Dissertations / Theses on the topic 'Resonant neurons'

Diese Dissertation befasst sich mit dem Einfluss von Resonanz und Synchronisation auf die Präzision und die Zuverlässigkeit von Rezeptorneuronen. Präzision von individuellen Neuronen an der Peripherie eines Nervensystems, beispielsweise in sensorischen Neuronen, ist äusserst wichtig für höhere Stufen der Verarbeitung. Verschiedene Formen von Resonanz können dazu führen, dass sich die Präzision eines Neurons erhöht. Hier wird neuronale Timing-Resonanz untersucht: diese kommt vor, wenn ein Neuron für Signale mit Frequenzen um seine Resonanzfrequenz - seiner Feuerrate - Aktionspotentiale (Spikes) mit höherer Präzision produziert, als für andere Frequenzen. Mit Hilfe von elektrophysiologischen Experimenten an auditorischen Rezeptorneuronen der Heuschrecke Locusta migratoria werden Spike-Antworten gewonnen, welche mit verschiedenen Zuverlässigkeitsmassen auf ihre Präzision untersucht werden. Verschiedene auditorische Stimulus-Typen und Stimulus-Parameter werden verwendet, um Kopplungsverhältnisse zwischen der Stimulusfrequenz und der Spike-Antwort und deren Einfluss auf Spike-Zeiten-Zuverlässigkeit, Phasen-Kopplung und Spike-Jitter zu untersuchen. Dabei werden durch Variation der Stimulusamplitude sogenannte Arnold-Zungen sichtbar. Der deutlichste Effekt ist für Stimulusfrequenzen in der Nähe der mittleren Feuerrate zu sehen, wo die Breite der Arnold-Zunge ansteigt, wenn die Stimulusamplitude erhöht wird und erhöhte Werte für die Zuverlässigkeitsmasse vorhanden sind.
This thesis deals with the effect of resonance and synchronisation on the precision and reliability of receptor neurons. Precision of individual neurons at the periphery of a nervous system, for example sensory neurons, is very important for later stages of processing. Different forms of resonance lead to an increase of precision in a neuron. Here, we examine neuronal timing resonance: a neuron produces action potentials (spikes) with greater precision around its resonance frequency - its firing rate - than at other frequencies. By using electrophysiological experiments on auditory receptor neurons of the locust Locusta migratoria, spike responses are generated whose precision is investigated using different reliability measures. Different types of auditory stimuli and stimulus parameters are used to examine locking of the spike response to the frequency of the stimulus, and the influence this locking has on spike time reliability, phase coupling and spike jitter. By varying the stimulus amplitude, so-called Arnold tongues become visible. The most prominent effect is seen for stimulus frequencies around the average firing rate, where the width of the Arnold tongue and the values of the reliability measures increases for increasing stimulus amplitudes.

Le concept de "cognition incarnée" considère que le schéma classique Perception-Cognition-Action proposant un flux séquentiel de traitement de l’information n'est pas approprié pour comprendre l'effet comportemental des troubles neurodégénératifs et trouver des solutions thérapeutiques innovantes. La découverte des neurones miroirs (NM) a donné un substrat biologique à cette théorie: on pense maintenant que les NM relient les connaissances sur les actions et les perceptions non seulement pour intégrer la perception dans la planification et l'exécution, mais aussi pour soutenir un large éventail de fonctions cognitives, par ex. empathie et langage. En même temps, il est maintenant clair que dans chaque maladie neurodégénérative les symptômes cognitifs et moteurs sont représentés le long d'un continuum. Les maladies neurodégénératives liées au vieillissement, comme la maladie d'Alzheimer (MA), la forme la plus courante de démence, sont devenues un enjeu social très important. Comme il n'y a pas de remède pour la MA, les études se concentrent sur la prévention. Une catégorie qui représente maintenant une cible privilégiée est le trouble cognitif léger (TCL), considéré comme une étape intermédiaire entre le vieillissement normal et la MA. Même si MA et TCL ont été caractérisées comme des maladies «cognitives» jusqu'à présent, un lien entre la fonction motrice et le risque de développer la MA a été reconnu.Le but principal de cette recherche est d'étudier l'intégrité du réseau NM dans la MA, le TCL et le vieillissement normal. La caractérisation de son fonctionnement dans les maladies neurodégénératives serait utile pour une meilleure compréhension de leurs mécanismes fonctionnels et manifestations cliniques. Cela permettrait également d’exploiter le NM dans la réhabilitation des symptômes.La thèse comprend deux parties : la première inclue une vaste recherche bibliographique destinée à décrire le cadre scientifique qui justifie une telle recherche.Nous avons d'abord passé en revue les preuves sur l'existence d'un système NM chez les singes et les humains, et ses multiples rôles possibles et après brièvement décrit le tableau clinique des principaux troubles neurodégénératifs, en montrant comment les symptômes cognitifs et moteurs s’entrecroisent. Ensuite, nous avons détaillé les résultats de la recherche documentaire sur les maladies neurodégénératives, NM et cognition incarnée, en les commentant à la lumière de cette théorie.La deuxième partie de la thèse décrit la procédure expérimentale qui a été réalisée dans le but de la recherche.Trois groupes appariés de 16 sujets chacun (CA-contrôles âgées, TCL amnésique avec atrophie hippocampique et MA) ont été évalués avec une batterie neuropsychologique centrée sur les fonctions liées au système NM, et une tâche IRMf spécifiquement créée pour tester les NM: celle- ci était constituée d'une tache d’observation, où aux sujets ont été montrés des vidéos d'une main droite saisissant différents objets, et d'une tache motrice où les sujets ont observé des images d'objets orientés pour être saisis avec la main droite, et ont fait le geste correspondant.Chez les CA, l'analyse de conjonction (comparant l'activation de l'IRMf pendant l'observation et l'exécution) a indiqué l'activation d'un réseau bilatéral fronto-pariétal dans les zones NM « classiques» et du gyrus temporal supérieur (STG), entrée visuelle corticale aux NM. Le groupe TCL a montré une activation similaire, cependant, les zones pariétales ont été moins activées et le STG n'a pas été activé, tandis que l'inverse était vrai pour la zone de Broca droite. Nous n'avons observé aucune activation du réseau fronto-pariétal chez le groupe MA. Dans tous les tests neuropsychologiques (y compris les tests de fonctions attribuées à NM), les sujets MA ont été plus mauvais que les CA, alors que les sujets TCL montraient seulement des troubles de mémoire épisodique et fluidité sémantique (...)
The concept of “embodied cognition” considers that the classical Perception-Cognition-Action architecture proposing a sequential flow of processing with clean cuts between all modules is not appropriate to understand the behavioral effect of neurodegenerative disorders and to find innovative therapeutic solutions. In the last decades, the discovery of the mirror neurons (MN) has given a biological substrate to this theoretical perspective: the MN are now thought linking together knowledge about actions and perceptions not only to integrate perception in action planning and execution but also as a neural mechanism supporting a wide range of cognitive functions, e.g. empathy and language. At the same time, it is now clear that in each neurodegenerative disease both cognitive and motor symptoms are represented along a continuum. In the current demographic context, neurodegenerative diseases linked to aging have become a very important social issue. Alzheimer Disease (AD), the most common form of dementia, is a neurodegenerative disease strictly linked to aging. As actually there is no cure, several studies are focusing on prevention. A category which now represents a preferential target of intervention is Mild Cognitive Impairment (MCI), considered as an intermediate stage between normal aging and AD. Even if AD and MCI have been characterized as “cognitive” diseases until now, a link between motor function and the risk of developing AD has been recognized.The main purpose of this research is to investigate the integrity of the MN network in AD, MCI and normal aging. Characterizing the functioning of the MN network in neurodegenerative diseases would be useful to better understand functional mechanisms and their clinical manifestations. It would also allow to capitalize on these kinds of neurons in the rehabilitation of motor and cognitive symptoms.The thesis consists of two parts: the first part includes an extensive bibliographic research intended to describe the scientific frame which justifies such a research.We first reviewed the evidence about the existence of a MN system in monkeys and humans, and its multiple possible roles in humans.We then briefly reviewed the clinical picture of the main neurodegenerative disorders, showing how cognitive and motor symptoms intersect in all of them.Next, we detailed the results of literature searching on neurodegenerative diseases, MN, and embodied cognition, commenting them at the light of this hypothesis.The second part of the thesis describe the experimental procedure which has been performed to evaluate the integrity of the MN network in normal elderly and people with AD and MCI, and its results.Three matched groups of 16 subjects each (normal elderly-NE, amnesic MCI with hippocampal atrophy and AD) were evaluated with a neuropsychological battery centered on functions thought to be linked to the MN system, and a fMRI task specifically created to test MN: that comprised of an observation run, where subjects were shown videos of a right hand grasping different objects, and of a motor run, where subjects observed visual pictures of objects oriented to be grasped with the right hand, and made the corresponding gesture.In NE subjects, the conjunction analysis (comparing fMRI activation during observation and execution), indicated the activation of a bilateral fronto-parietal network in “classical” MN areas, and of the superior temporal gyrus (STG), an area thought to provide the cortical visual input to the MN. The MCI group showed the activation of areas belonging to the same network, however, parietal areas were activated to a lesser extent and the STG was not activated, while the opposite was true for the right Broca’s area. We did not observe any activation of the fronto-parietal network in AD participants (...)

