado. Também proponho a existência de um controle de ganho no Corpo Cogumelar que seria responsável pela manutenção dos ingredientes principais para reconhecimento de padrões e aprendizado. Ambas estas partes contribuem para o compreendimento de como sistemas sensoriais operam e quais os mecanismos fundamentais que os fornecem performance invejável. / With the advances in digital and analogical electronics in the last 50 years, neuroscience gained great momentum and one of its most well-financed sub-areas was born: computational neuroscience. Studies in this area, still considered recent by many, range from the ionic balance in the molecular level (scale of few nanometers), up to how neural populations influence behavior of large mammalians (scale of meters). The computational neuroscience core is highly based on inter- and multi-disciplinary techniques, involving systems biology, biochemistry, mathematical modeling, thermodynamics, statistical physics, etc. The impact in areas of current great interest, like in pharmaceutical drugs development and military devices, is its major flagship. Specifically for the later, deep understanding of neural code and how sensory information is filtered by neural populations is essential. And we are still grasping at the surface of really understanding many of the complex sensory systems we know. An example of such sensory modality that coexisted among all kinds of life forms is olfaction. In mammalians, the number of studies in this area seems to be growing steadily. However, we are still far from a complete agreement on how the basic mechanisms in olfaction work. There is a large literature of biochemical and behavioral studies, yet there is not a single model that comprises all this information and reproduces any olfactory system completely. In this thesis, I discuss in two parts sensory systems following a general line of argument based on olfaction. In the first part, a formal model that resembles the olfactory bulb (mammalians) is considered to investigate the relationship between performance in information coding and the existence of a critical dynamics. I show that, while the performance of neural networks may be intrinsically linked to a critical point, power laws are not exactly linked to neither critical points or performance optimization. Recent experiments corroborate this observation. In the second part, I discuss and propose a first dynamical model to the central organ responsible for olfactory learning in insects: the Mushroom Bodies. The novelty in this model is in the time integration, besides being able of pattern recognition (which odor) and concentration estimation at the same time. With this model, I propose an explanation for a seemingly paradoxical observation of coding anticipation in the Mushroom Bodies, where the last neural layer seems to trail the input layer. I propose the existence of a balance between accuracy and speed of pattern recognition in the Mushroom Bodies based on its fundamental morphological structure. I also propose the existence of a robust gain-control structure that sustain the key ingredients for pattern recognition and learning. This balance can be empirically tested. Both parts contribute to the understanding of the basic mechanisms behind sensory systems.

La caractérisation de la structure du cerveau humain et les motifs fonctionnelles qu’il fait apparaitre est un défi central pour une meilleure compréhension des pathologies du réseau cérébral telle que l’épilepsie du lobe temporal. Cette caractérisation pourrait aider à améliorer la prédictibilité clinique des résultats d’une chirurgie visant à traiter l’épilepsie.Le fonctionnement du cerveau peut être étudié par l’électroencéphalographie (EEG) ou par l’imagerie de résonance magnétique fonctionnelle (IRMf), alors que la structure peut être caractérisé par l’IRM de diffusion (IRMd). Nous avons utilisé ces modalités pour mesurer le fonctionnement du cerveau pendant une tache de reconnaissance de visages et pendant le repos dans le but de faire le lien entre les modalités d’une façon optimale en termes de résolution temporale et spatiale. Avec cette approche on a mis en évidence une perturbation des relations structure-fonction chez les patients épileptiques.Ce travail a contribué à améliorer la compréhension de l’épilepsie comme une maladie de réseau qui affecte le cerveau à large échelle et non pas au niveau d’un foyer épileptique local. Dans le futur, ces résultats pourraient être utilisés pour guider des interventions chirurgicales mais ils fournissent également des approches nouvelles pour évaluer des traitements pharmacologiques selon ses implications fonctionnelles à l’échelle du cerveau entier. / The understanding human brain structure and the function patterns arising from it is a central challenge to better characterize brain network pathologies such as temporal lobe epilepsies, which could help to improve the clinical predictability of epileptic surgery outcome.Brain functioning can be accessed by both electroencephalography (EEG) or functional magnetic resonance imaging (fMRI), while brain structure can be measured with diffusion MRI (dMRI). We use these modalities to measure brain functioning during a face recognition task and in rest in order to link the different modalities in an optimal temporal and spatial manner. We discovered disruption of the network processing famous faces as well a disruption of the structure-function relation during rest in epileptic patients.This work broadened the understanding of epilepsy as a network disease that changes the brain on a large scale not limited to a local epileptic focus. In the future these results could be used to guide clinical intervention during epilepsy surgery but also they provide new approaches to evaluate pharmacological treatment on its functional implications on a whole brain scale.

Comprendre le comportement du système visuel rétino-thalamo-cortico-colliculaire (i.e. précoce) dans une situation d'images naturelles est d'une importance capitale pour comprendre ce qui se passe ensuite dans le cerveau. Pour comprendre ces comportements, les neurobiologistes ont étudié les voies standard, Parvocellulaires et Magnocellulaires, depuis des décennies. Cependant, il y a aussi la voie non-standard, ou Koniocellulaire, qui joue un rôle modulateur important dans les traitements local, global, et entremêlé, pour atteindre de tels comportements. Particulièrement, l'analyse standard du mouvement réalisée par la voie Magno est alternée avec des réactions rapides, comme la fuite ou l'approche à des mouvements spécifiques, qui sont pré-câblés dans la voie Konio. De plus, l'étude d'une tâche de fixation dans une situation réelle, par exemple quand un prédateur s'approche lentement de sa proie, implique non seulement un mécanisme de mouvement, mais nécessite également l'utilisation de la voie Parvo, qui analyse, au moins, le contraste de l'image. Ici, nous étudions dans un modèle neuronal de calcul bio-inspiré comment ces voies peuvent être modélisées avec un ensemble minimal de paramètres, afin de fournir des résultats numériques robustes lors d'une tâche réelle. Ce modèle repose sur une étude approfondie pour intégrer des éléments biologiques dans l'architecture des circuits, les constantes de temps et les caractéristiques de fonctionnement des neurones. Nos résultats montrent que notre modèle, bien que fonctionnant via des calculs locaux, montre globalement un bon comportement de réseau en termes d'espace et de temps, et permet d'analyser et de proposer des interprétations de l'interaction entre le thalamus et le cortex. À une échelle plus macroscopique, les comportements du modèle sont reproductibles et peuvent être qualitativement comparés à des mesures de fixation oculaire chez l'homme. Cela est également vrai lorsque l'on utilise des images naturelles, où quelques paramètres sont légèrement modifiés, en gardant des résultats qualitativement humains. Les résultats de robustesse montrent que les valeurs précises des paramètres ne sont pas critiques, mais leur ordre de grandeur l'est. Une instabilité numérique ne se produit qu'après une variation de 100% d'un paramètre. Nous pouvons donc conclure que cette approche systémique est capable de représenter les changements de l'attention en utilisant des images naturelles, tout en étant algorithmiquement robuste. Cette étude nous donne ainsi une interprétation possible sur le rôle de la voie Konio, tandis qu'en même temps elle nous permet de participer au débat sur les low et high-roads des flux attentionnel et émotionnel. Néanmoins, d'autres informations, comme la couleur, sont également présentes dans le système visuel précoce, et pourraient être prises en considération, ainsi que des mécanismes corticaux plus complexes, dans les perspectives de ce travail / Understanding the behavior of the retino-thalamo-cortico-collicular (i.e. early) visual system in a natural images situation is of utmost importance to understand what further happens in the brain. To understand these behaviors, neuroscientists have looked at the standard Parvocellular and Magnocellular pathways for decades. However, there is also the non-standard Koniocellular pathway, which plays an important modulating role in the local, global, and intermingled processing carried out to achieve such behaviors. Particularly, the standard motion analysis carried out by the Magno pathway is alternated with rapid reactions, like fleeing or approaching to specific motions, which are hard-wired in the Konio pathway. In addition, studying a fixation task in a real situation, e.g., when a predator slowly approaches its prey, not only involves a motion mechanism, but also requires the use of the Parvo pathway, analyzing, at least, the image contrast. Here, we study in a bio-inspired computational neural model how these pathways can be modeled with a minimal set of parameters, in order to provide robust numerical results when doing a real task. This model is based upon an important study to integrate biological elements about the architecture of the circuits, the time constants and the operating characteristics of the different neurons. Our results show that our model, despite operating via local computations, globally