Dans la perspective de mieux comprendre les phénomènes de compartimentation cellulaire au sein du système nerveux central, nous avons choisi de définir les caractéristiques métaboliques des neurones et des astrocytes en culture primaire homogène. Le devenir métabolique du [1-13C]glucose dans les neurones et les astrocytes cérébelleux, de même que dans les astrocytes corticaux, a été caractérisé par spectroscopie de RMN. Les astrocytes, contrairement aux neurones, synthétisent la glutamine. La maturation des voies de biosynthèse de cet acide aminé est retardée dans les astrocytes cérébelleux par rapport aux astrocytes corticaux. La quantification des flux du catabolisme du glucose exogène a été réalisée. Ces résultats ont montré l'utilisation quasi-exclusive du glucose comme source de carbone par les neurones, alors que les astrocytes utilisent des sources plus diversifiées (glucose, acides aminés exogènes, sources endogènes de carbone). De même, l'activité de la voie de la pyruvate carboxylase est de faible importance dans les neurones, ce qui implique la nécessité d'un apport de carbone extérieur pour ces cellules. Cette étude nous a permis de mettre en évidence des composés synthétisés et libérés par les astrocytes dans le milieu extracellulaire, l'alanine et le citrate, susceptibles de servir de navettes de carbone et/ou d'azote, autres que la glutamine, entre les neurones et les astrocytes. Les données acquises par RMN du 31P ont révélé des charges énergétiques très similaires dans les neurones et les astrocytes cérébelleux, de même que dans le cervelet entier. Des différences concernant les composés liés au métabolisme des membranes ont pu être observées. Une étude du développement du cervelet de rat a été réalisée par RMN du 31P et du 1H, démontrant l'existence d'un contenu élevé en acétate dans le cervelet à la naissance. Celui-ci décroît lors des 1ers jours postnataux, alors que la concentration en NAA augmente
In order to investigate the cellular compartmentation of the central nervous system, we first defined the metabolic properties of neurons and astrocytes in homogenous primary culture. The metabolic fate of [1-13C]glucose in cerebellar neurons and astrocytes, as well as in cortical astrocytes, was characterized by NMR spectroscopy. The astrocytes, contrary to neurons, synthesized glutamine. The maturation of the glutamine synthesis pathway was delayed in cerebellar astrocytes, as compared to cortical astrocytes. The fluxes involved in exogenous glucose utilization were quantified. The results demonstrated that if neurons used exclusively glucose as carbon source to fuel the Krebs cycle, the carbon sources for astrocytes were diversified (glucose, exogenous amino acids, endogenous carbon sources). In the same way, the pyruvate carboxylase activity was of minor importance in neurons, that implied the need for these cells of exogenous carbon substrates. We evidenced that alanine and citrate were also synthesized by astrocytes and exported to their extracellular medium. These metabolites may play a role as carbon and/or nitrogen shuttles betwen neurons and astrocytes. 31P NMR data showed similar energy charges in cerebellar neurons, astrocytes and in the cerebellum. Differences in the content of metabolites linked to membrane metabolism were observed. The postnatal development of the cerebellum was studied using 31P and 1H NMR spectroscopy. A large content of acetate was evidenced at birth, that decreased during the first postnatal days whereas the NAA content increased

La science a avancé depuis le XIX ème siècle. Des nouveaux outils sont apparus : la microscopie optique nous donne la vision des cellules, la microscopie électronique nous entraine au cœur de celles-ci. L’imagerie par résonance magnétique est apparue dans les années soixante-dix. Depuis son évolution, l’IRM nous entraine de plus en plus loin dans les profondeurs secrètes de nos cerveaux. La possibilité d’observer l’activité neuronale à l’aide de l’imagerie fonctionnelle est une grande révolution. Cette thèse montrera la possibilité que l’on a d’observer l’activité d’un neurone individuel sans modification de son réseau grâce à l’imagerie rehaussée au manganèse. L’étude sera effectuée sur des Aplysies à très haut champ (17T). Ces animaux sont des mollusques marins gastropodes qui possèdent une particularité : leurs neurones sont de tailles importantes, ils peuvent atteindre 1 mm de diamètre. Leurs neurones sont regroupés en plusieurs ganglions. Mon étude portera sur le ganglion buccal, qui est le ganglion le plus étudié dans les recherches en électrophysiologie. Avant de réaliser les acquisitions, j’ai dû concevoir plusieurs antennes de tailles microscopiques adaptées à la taille des ganglions. En réduisant la taille des antennes, le rapport signal sur bruit augmente. Dans un deuxième temps, une double antenne a été développée permettant l’acquisition de deux échantillons simultanément. Cette antenne a nécessité de créer des préamplificateurs fonctionnant à 730 MHz. La première série d’expériences a permis d’observer l’évolution de l’activité neuronale selon différents stimulus liés au comportement alimentaire des aplysies in-vivo. J’ai montré grâce à la technique mise en place que l’on peut distinguer par IRM l’activité de chaque neurone face à un stimulus. Par la suite, pour continuer ce travail, une deuxième série d’expériences a été effectuée in-vitro. J’ai étudié le comportement des neurones selon les neuro-stimulateurs perfusés : la dopamine et la sérotonine, tous les deux présents naturellement dans l’aplysie. Globalement les neurones ont été activés mais après les avoir observés individuellement, j’ai remarqué quelques différences selon les neurotransmetteurs. Cette technique peut maintenant être utilisée pour étudier d’autres conduites de l’aplysie comme le comportement compulsif. L’étude sur la mémoire peut être aussi envisagée. Les origines comportementales ont probablement des mécanismes identiques entre les différentes espèces animales et donc avec l’Homme comme l’a démontré les études d’Eric Kandel sur la mémoire
Science has evolved since the 19th century. New tools have appeared such as optical microscopy which gives us the vision of cells and electronic microscopy which leads us into their hearts. The magnetic resonance imaging appeared in the seventies. Evolving over time, the MRI has taken us farther and farther into the secret depths of our brains. The possibility of observing the neuronal activity thanks to the functional imaging is a major evolution. This thesis will show the possibility we have to observe the activity of a single neuron without modification of its network thanks to the manganese enhanced magnetic resonance imaging technique. The study was done on the Aplysia at very high field magnet (17T). These animals are marine gastropod mollusks with a peculiarity: their neurons are of important size and can reach 1 mm in diameter. Their neurons are grouped into several ganglia. My study concerns the buccal ganglion which is the most studied ganglia in the research in electrophysiology. Before making any acquisitions, I had to conceive several microscopic coils adapted to the size of the ganglions. By reducing the size of the coils, the signal of the noise ratio increases. Then, a double coil allowing the simultaneous acquisition of two samples was built. This antenna required the construction of pre-amplifiers operating at 730 MHz. The first series of experiments helped observe the evolution of the neuronal activity according to different stimuli linked to the eating habits of the Aplysia in vivo. Thanks to the technique implemented, I shall show that, using MRI, it is possible to distinguish the activity of each neuron with respect to a stimulus. Afterwards, to continue this work, a second series of experiments was made in vitro. I studied the behavior of neurons when perfused with neural stimulators: dopamine and serotonin, both naturally present in the Aplysia. Generally, all neurons were activated but when observing them individually, I noticed some differences. Studies in electrophysiology will allow us to get a better understanding and a confirmation of the results of this study. The MEMRI technique can be used in the future to study various disorders such as compulsive behaviors, which are present in the Aplysia, and probably have the same origins as in humans, given that many fundamental processes (such as memory studied by Eric Kandel who he demonstrated that human and Aplysia memories works with the same mechanism) are similar between the two species