shows a good network behavior in terms of space and time, and allows to analyze and propose interpretations to the interplay between thalamus and cortex. At a more macroscopic scale, the behaviors emerging from the model are reproducible and can be qualitatively compared to human-made fixation measurements. This is also true when using natural images, where just a few parameters are slightly modified, keeping the qualitatively human-like results. Robustness results show that the precise values of the parameters are not critical, but their order of magnitude matters. Numerical instability occurs only after a 100% variation of a parameter. We thus can conclude that such a reduced systemic approach is able to represent attentional shifts using natural images, while also being algorithmically robust. This study gives us as well a possible interpretation about the role of the Konio pathway, while at the same time allowing us to participate in the debate between low and high-roads in the attentional and emotional streams. Nevertheless, other information, such as color, is also present in the early visual system, and should be addressed together with more complex cortical mechanisms in a sequel of this work

De nos jours, les opérations chirurgicales sont impossibles à imaginer sans anesthésie générale, qui implique la perte de conscience, l'immobilité, l'amnésie et l'analgésie. La compréhension des mécanismes sous-jacents de chacun de ces effets garantit un traitement médical bien contrôlé. Cette thèse se concentre sur l'effet analgésique de l'anesthésie générale, précisément, sur la réaction du patient aux stimuli nociceptifs. Nous étudions également les différences des réactions entre différents médicaments anesthésiques. L'étude a été effectuée sur un ensemble de données constituées de 230 signaux EEG : enregistrements pré- et post-incision obtenus sur 115 patients qui ont reçu du desflurane et du propofol. La première phase de l'étude comprend l'analyse spectrale de puissance, qui est une méthode très répandue dans le traitement du signal. L'information spectrale a été décrite en ajustant l'activité de fond, qui exhibe un comportement $1/f$, aux estimations de la densité spectrale de puissance des signaux d'EEG et en mesurant la puissance contenue dans des bandes delta et alpha par rapport à la puissance de l'activité de fond. Une autre amélioration a été réalisée par l'expansion des spectres avec des informations de temps en raison de la nature non stationnaire observée dans les signaux EEG. Pour obtenir les représentations temps-fréquence des signaux nous appliquons trois méthodes différentes: scalogramme (basé sur la transformée en ondelettes continue), spectrogramme classique, et réaffectation de spectrogramme. Celle-ci permet d'améliorer la lisibilité d'une représentation temps-fréquence en réaffectant l'énergie contenue dans le spectrogramme à des positions plus précises. Par la suite, les spectrogrammes obtenus ont été utilisés pour la reconstruction de l'espace de phase, pour l'analyse récurrence et pour sa quantification par une mesure de complexité. L'analyse de récurrence permet de décrire et visualiser les dynamiques récurrentes d'un système et de découvrir des motifs structurels contenus dans les données. Ici, les diagrammes de récurrence ont été utilisés comme réécriture de grammaire pour transformer le signal original en une séquence symbolique, où chaque symbole représente un certain état du système. Trois mesures de complexité différentes sont alors calculées à partir de ces séquences symboliques afin de les utiliser comme éléments de classification. Enfin, en combinant les caractéristiques obtenues avec l'analyse spectrale de puissance et avec l'analyse symbolique de récurrence, nous effectuons la classification des données en utilisant deux méthodes de classification~: l'analyse discriminante linéaire et les machines à vecteurs de support. La classification a été effectuée sur des problèmes à deux classes, la distinction entre les signaux EEG pré- / post-incision, ainsi qu'entre les deux différents médicaments anesthésiques, desflurane et propofol. / Nowadays, surgical operations are impossible to imagine without general anesthesia, which involves loss of consciousness, immobility, amnesia and analgesia. Understanding mechanisms underlying each of these effects guarantees well-controlled medical treatment. This