Aquesta tesi estudia les propietats de sincronització de motius de neurones o poblacions de neurones acoblats amb retard. S'ha trobat que dos elements indirectament connectats de forma bidireccional a través d'un mediador de retransmissió dinàmica poden sincronitzar robustament la seva activitat a retard zero. L'efecte és estudiat en circuits especialment ben coneguts del cervell: xarxes corticals, circuits talamo-corticals i xarxes hipocampo-corticals. Els fonaments del mecanisme s'atribueixen a la influència de fonts ressonants: un parell d'elements directament acoblats bidireccionalment. En la presència de latència no negligible, la parella acoblada tendeix a sincronitzar en anti-fase. Aquesta propietat prevalent, intrínsecament dota a cada element amb una capacitat potencial de induir òptimament una sincronització isocrònica entre dos elements comunament conduïts. Aquest efecte, anomenat Sincronització Induïda per Ressonància, s'observa consistentment en diversos sistemes, sempre que hi hagi absència de frustració geomètrica en l'arquitectura estructural
Esta tesis estudia las propiedades de sincronización de motivos de neuronas, o de las poblaciones neuronales, acopladas con un cierto retraso. Se ha encontrado que dos elementos indirectamente conectados de forma bidireccional, a través de un mediador dinámico, pueden sincronizar de forma robusta sus actividades a tiempo cero. El efecto se estudia en circuitos del cerebro que se sabe juegan un papel fundamental: las redes corticales, circuitos tálamo-corticales y las redes hipocampo-corteza. Los fundamentos del mecanismo se atribuyen a la influencia de las fuentes de resonancia: un par de elementos bidireccionalmente acoplados. En presencia de tiempos de latencia no despreciable, el par de neuronas o poblaciones de neuronas acopladas tiende a sincronizar en oposición de fase. Esta característica predominante intrínsecamente dota a cada uno de los elementos con una capacidad potencial de inducir, de manera óptima, sincronización isócrona entre los elementos comúnmente dirigidos. Esta sincronización inducida por resonancia se observa consistentemente en varios sistemas, cuando ocurre que la frustración geométrica está ausente de la arquitectura estructural.
This thesis studies the synchronization properties of delay-coupled motifs of neurons or neuronal populations. It is found that two elements indirectly bidirectional-connected through a dynamical-relaying mediator can robustly synchronize their activity at zero-lag. The effect is studied in special well-known circuits of the brain: cortical networks, thalamocortical circuits, and hippocampal-cortical networks. The foundations of the mechanism are ascribed to the influence of resonant sources: a pair of directly bidirectional-coupled elements. In the presence of non-negligible latency, the coupled pair tends to synchronize in anti-phase. This prevalent property intrinsically endows each of the elements with a potential capability to optimally induce isochronous synchronization between commonly driven elements. This, so-called Resonance-Induced Synchronization, is consistently observed in distinct systems, whenever geometrical frustration is absent of the structural architecture

Recent progress in magnetic resonance imaging (MRI) has opened the way for micron-scale resolution, and thus for imaging biological cells. In this thesis work, we performed magnetic resonance microscopy (MRM) on the nervous system of Aplysia californica, a model particularly suited due to its simplicity and to its very large neuronal cell bodies, in the aim of studying cellular-scale processes with various MR contrasts. Experiments were performed on a 17.2 Tesla horizontal magnet, at resolutions down to 25 µm isotropic. Initial work consisted in conceiving and building radiofrequency microcoils adapted to the size of single neurons and ganglia. The first major part of the project consisted in using the manganese ion (Mn2+) as neural tract tracer in the buccal ganglia of Aplysia. Manganese is an MR contrast agent that enters neurons via voltage-gated calcium channels. We performed the mapping of axonal projections from motor neurons into the peripheral nerves of the buccal ganglia. We also confirmed the existence of active Mn2+ transport inside the neural network upon activation with the neurotransmitter dopamine. In the second major part of the project, we tested the potential of two diffusion MRI sequences for microscopy. On the one hand, we explored a very original mechanism for diffusion weighting, DESIRE (Diffusion Enhancement of SIgnal and REsolution), particularly suited for small samples. The two-dimensional DESIRE sequence was implemented and successfully tested on phantoms. The measured enhancement was consistent with theoretical predictions. Using this sequence to produce diffusion weighted images with an unprecedented contrast in biological tissue remains a challenge. On the other hand, a more "standard" sequence was implemented to measure the apparent diffusion coefficient (ADC) in nervous tissue with MRM. This sequence was a three-dimensional DP-FISP (Diffusion Prepared Fast Imaging with Steady-state free Precession), which met criteria for high resolution in a short acquisition time, with minimal artifacts. Using this sequence, we studied the changes in water ADC at different scales in the nervous system, triggered by cellular challenges. The challenges were hypotonic shock or exposure to ouabain. ADC measurements were performed on single isolated neuronal bodies and on ganglia tissue, before and after challenge. Both types of stress produced an ADC increase inside the cell and an ADC decrease at tissue level. The results favor the hypothesis that the increase in membrane surface area associated with cell swelling is responsible for the decrease of water ADC in tissue, typically measured in ischemia or other conditions associated with cell swelling.

A stochastic leaky integrate-and-fire neuron model was implemented in this study to simulate the spiking activity of the electrosensory "P-unit" receptor neurons of the weakly electric fish Apteronotus leptorhynchus. In the context of sensory coding, these cells have been previously shown to respond in experiment to natural random narrowband signals with either a linear or nonlinear coding scheme, depending on the intrinsic firing rate of the cell in the absence of external stimulation. It was hypothesised in this study that this duality is due to the relation of the stimulus to the neuron's excitation threshold. This hypothesis was validated with the model by lowering the threshold of the neuron or increasing its intrinsic noise, or randomness, either of which made the relation between firing rate and input strength more linear. Furthermore, synchronous P-unit firing to a common input also plays a role in decoding the stimulus at deeper levels of the neural pathways. Synchronisation and desynchronisation between multiple model responses for different types of natural communication signals were shown to agree with experimental observations. A novel result of resonance-induced synchrony enhancement of P-units to certain communication frequencies was also found.

Doutoramento em Física
Communication and cooperation between billions of neurons underlie the power of the brain. How do complex functions of the brain arise from its cellular constituents? How do groups of neurons self-organize into patterns of activity? These are crucial questions in neuroscience. In order to answer them, it is necessary to have solid theoretical understanding of how single neurons communicate at the microscopic level, and how cooperative activity emerges. In this thesis we aim to understand how complex collective phenomena can arise in a simple model of neuronal networks. We use a model with balanced excitation and inhibition and complex network architecture, and we develop analytical and numerical methods for describing its neuronal dynamics. We study how interaction between neurons generates various collective phenomena, such as spontaneous appearance of network oscillations and seizures, and early warnings of these transitions in neuronal networks. Within our model, we show that phase transitions separate various dynamical regimes, and we investigate the corresponding bifurcations and critical phenomena. It permits us to suggest a qualitative explanation of the Berger effect, and to investigate phenomena such as avalanches, band-pass filter, and stochastic resonance. The role of modular structure in the detection of weak signals is also discussed. Moreover, we find nonlinear excitations that can describe paroxysmal spikes observed in electroencephalograms from epileptic brains. It allows us to propose a method to predict epileptic seizures. Memory and learning are key functions of the brain. There are evidences that these processes result from dynamical changes in the structure of the brain. At the microscopic level, synaptic connections are plastic and are modified according to the dynamics of neurons. Thus, we generalize our cortical model to take into account synaptic plasticity and we show that the repertoire of dynamical regimes becomes richer. In particular, we find mixed-mode oscillations and a chaotic regime in neuronal network dynamics.
A comunicação e a cooperação entre milhares de milhões de neurónios está na base do poder do cérebro. Como é que funções cerebrais complexas emergem da dinâmica celular? Como é que grupos de neurónios se auto-organizam em padrões de atividade? Estas são questões cruciais em neurociências. Para as responder é necessário ter um sólido conhecimento teórico sobre como os neurónios comunicam ao nível microscópico, bem como de que forma ocorre atividade coletiva. Nesta tese pretendemos compreender como ´e que fenómenos coletivos complexos podem emergir num modelo simples de redes neuronais. Usando um modelo com excitação e inibição balanceadas e uma arquitectura de rede complexa, desenvolvemos métodos analíticos e numéricos para descrever a sua dinâmica neuronal. Estudamos como é que a interação entre neurónios gera vários fenómenos coletivos, tais como o aparecimento espontâneo de oscilações de rede e convulsões epilépticas, assim como também examinamos a forma de antecipar as transições para esses estados. No nosso modelo mostramos que os vários regimes dinâmicos são separados por transições de fase, e investigamos as correspondentes bifurcações e fenómenos críticos. Isto permite-nos sugerir uma explicação qualitativa do efeito Berger, e investigar fenómenos tais como avalanches, filtro passa-faixa, e ressonância estocástica. O papel da estrutura modular na deteção de sinais fracos é também discutido. Além disso, encontramos excitações não-lineares que podem descrever spikes paroxísticos observados em eletroencefalogramas de cérebros epilépticos. Tal permite-nos propor um método para prever convulsões epilépticas. A memória e a aprendizagem são funções chave no cérebro. Existem evidências de que estes processos resultam de alterações dinâmicas na estrutura cerebral. Ao nível microscópico, as conexões sinápticas são plásticas e são modificadas de acordo com a dinâmica dos neurónios. Por isso, generalizamos o nosso modelo de modo a considerar a plasticidade sináptica e mostramos que o conjunto de regimes dinâmicos se torna mais rico. Em particular, encontramos oscilações de modo misto e um regime de atividade neuronal caótica.

Neurons are the basic elements of the networks that constitute the computational units of the brain. They dynamically transform input information into sequences of electrical pulses. To conceive the complex function of the brain, it is crucial to understand this transformation and identify simple neuron models which accurately reproduce the known features of biological neurons. This thesis addresses three different features of neurons. We start by exploring the effect of subthreshold resonance on the response of a periodically forced neuron using a simple threshold model. The response is studied in terms of an implicit one-dimensional time map that corresponds to the Poincar´e map of the forced system. Qualitatively distinct responses are found, including mode locking and chaos. We analytically find the stability regions of mode-locking solutions, and identify the transition to chaos through period-adding bifurcations. We show that the response becomes chaotic when the forcing frequency is close to the resonant frequency. Then we will consider an experimentally verified model with realistic spikegenerating mechanism and study the effect of filtered synaptic fluctuations on the firing-rate response of the neuron. Using a population density method as well as an efficient numerical method, we find the steady-state firing rate in two limits of fast and slow synaptic inputs and present the linear response theory for the firing rate of the model in response to both time-dependent mean inputs and time-dependent noise intensity. Finally, a novel model is introduced that incorporates threshold variability of neurons. We determine the modulation of the input-output properties of the model due to oscillatory inputs and in the presence of filtered synaptic fluctuations.

Orientador: Naasson Pereira de Alcântara Junior
Banca: José Carlos Sartoti
Banca: José Alfredo Covolan Ulson
Resumo: Neste trabalho, apresenta-se o desenvolvimento de uma metodologia utilizando redes neurais artificiais, para auxiliar na determinação da freqüência de ressonância no projeto de antenas tipo microfita de equipamentos móveis, tanto para antenas retangulares como para antenas triangulares. Compararam-se modelos deterministas e modelos empíricos baseados em Redes Neurais Artificiais (RNA) da literatura pesquisada com os modelos apresentados neste trabalho. Apresentam-se modelos empíricos baseados em RNAs tipo Perceptron Multicamadas (PMC). Os modelos propostos também são capazes de serem integrados em um ambiente CAD (Computed Aided Design) para projetar antenas tipo microfita de equipamentos móveis.
Abstract: This work presents the development of models that can be used in the design of microstrip antennas for mobile communications. The antennas can be triangular or rectangular. The presented models are compared with deterministic and empirical models based on artificial neural networks (ANN) presented in the literature. The models are based on Perceptron Multilayer (PML). The models can be embedded in CAD systems, in order to design microstrip antennas for mobile communications.
Mestre

Die vorliegende Arbeit befasst sich mit dem Einfluss von nicht-weißem oder nicht-Gauß’schem synaptischen Rauschen auf die Informationsübertragung in stochastischen Neuronenmodellen. Ziel ist es, zu verstehen, wie eine Nervenzelle ein Signal in ihrer Pulsaktivität kodiert. Synaptisches Rauschen beschreibt hier den Einfluss anderer Nervenzellen, die nicht das interessierende Signal tragen, aber seine Übertragung durch ihre synaptische Wirkung auf die betrachtete Zelle beeinflussen. In stochastischen Neuronenmodellen wird diese Hintergrundaktivität durch einen stochastischen Prozess mit geeigneter Statistik beschrieben. Ist die Rate, mit der präsynaptische Pulse auftreten, hoch und zeitlich konstant, die Wirkung einzelner Pulse aber verschwindend gering, so wird das synaptische Rauschen durch einen Gauß’schen Prozess beschrieben. Oft wird zudem angenommen, dass das Rauschen unkorreliert (weiß) ist. In dieser Arbeit wird neuronale Signalübertragung in dem Fall untersucht, dass eine solche Näherung nicht mehr gerechtfertigt ist, d.h. wenn der synaptische Hintergrund durch einen stochastischen Prozess beschrieben werden muss, der nicht weiß, nicht Gauß’sch, oder weder weiß noch Gauß’sch ist. Mittels Simulationen und analytischer Rechnungen werden drei Szenarien behandelt: Zunächst betrachten wir eine Zelle, die nicht ein, sondern zwei Signale empfängt, welche zusätzlich durch synaptische Kurzzeitplastizität gefiltert werden. In diesem Fall muss der Hintergrund durch ein farbiges Rauschen beschrieben werden. Im zweiten Szenario betrachten wir den Fall, dass der Effekt einzelner Pulse nicht mehr als schwach angenommen werden kann. Das Rauschen ist dann nicht mehr Gauß’sch, sondern ein Schrotrauschen. Schließlich untersuchen wir den Einfluss einer präsynaptischen Population, deren Feuerrate nicht zeitlich konstant ist, sondern zwischen Phasen hoher und niedriger Aktivität, sogenannten up und down states, springt. In diesem Fall ist das Rauschen weder weiß noch Gauß’sch.
This thesis is concerned with the effect of non-white or non-Gaussian synaptic noise on the information transmission properties of single neurons. Synaptic noise subsumes the massive input that a cell receives from thousands of other neurons. In the framework of stochastic neuron models, this input is described by a stochastic process with suitably chosen statistics. If the overall arrival rate of presynaptic action potentials is high and constant in time and if each individual incoming spike has only a small effect on the dynamics of the cell, the massive synaptic input can be modeled as a Gaussian process. For mathematical tractability, one often assumes that furthermore, the input is devoid of temporal structure, i.e. that it is well described by a Gaussian white noise. This is the so-called diffusion approximation (DA). The present thesis explores neuronal signal transmission when the conditions that underlie the DA are no longer met, i.e. when one must describe the synaptic background activity by a stochastic process that is not white, not Gaussian, or neither. We explore three distinct scenarios by means of simulations and analytical calculations: First, we study a cell that receives not one but two signals, additionally filtered by synaptic short-term plasticity (STP), so that the background has to be described by a colored noise. The second scenario deals with synaptic weights that cannot be considered small; here, the effective noise is no longer Gaussian and the shot-noise nature of the input has to be taken into account. Finally, we study the effect of a presynaptic population that does not fire at a rate which is constant in time but instead undergoes transitions between states of high and low activity, so-called up and down states.