thesis focuses on analgesia effect of general anesthesia, more specifically, on patients reaction to nociceptive stimuli. We also study differences in the reaction between different anesthetic drugs. The study was conducted on dataset consisting of 230 EEG signals: pre- and post-incision recordings obtained form 115 patients, who received desflurane and propofol. The first stage of the study comprise power spectral analysis, which is a widespread approach in signal processing. Spectral information was described by fitting the background activity, that exposes $1/f$ behavior, to power spectral density estimates of the EEG signals and measuring power contained in delta and alpha bands relatively to the power of background activity. A further improvement was done by expanding spectra with time information due to observed non-stationary nature of EEG signals. To obtain time-frequency representations of the signals we apply three different methods: scalogram (based on continuous wavelet transform), conventional spectrogram, and spectrogram reassignment. The latter allows to ameliorate readability of a time-frequency representation by reassigning energy contained in spectrogram to more precise positions. Subsequently, obtained spectrograms were used as phase space reconstruction in recurrence analysis and its quantification by complexity measure. Recurrence analysis allows to describe and visualize recurrent dynamics of a system and discover structural patterns contained in the data. Here, recurrence plots were used as rewriting grammar to turn an original signal into a symbolic sequence, where each symbol represents a certain state of the system. After computing three different complexity measures of resulting symbolic sequences they are used as features for classification. Finally, combining features obtained with power spectral analysis and recurrence symbolic analysis, we perform classification of the data using two classification methods: linear discriminant analysis and support vector machines. Classification was carried out on two-class problem, distinguishing between pre-/post-incision EEG signals, as well as between two different anesthetic drugs, desflurane and propofol.

Das menschliche Gehirn funktioniert weitgehend zuverlässig – egal ob man im Schneegestöber nach einer schützenden Unterkunft sucht oder im Hochsommer einen Marathon läuft. Der Grund hierfür liegt im Erhalt einer nahezu konstanten Körpertemperatur, der für den menschlichen Organismus einen hohen Energieaufwand darstellt. Dadurch verliert die Temperaturabhängigkeit chemischer Prozesse auf mikroskopischer Ebene für den Menschen an Bedeutung – im Gegensatz zu allen wechselwarmen Lebewesen, deren Körpertemperatur sich der Umgebungstemperatur umgehend anpasst. Dass lebenswichtige Körper- und Gehirnfunktionen vieler Wechselwarmer dennoch über einen breiten Temperaturbereich funktionieren, legt nahe, dass sich diese Tiere Mechanismen zu Nutze machen, die die Temperaturabhängigkeit auf mikroskopischer Ebene ausgleichen. Die vorliegende Arbeit beschreibt Möglichkeiten der so genannten Temperaturkompensation am Beispiel des Hörsystems der Heuschrecke. Für einige Heuschreckenarten ermöglicht das Hörsystem die Lokalisierung und Identifizierung möglicher Partner anhand von Werbegesang, auch bei schlechten Sichtverhältnissen in hoher Vegetation. Insbesondere funktioniert die akustische Kommunikation über eine Temperaturspanne von bis zu 15°C. Diese Doktorarbeit erklärt zum einen, wie einzelne Nervenzellen mit temperaturabhängigen Ionenkanälen eine temperaturkompensierte Stimulusrepräsentation erzeugen können. Weiterhin wird gezeigt, dass der zugrundeliegende zell-intrinsische Kompensationsmechanismus nicht den neuronalen Energieverbrauch beeinträchtigen muss. Zum anderen wird belegt, dass die Schallverarbeitung auf höheren Verarbeitungsstufen selbst nicht temperaturkompensiert ist. Anhand mathematischer und computergestützter Modelle wird erläutert wie dennoch mit der gemessenen Temperaturabhängigkeit der neuronalen Verarbeitung temperaturkompensierte Gesangserkennung ermöglicht wird. Die vorgeschlagenen Mechanismen können auf alle wechselwarmen Organismen verallgemeinert werden. / The human brain largely remains functional regardless of whether one is searching for the shortest path to a warming shelter in a snowstorm or running a marathon on a summer’s