Thesis (Ph. D.)--University of California, San Diego, 2008.
Title from first page of PDF file (viewed September 24, 2008). Available via ProQuest Digital Dissertations. Vita. Includes bibliographical references (p. 133-135).

Parkinson’s disease (PD) is a progressive neurodegenerative disorder characterised by a loss of pigmented dopaminergic neurons in the substantia nigra (SN) pars compacta and loss of pigmented noradrenergic neurons in the locus coeruleus (LC). Diagnosing PD can be challenging, especially in the early stages particularly when the typical movement disorder symptoms such as tremor, rigidity, bradykinesia and postural instability are not easily identifiable. Despite well-established PD clinical diagnostic criteria there is a misdiagnosis rate of up to 15% by neurology specialists and 25 % by general practitioners. The only approved diagnostic test to confirm suspected PD in a tremulous patient is dopamine transporter single photon emission tomography (DaTScanTM). This test is costly (£800 – 1500 in the UK) and has limited geographical availability in the UK. It involves exposure to ionising radiation and can only be used to assess the integrity of the dopaminergic system. Therefore there is a strong need for better and more accessible diagnostic tests for PD. The aim of this thesis is to investigate the sensitivity and specificity of three different MRI techniques as potential biomarkers of PD. MRI at 3T field strength was used in this thesis to demonstrate PD pathology in the pigmented brain stem nuclei of SN, LC and the ventral tegmental area (VTA). The objective was to develop new, easily accessible and affordable disease markers to help clinicians to establish the correct diagnosis early. A promising technique, which is based on the assessment of free motion of water-associated protons in tissue, is termed diffusion tensor imaging (DTI). The amount of free motion in all directions of protons in tissues like the brain can be described using mean diffusivity (MD) as a measure. Diffusion in tissues like the brain is often limited (“restricted”) in certain directions. For example diffusion across the myelin sheaths of nerve-fibres in the brain white matter is constrained, whereas along the direction of the nerve fibre protons can diffuse freely. This is termed anisotropic diffusion and can be described using fractional anisotropy as a measure (FA). Microstructural PD pathological processes may alter these measures of diffusivity especially in the area of the early affected brain region of the SN. In a prospective case control study of 30 patients and 22 controls diffusion tensor imaging alterations of the SN were investigated by measuring regional alterations of fractional FA and MD. In addition, a systematic literature review and meta-analysis was performed to determine the evidence for nigral DTI alterations throughout the literature. The case control study did demonstrate a small but significant increase of nigral MD; however the meta-analysis did not confirm this result when synthesizing effect sizes of nine identified relevant studies. No significant PD induced FA alterations were found in the prospective case control study. The meta-analysis of nigral FA changes did likewise not show significant FA decrease after correcting for studies with unusual high FA measures in the control arm population. In summary the meta-analysis and the results of the case control study did not confirm that standard DTI measurements of the SN are reliable biomarkers of PD pathology. In a further case-control study MRI sequences tracking the neuromelanin content of the pigmented brain stem nuclei like the SN, LC and the ventral tegmental area were investigated. PD induced decline of neurons in these nuclei causes depigmentation due to loss of neuromelanin content. In this study (including data from 24 PD patients and 20 controls) I found that only little neuromelanin related signal could be observed in the ventral tegmental area and there was no significant difference between patients with PD and controls. However, there were significant signal alterations of the SN and LC signal when comparing between the two groups. The neuromelanin related signal loss was most pronounced in the posterior SN even in the earlier stages of the disease. The signal loss in the anterior SN was less severe and correlated with the unified PD rating scale (UPDRS) and Hoehn and Yahr score as a measure of disease severity. The neuromelanin related signal reduction was significant but less extensive in the region of the LC when compared to the SN. The signal alterations in the LC did not correlate with the UPDRS or the Hoehn and Yahr score. In the third part of the experimental section of this thesis, a further prospective case-control study of 19 participants (10 patients with PD) and retrospective study of 105 clinical cases (9 patients with PD) was performed. A high resolution SWI/T2* ‘iron sensitive’ sequence was used to assess MRI changes of the nigrosome-1. Nigrosomes are little islands of dopaminergic cells with physiologically low iron content. The healthy hyperintense signal of the linear shaped nigrosome-1 surrounded by the iron containing low signal SN regions has great resemblance to the appearance of a swallow tail. The PD induced pathological signal reduction within nigrosome-1 resulted in a loss of the typical ‘swallow tail appearance’. Visual qualitative assessment of the MRI scans for absence and presence of nigrosome-1 revealed high sensitivity and specificity (80-100% and 86-89% respectively) to allow differentiation of PD from healthy controls and non-PD patients. In summary I found that standard nigral DTI is not reliable as a PD biomarker. Nigrosome and neuromelanin weighted MRI offers great potential for development into a clinically useful biomarker. Comparing the two techniques, nigrosome imaging has some advantages over neuromelanin weighted imaging: the high resolution SWI/T2* sequence is shorter (2-5 min versus 7-14 min neuromelanin MRI. However, further optimization of neuromelanin MRI sequences may be able to shorten the acquisition time. A further advantage of nigrosome MRI is that the images can be visually assessed for pathological alterations without the need for complicated analysis or data processing. A disadvantage of high resolution SWI/T2* is that it is more prone to artefacts. An advantage of neuromelanin weighted MRI is that changes especially in the anterior substantia nigra correlate to measures of disease severity like the UPDRS, although there is some early evidence from pilot studies that nigrosome imaging (at field strengths of 7T) may also be useful to assess disease severity related changes. Which of the two techniques is better suited to monitor longitudinal progressive PD related changes has to be assessed in future studies. In conclusion standard nigral DTI measures have no proven value as a reliable diagnostic marker of PD. High resolution T2*/SWI MRI and neuromelanin weighted MRI of PD induced alteration of pigmented brain stem neurons distinguish PD from non-PD and control subjects with high sensitivity and specificity. Neuromelanin related alterations especially of the anterior SN correlate to disease severity measures like the UPDRS and therefore have potential as disease progression marker. The easy applicability of the ‘swallow tail sign’ to indicate a healthy nigrosome-1 in the SN may well prove a useful marker to help the clinical diagnosis of PD. If future studies confirm a similar diagnostic accuracy as the current clinical gold standard DaTScanTM, nigrosome MRI may replace DaTScanTM in the standard clinical setting.

Non linear dynamical systems can present a diversity of unconventional features when perturbed by an external noise, such as an enhancing of transport properties, stabilization of spatial patterns and noise induced phase transitions. In particular, the external noise can improve the system s response to weak external periodic pulses, a phenomenon termed as stochastic resonance due to its similarity with the resonance phenomena displayed by deterministic dynamical systems. Stochastic resonance ideas have been widely applied to better understand the behavior of many physical, chemical and biological systems, such as optical, electronic and magnetic systems, chemical reactions, as well as several features regarding neuro-physiological aspects of sensory systems. In this work, we study the stochastic resonance phenomenon in the integrate-fire model for the neuronal response by a sub-threshold periodic signal. In the traditional approach the threshold level is reached by superposing a gaussian noise to the periodic drive. As non gaussian noises have been shown to be quite overspread in natural systems, we investigate the sensitivity of the stochastic resonance condition upon the noise s probability distribution function. To generate a power-law distributed noise, we considered a stochastic process with both additive and multiplicative noises which allows for a fine tuning of the asymptotic power-law decay exponent. We employed both analogical and computational solutions of the stochastic differential equations which produced similar results. The dependence of the optimal noise intensity for the stochastic resonance condition on the power-law exponent of the non gaussian noise is reported. Our main finding is that the stochastic resonance condition is achieved with a minimum intensity of the input noise when it has a probability distribution with a finite decay exponent. Therefore, neural systems can explore the non gaussian character of the input noise to improve the ability to identify sub-threshold signals. Instituto de F´ısica - UFAL
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior
Sistemas dinâmicos não lineares podem apresentar uma diversidade de características não convencionais quando perturbados por ruídos externos, tais como uma otimização das propriedades de transporte, estabilização de padrões espaciais e transições de fase. Em particular, o ruído pode melhorar a resposta do sistema a pulsos periódicos externos fracos, um fenômeno conhecido como ressonância estocástica devido a sua similaridade com o fenômeno de ressonância mostrado por sistemas dinâmicos determinísticos. A idéia de ressonância estocástica foi largamente aplicada para se entender o comportamento de muitos sistemas físicos, químicos e biológicos, tais como sistemas magnéticos, ópticos, eletrônicos, reações químicas, assim como vários aspectos neurofisiológicos de sistemas sensoriais. Nesta dissertação, nós estudamos o fenômeno de ressonância estocástica em um modelo integra-dispara para resposta neuronal estimulada por um sinal periódico sub-limiar. No enfoque tradicional, o nível de limiar de disparo é alcançado por uma superposição de um ruído gaussiano com um estímulo periódico. Como ruídos não gaussianos surgem em sistemas naturais com elevada freqüência, nós investigamos a sensibilidade da condição de ressonância estocástica em relação à função distribuição de probabilidade do ruído. Para gerarmos um ruído distribuído tipo lei de potência, nós consideramos um processo estocástico com ruído multiplicativo e aditivo que permite o ajuste fino do expoente de decaimento assintótico da lei de potência. Utilizamos tanto solução analógica quanto digital de equações diferenciais estocásticas que produzem resultados similares. A dependência da intensidade ótima de ruído para a condição de ressonância estocástica com o expoente da lei de potência de um ruído não gaussiano é relatada. Em particular, obtivemos que a condição de ressonância é atingida com o mínimo ruído possível para ruídos que apresentam um expoente da lei de decaimento finito. Portanto, a natureza não gaussiana do ruído pode ser explorada para otimizar a identificação de sinais sub-limiares por sistemas neuronais.

Possible changes in brain metabolites are associated with Motor Neurone Disease (MND). Magnetic Resonance Spectroscopy (MRS) has been performed on the brains of MND patients and control volunteers to acquire signals which contain information about brain metabolites from within the motor cortex area. Discrimination between JvThD and normal groups may help to understand the pathogenic mechanisms in MND and may be useful for monitoring the effects of future trial treatment regimens. The research described in this thesis presents the development of a pattern recognition system based on neural networks to correctly distinguish between motor neurone disease (MND) patients and controls when presented with a nuclear magnetic resonance (NMR) spectrum. The NMR spectra are pre-processed to obtain consistent data, and statistical parameters are extracted and selected from each spectrum. Four statistical neural network classifiers are used to provide information and initial decisions (MND/normal). A neural network is then used to combine these to give a final decision. Experimental results indicate that the system can achieve high performance classification on the spectra, including spectra not seen by the system during training. The experiment was repeated on different training and test sets to validate the method and the repeated design shows that the final system was able to achieve high performance classification. A ifizzy rule-based system teclmique is applied to translate and extract rules encoded in weights of neural network classifiers. The neural networks are translated into a few comprehensible rules to understand how the network performs the final decision.

Made available in DSpace on 2014-06-11T19:24:38Z (GMT). No. of bitstreams: 0 Previous issue date: 2005-10-14Bitstream added on 2014-06-13T18:20:52Z : No. of bitstreams: 1 brinhole_er_me_bauru.pdf: 2082084 bytes, checksum: e9da8790c90b7d67f56ee24ea0858f1c (MD5)
Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES)
Cefet - Cornelio Procopio
Neste trabalho, apresenta-se o desenvolvimento de uma metodologia utilizando redes neurais artificiais, para auxiliar na determinação da freqüência de ressonância no projeto de antenas tipo microfita de equipamentos móveis, tanto para antenas retangulares como para antenas triangulares. Compararam-se modelos deterministas e modelos empíricos baseados em Redes Neurais Artificiais (RNA) da literatura pesquisada com os modelos apresentados neste trabalho. Apresentam-se modelos empíricos baseados em RNAs tipo Perceptron Multicamadas (PMC). Os modelos propostos também são capazes de serem integrados em um ambiente CAD (Computed Aided Design) para projetar antenas tipo microfita de equipamentos móveis.
This work presents the development of models that can be used in the design of microstrip antennas for mobile communications. The antennas can be triangular or rectangular. The presented models are compared with deterministic and empirical models based on artificial neural networks (ANN) presented in the literature. The models are based on Perceptron Multilayer (PML). The models can be embedded in CAD systems, in order to design microstrip antennas for mobile communications.

Imaging techniques have become much more in demand in modern medicine, especially in fields of disease prognosis, diagnosis and therapeutics. This is because a better understanding of different diseases, characteristics of each patient and further optimizing treatment planning, are all enhanced by advanced imaging techniques. Since each imaging modality has its own merits and intrinsic limitations, combining two or more complementary imaging modalities has become an interesting research area. In this study, gadolinium (Gd3+) doped CdTe quantum dots (QDs) were synthesized and used as multimodal imaging probes of two highly complementary imaging modalities: optical imaging and magnetic resonance imaging. The new imaging probes were characterized by X-ray diffraction (XRD), transmission electron microscopy (TEM), UV-vis absorbance spectra, fluorescence spectra (FL) and magnetic resonance imaging (MRI). The optical / MRI imaging probes were further functionalized by conjugating with the axonal tracer dextran amine (10 kDa) for non-invasive axonal tracing observations. Biocompatibility and MRI contrast effect of prepared multimodal imaging probes were investigated by in vitro cell experiments and MRI scanner. Ultimately, it is hoped that this imaging probe will help us better understand the regeneration mechanisms in real time without sacrificing animals at intervening time-points.

In vielen Kommunikationssytemen wird Information durch die zeitliche Struktur von Signalen kodiert. Ein neuronales System, welches rhythmische Signale verarbeitet, sollte davon profitieren, seine inhärenten Filtereigenschaften den Frequenzcharakteristika dieser Signale anzupassen. Die Grille Gryllus bimaculatus stellt ein einfaches biologisches System dar, für welches nur wenige, spezifische Modulationsfrequenzen verhaltensrelevant sind. Ich habe einzelne Neuronen im peripheren und höheren auditorischen System der Grille hinsichtlich einer möglichen Anpassung auf diese Frequenzen untersucht. Hierfür habe ich extrazelluläre Elektrophysiologie mit verschiedenen Stimulationsparadigmen kombiniert, welche auf amplitudenmodulierten Tönen basierten. Die Analyse der experimentellen Daten ergab, dass bereits in der auditorischen Peripherie einige der untersuchten Neurone Bandpasseigenschaften aufwiesen, da sie verhaltensrelevante Modulationsfrequenzen mit den höchsten Feuerraten beantworteten. Anhand einfacher mathematischer Modelle demonstriere ich, wie weitverbreitete, zellintrinsische und netzwerkbasierte Mechanismen die beobachteten Feuerratenresonanzen erklären könnten. Diese Mechanismen umfassen unterschwellige Resonanz von Membranströmen, aktivitätsabhängige Adaptation, sowie das Zusammenwirken von Exzitation und Inhibition. Ich zeige, wie eine serielle Kombination solcher elementarer Filter die deutliche Selektivität im Verhalten der Grille erklären könnte, ohne dabei auf ein dediziertes Filterelement zurückzugreifen. Allgegenwärtige neuronale Mechanismen könnten demnach eine dezentralisierte Filterkaskade in einem hochspezialisierten und größenbeschränkten neuronalen System begründen.
In many communication systems, information is encoded in the temporal pattern of signals. For rhythmic signals that carry information in specific frequency bands, a neuronal system may profit from tuning its inherent filtering properties towards a peak sensitivity in the respective frequency range. The cricket Gryllus bimaculatus is a simple biological system for which only a narrow range of modulation frequencies is behaviorally relevant. I examined individual neurons in the peripheral and higher auditory system for tuning towards specific temporal parameters and modulation frequencies. To this end, I combined extracellular electrophysiology with different stimulation paradigms involving amplitude-modulated sounds. Analysis of the experimental data revealed that—even in the auditory periphery—some of the examined neurons acted as tuned band-pass filters, yielding highest firing-rates for behaviorally relevant modulation frequencies. Using simple computational models, I demonstrate how common, cell-intrinsic or network-based mechanisms could account for the experimentally observed firing-rate resonances. These mechanisms include subthreshold resonances, spike-triggered adaptation, as well as the interplay of excitation and inhibition. I present how a serial combination of such elementary filters could explain the strong selectivity evident in the cricket’s behavior—without the need for a dedicated filter element. Pervasive neuronal mechanisms could therefore constitute an efficient, distributed frequency filter in a highly specialized, size-constrained neuronal system.