day. This robustness of brain functionality can be attributed to the maintenance of a constant body temperature, which requires a large investment of energy. Due to homeothermy, the temperature dependence of all chemical reactions, including those inside the body, loses relevance as a constraint for humans. For poikilotherms, in contrast, a rise in ambient temperature translates to an increase in body temperature, which speeds up all chemical processes. Yet, many poikilotherms exhibit robustness of vital behaviors across a broad range of temperatures, which suggests the existence of mechanisms that compensate for temperature dependencies at the microscopic level. The present thesis proposes mechanisms for such temperature compensation, using the auditory system of the grasshopper as a model system. For various grasshopper species, the auditory system facilitates localization and recognition of conspecifics under conditions of low visibility. In particular, communication and recognition remain functional across a temperature range of up to 15 C. Here, we show on the one hand how single nerve cells with temperature-dependent ion channels can generate a temperature-compensated stimulus representation. Importantly, we reveal that the underlying cell-intrinsic compensation mechanism need not impair neuronal energy efficiency. On the other hand, we show that sound processing in higher-order neurons does not exhibit the degree of compensation that is found at the input level. Using a combination of mathematical modeling and simulations we show how temperature compensation of song recognition can be achieved at the network level, with temperature-dependent neural filters. In principle the proposed mechanisms are applicable to all poikilothermic species.

Submitted by Maria Cristina (library@lncc.br) on 2015-10-08T14:40:37Z No. of bitstreams: 1 thesiskarine.pdf: 6490137 bytes, checksum: 5cfc4bafd419dffab8b58cc3783124ae (MD5) / Approved for entry into archive by Maria Cristina (library@lncc.br) on 2015-10-08T14:40:53Z (GMT) No. of bitstreams: 1 thesiskarine.pdf: 6490137 bytes, checksum: 5cfc4bafd419dffab8b58cc3783124ae (MD5) / Made available in DSpace on 2015-10-08T14:41:02Z (GMT). No. of bitstreams: 1 thesiskarine.pdf: 6490137 bytes, checksum: 5cfc4bafd419dffab8b58cc3783124ae (MD5) Previous issue date: 2015-03-30 / Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) / In this work we develop a neurocomputational model based on ordinary differential equations which describes the interaction between the reward circuit and the thalamocortical circuit, taking into account the influence of astrocyte. The physiology for these circuits is studied by a coupled model, used to obtain numerical results that describe the action potential behavior associated to each neuron in the neural network. The initial value equations of the proposed models are discretized using classical numerical methods. Thus, it is possible to study the attentional focus behavior when an exogenous substance is added to the system, in particular, to study the influence of nicotine on the attentional focus. The proposed modeling is applied on problems arising in medicine, specifically, in neuropsychiatry. The study cases refer to patients with chemical dependence in nicotine and attention deficit hyperactivity disorder (ADHD) / Neste trabalho desenvolvemos um modelo neurocomputacional baseado em equações diferenciais ordinárias, que descreve a interação entre o circuito da recompensa e o circuito tálamo-cortical, considerando a influência do astrócito. O estudo da fisiologia destes circuitos inspira a construção de um modelo acoplado para ser usado na obtenção de resultados numéricos que descrevem o comportamento do potencial de ação associado a cada neurônio da rede neural. Os problemas de valor inicial que representam os modelos estudados são discretizados usando métodos numéricos clássicos. Desta forma, é possível estudar o comportamento do foco de atenção quando uma substância exógena é adicionada ao sistema, em particular, estudar a influência da nicotina no foco de atenção. A modelagem aqui proposta é aplicada em problemas advindos da medicina, especificamente, da área de neuropsiquiatria. Os casos de estudos estudo estão restritos a pacientes com problemas de dependência química em nicotina e pacientes com transtorno de déficit de atenção e hiperatividade (TDAH).