Orientador: Li Li Min
Dissertação (mestrado) - Universidade Estadual de Campinas, Faculdade de Ciencias Medicas
Made available in DSpace on 2018-08-09T08:01:58Z (GMT). No. of bitstreams: 1 Jacini_WantuirFranciscoSiqueira_M.pdf: 1344884 bytes, checksum: c81854d2225fd240d4f60b8fa20ab14e (MD5) Previous issue date: 2007
Resumo: Estudos experimentais têm sugerido que a prática de exercícios físicos associados a treinamento, planejamento e execução de movimentos complexos causam alterações estruturais no sistema nervoso. Essas podem refletir modificações benéficas à plasticidade cerebral, em resposta ao aumento da atividade imposta pela aprendizagem motora e exercícios físicos. Todavia, efeitos estruturais, com potencial implicação para o auxílio em reabilitação de problemas neurológicos, não são bem entendidos. A prática de judô demanda planejamento de atividades motoras complexas e exercícios físicos. A prática de corrida de fundo demanda pouca aprendizagem motora e contém exercícios físicos. Nosso objetivo foi Investigar variações no volume de substância cinzenta estimulada pela prática de judô e corrida de fundo através de morfometria baseada em voxels (VBM) em imagens de ressonância magnética (RM). Imagens de RM foram feitas em atletas de judô, de corrida de fundo e não atletas em um scanner de 2T (Elscint Prestige, Haifa, Israel). VBM foi feita usando os programas de analise matemática estatística de imagem (Statistical Parametric Mapping, SPM 2 e Matlab 7.0), utilizando imagens em T1 e voxels isotrópicos de 1mm. Os mapas probabilísticos com a substância cinzenta foram comparados voxel a voxel usando teste t-student para determinar diferenças significativas na CSC entre os grupos. Os resultados finais foram corrigidos por um algoritmo de correção para achados falso-positivos (False Discovery Rate - FDR) que minimiza o efeito das múltiplas comparações e eleva a sensibilidade do resultado final. Foram encontrados aumentos significativos no volume de substância cinzenta, no grupo de judocas em diversas áreas, descritas a seguir: Motoras primárias e suplementar no hemisfério esquerdo: giro pré-central (227 voxels); giro frontal superior (398 Voxels); giro frontal médio (34 Voxels); área motora suplementar (196 Voxels). E no hemisfério direito: área motora suplementar (51 Voxels). Áreas relacionadas a planejamento no hemisfério esquerdo: giro frontal inferior triangular (418 Voxels); rolando opercular (62 Voxels); giro frontal médio orbital (56 Voxels); giro reto (69 Voxels). No hemisfério direito, giro frontal inferior triangular, (162 Voxels) e giro reto (517 Voxels). Áreas relacionadas a concentração no hemisfério esquerdo: giro frontal superior orbital (357 Voxels); giro frontal médio orbital (305 Voxels); giro frontal inferior opercular (530 Voxels). No hemisfério direito: giro frontal superior orbital (283 Voxels); giro frontal inferior opercular (472 Voxels). Região dos lobos occipital e parietal, em áreas relacionadas à visão e associativas no hemisfério esquerdo: giro occipital médio (133 Voxels); giro parietal superior (435 Voxels); giro parietal inferior (46 Voxels). No hemisfério direito: giro occipital superior (169 Voxels); giro parietal superior (114 Voxels). Região do lobo temporal em áreas relacionadas à memória do hemisfério esquerdo: lóbulo para central (55 Voxels); giro temporal médio (326 Voxels); giro temporal inferior (1040 Voxels). No hemisfério direito: giro temporal médio (122 Voxels); giro temporal inferior (398 Voxels). Também foram encontrados aumento na concentração de substância cinzenta, no cerebelo em áreas relacionadas a aprendizagem motora no hemisfério esquerdo: cerebelo 1 (32 Voxels); cerebelo 7b (51 Voxels). No hemisfério direito: cerebelo 6 (29 Voxels). No grupo de corredores fundistas, encontramos alterações significativas, com aumentos e diminuições na concentração de substância cinzenta em diversas áreas. Diferentemente do que foi descrito no grupo de judocas, o grupo de corredores apresentou aumento e diminuição na concentração de substância cinzenta em áreas muito difusas. As alterações foram encontradas nas seguintes áreas: Na região do lobo frontal: houve aumento no volume de substância cinzenta de 14.457 voxels, entretanto apresentou uma diminuição de 5.720 voxels. A diferença entre aumento e diminuição no número de voxels que representaria ganho, foi de 8.737 voxels. Na região do lobo temporal e na insula foram encontradas mais diminuição que aumento: houve aumento no volume de substância cinzenta de 1.629 voxels, entretanto apresentou uma diminuição muito maior, de 9.580 voxels. A diferença entre aumento e diminuição no número de voxels que representaria perda, foi negativa em 7951 voxels. Região do lobo parietal foi encontrado aumento superior a diminuição no volume de substância cinzenta: houve aumento no volume de substância cinzenta de 11.298 voxels, enquanto apresentou uma diminuição de 10.725 voxels. A diferença entre aumento e diminuição no número de voxels que representaria ganho, foi de 573 voxels. Região do lobo occipital foi encontrado aumento superior a diminuição no volume de substância cinzenta: houve aumento no volume de substância cinzenta de 4.241 voxels, enquanto apresentou uma diminuição de 1.047 voxels. A diferença entre aumento e diminuição no número de voxels que representaria ganho, foi de 3.194 voxels. Também foram encontrados mais aumento que diminuição no volume de substância cinzenta no cerebelo: houve aumento no volume de substância cinzenta de 10.117 voxels, enquanto apresentou uma diminuição de 7.778 voxels. A diferença entre aumento e diminuição no número de voxels que representaria ganho, foi de 2.339 voxels. Sugerimos que a prática de esportes resulte em alterações morfológicas no córtex motor primário, cerebelo e áreas associativas. Nossos achados sugerem que planejamento motor e execução de movimento, associada a prática de exercícios físicos podem induzir melhoras na plasticidade cerebral, e justificar o uso de esportes na reabilitação de doenças neurológicas. Treinamento físico em atletas judocas e corredores fundistas induz a plasticidade neural. A magnitude da plasticidade neural são diferentes nessas duas modalidades, sugerindo que as mudanças podem ser específicas de acordo com a modalidade desportiva. O mecanismo da indução da plasticidade neural pode ser relacionado a aprendizagem motora como visto nos judocas ou relacionado a possíveis efeitos de fatores troficos e hormonais no cérebro liberados durante o exercício físico prolongado como visto nos corredores fundistas
Abstract: The regular practice of physical exercises can stimulate the cerebral plasticity, caused a great benefit consequently, to a healthy nervous system. This phenomenon are induced by increases the liberation of many local trophic factor as nerve growth factor (NGF), brain derived neurotrophic factor (BDNF) and fibroblast growth factor - 2 (FGF - 2). These local trophic factor are responsible for cell surviving, a best resistance to nervous system insults and increases the neuronal connections. Moreover, motor learning was described at literature as an interesting alternative to improve cerebral plasticity, also increasing local tropic factors. Experimental studies have suggested that physical exercise encompassing practice, planning and execution of complex movements is associated with changes in brain structure. These changes possibly reflect plastic modifications of the cortical mantle in response to an enhanced demand imposed by the motor task. In humans, cortical plasticity in relation to physical activities is yet to be fully determined and quantified. Moreover, the effects of physical exercise on brain structure, with intrinsic potential implications for rehabilitation of brain damage, are not fully understood. The practice of judo, as physical exercise, demands complex motor planning and control. The practice of long distance running demands mainly physical exercise. To investigate the variations of gray matter volume stimulated by the practice of judo, a physical exercise involved with complex motor planning and control, by using MRI voxel based morphometry (VBM). Eight judo, eight long distance runners and healthy controls were enrolled in this study. MRI was performed in a 2T scanner (Elscint Prestige Haifa, Israel). VBM was performed using the software package SPM 2, on T1-weighted images with 1 mm isotropic voxels. The resulting gray matter concentration (GMC) probabilistic maps were compared, voxel-by-voxel, using a t-test, in order to determine differences in white and gray matter concentration between controls and judo practitioners. Contrasts were defined in order to estimate the probability of each voxel being gray matter. Statistical results were correct by for multiple comparisons using FDR (false discovery rate). We found significative changes compatible with gray matter increases in the judo players group in many different areas. The increases in gray matter volume are related with the enhanced areas where are neuron cells present, and their was found areas such as: In the region of the frontal lobe; motor areas primary and supplementary in the left hemisphere: pre-central gyrus, (227 voxels); frontal superior gyrus (398 Voxels); frontal middle gyrus (34 Voxels); supplementary motor area (196 Voxels). In the right hemisphere: supplementary motor area (51 Voxels). We also found planning related areas compatible with motor actions in the left hemisphere: frontal inferior triangular gyrus (418 Voxels); rolandic opercular, (62 Voxels); frontal middle orbital gyrus (56 Voxels); rectus gyrus (69 Voxels). In the right hemisphere: frontal inferior triangular gyrus (162 Voxels); rectus gyrus, (517 Voxels). Concentration related areas in the left hemisphere: frontal superior orbital gyrus (357 Voxels); frontal middle orbital gyrus (305 Voxels); frontal inferior opercular gyrus (530 Voxels). In the right hemisphere: frontal superior orbital gyrus (283 Voxels); frontal inferior opercular gyrus (472 Voxels). Region of the parietal e occipital lobes, vision and associative related areas in the left hemisphere: occipital middle gyrus (133 Voxels); parietal superior gyrus (435 Voxels); parietal inferior gyrus (46 Voxels). In the right hemisphere: occipital superior gyrus (169 Voxels); parietal superior gyrus (114 Voxels). Region of the temporal lobe; memory related areas in the left hemisphere: para central temporal middle gyrus (326 Voxels); temporal inferior gyrus (1040 Voxels). In the right hemisphere: temporal middle gyrus (122 Voxels); temporal inferior gyrus (398 Voxels). There were also found GMC increases in the cerebella left hemisphere: cerebellum 1 (32 Voxels); cerebellum 7b (51 Voxels). In the right hemisphere cerebellum 6 (29 Voxels).; On the running group we found many significative alterations when comparing with the health control group. We found increases and diminution on gray matter volume in several areas. The increases or diminution on gray matter volume is compatible with the increases or diminution in the areas where are the presence of more neuronal cell concentration. Differently of what we found in our previous study with judo group, the running group demonstrate an increase and diminution on gray matter volume in many and difuse inespecific areas. That alteration was mostly incompatible between this areas functions and the stimulus of running. The areas where was found increases and diminution of gray matter concentration was: in the region of frontal lobe we found an increase of 14.457 voxels, however this region demonstrated a decrease of 5.720 voxels. The difference between increase and decrease was 8.737 voxels. In the region of temporal lobe and insula was found more diminution then increases in related memory areas Increase in grey matter volume: the increase in grey matter volume was 1.629 voxels, however the decrease was much bigger, 9.580 voxels. The difference between increase and decrease was negative in 7.951 voxels. In the region of parietal lobe was found more increases then diminution in grey matter volume: the increase in grey matter volume was 11.298 voxels, however the decrease was 10.725 voxels. The difference between increase and decrease that represents a gain was 573 voxels. In the region of parietal lobe was found more increases then diminution in grey matter volume: the increase in grey matter volume was 4.241 voxels, however the decrease was 1.047 voxels. The difference between increase and decrease that represents a gain was 3.194 voxels. There were also found increases in the gray matter concentration in the cerebella in motor learning, motor memory and postural control related areas Increase in GMC: the increase in grey matter volume was 10.117 voxels, however the decrease was 7.778 voxels. The difference between increase and decrease that represents a gain was 2.339 voxels. Our findings suggest that motor planning and execution, embedded in sport practice, including judo, can induce plasticity related changes in the brain. These findings can justify the use of sports for rehabilitation of brain injuries such as cerebral palsy or stroke. The physical practice in judo players and long distance runners can increase neural plasticity. The increases of neural plasticity are different between this two sport modalities, suggesting that the change can be specific according to sport modalities. The mechanisms of neural plasticity increases can be related to motor skill as was shown in the judo players group or related to possible effects of tropic factors or hormonal released during prolonged physical exercises as shown in the long distance runners grou
Mestrado
Neurociencias
Mestre em Fisiopatologia Médica