Medidas de dependência entre séries temporais são estudadas com a perspectiva de evidenciar como diferentes regiões do cérebro interagem, por meio da aplicação a sinais eletrofisiológicos. Baseado na representação auto-regressiva e espectral de séries temporais, diferentes medidas são comparadas entre si, incluindo coerência espectral e a coerência parcial direcionada, e introduz-se uma nova medida, denominada transferência parcial direcionada. As medidas são analisadas pelas propriedades de parcialização, relações diretas ou indiretas e direcionalidade temporal, e são mostradas suas relações com a correlação quadrática. Conclui-se que, entre as medidas analisadas, a coerência parcial direcionada e a transferência parcial direcionada possuem o maior número de características desejáveis, fundamentadas no conceito de causalidade de Granger. A estatística assintótica é desenvolvida para todas as medidas, incluindo intervalo de confiança e teste de hipótese nula, assim como sua implementação computacional. A aplicação a séries simuladas e a análise de dados eletrofisiológicos reais ilustram o estudo comparativo e a aplicabilidade das novas estatísticas apresentadas. / Measures of dependence between temporal series are studied in the context of revealing how different brain regions interact, through their application to electrophysiology. Based on the spectral and autoregressive model of time series, different measures are compared, including coherence and partial directed coherence, and a new measure is introduced, named partial directed transfer. The measures are analyzed through the properties of partialization, direct or indirect relations and temporal directionality, and their relation to quadratic correlation is shown. It results that among the presented measures, partial directed coherence and partial directed transfer reveal the highest number of desirable properties, being grounded on the concept of Granger causality. The asymptotic statistics for all measures are developed, including confidence intervals and null hypothesis testing, as well as their computational implementation. The application to simulated series and the analysis of electrophysiological data illustrate the comparative study and the applicability of the newly presented statistics.

Neurons can respond to input by firing isolated action potentials or spikes. Sequences of spikes have been linked to the encoding of neuron input. However, many neurons also fire bursts; mechanistically distinct responses consisting of brief high-frequency spike firing. Bursts form separate response symbols but historically have not been thought to encode input. However, recent experimental evidence suggests that bursts can encode input in parallel with tonic spikes. The recognition of bursts as distinct encoding symbols raises important questions; these form the basic aims of this thesis: (1) What inputs do bursts encode? (2) Does burst structure provide extra information about different inputs. (3) Is burst coding robust against the presence of noise; an inherent property of all neural systems? (4) What mechanisms are responsible for burst input encoding? (5) How does burst coding manifest in in-vivo neurons. To answer these questions, bursting is studied using a combination of neuron models and in-vivo hippocampal neuron recordings. Models ranged from neuron-specific cell models to models belonging to three fundamentally different burst dynamic classes (unspecific to any neural region). These classes are defined using concepts from non-linear system theory. Together, analysing these model types with in-vivo recordings provides a specific and general analysis of burst encoding. For neuron-specific and unspecific models, a number of model types expressing different levels of biological realism are analysed. For the study of thalamic encoding, two models containing either a single simplified burst-generating current or multiple currents are used. For models simulating three burst dynamic classes, three further models of different biological complexity are used. The bursts generated by models and real neurons were analysed by assessing the input they encode using methods such as information theory, and reverse correlation. Modelled bursts were also analysed for their resilience to simulated neural noise. In all cases, inputs evoking bursts and tonic spikes were distinct. The structure of burst-evoking input depended on burst dynamic class rather than the biological complexity of models. Different n-spike bursts encoded different inputs that, if read by downstream cells, could discriminate complex input structure. In the thalamus, this n-spike burst code explains informative responses that were not due to tonic spikes. In-vivo hippocampal neurons and a pyramidal cell model both use the n-spike code to mark different LFP features. This n-spike burst may therefore be a general feature of bursting relevant to both model and in-vivo neurons. Bursts can also encode input corrupted by neural noise, often outperforming the encoding of single spikes. Both burst timing and internal structure are informative even when driven by strongly noise-corrupted input. Also, bursts induce input-dependent spike correlations that remain informative despite strong added noise. As a result, bursts endow their constituent spikes with extra information that would be lost if tonic spikes were considered the only informative responses.

Ma Ting. / "July 2004." / Thesis (Ph.D.)--Chinese University of Hong Kong, 2004. / Includes bibliographical references (p. 181-196). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Mode of access: World Wide Web. / Abstracts in English and Chinese.