Die vorliegende Arbeit beschäftigt sich mit Mechanismen, die in Einzelzellmodellen zu einer frequenzabhängigen Informationsübertragung führen können. Um dies zu untersuchen, werden Methoden aus der theoretischen Physik (Statistische Physik) und der Informationstheorie angewandt. Die Informationsfilterung in mehreren stochastischen Neuronmodellen, in denen unterschiedliche Mechanismen zur Informationsfilterung führen können, werden numerisch und, falls möglich, analytisch untersucht. Die Bandbreite der betrachteten Modelle erstreckt sich von reduzierten strombasierten ’Integrate-and-Fire’ (IF) Modellen bis zu biophysikalisch realistischeren leitfähigkeitsbasierten Modellen. Anhand numerischer Untersuchungen wird aufgezeigt, dass viele Varianten der IF-Neuronenmodelle vorzugsweise Information über langsame Anteile eines zeitabhängigen Eingangssignals übertragen. Der einfachste Vertreter der oben genannten Klasse der IF-Neuronmodelle wird dahingehend erweitert, dass ein Konzept von neuronalem ’Gedächtnis’, vermittelst positiver Korrelationen zwischen benachbarten Intervallen aufeinander- folgender Spikes, integriert wird. Dieses Model erlaubt eine analytische störungstheoretische Untersuchung der Auswirkungen positiver Korrelationen auf die Informationsfilterung. Um zu untersuchen, wie sich sogenannte ’unterschwelligen Resonanzen’ auf die Signalübertragung auswirken, werden Neuronenmodelle mit verschiedenen Nichtlinearitäten anhand numerischer Computersimulationen analysiert. Abschließend wird die Signalübertragung in einem neuronalen Kaskadensystem, bestehend aus linearen und nichtlinearen Elementen, betrachtet. Neuronale Nichtlinearitäten bewirken eine gegenläufige Abhängigkeit (engl. "trade-off") zwischen qualitativer, d.h. frequenzselektiver, und quantitativer Informations-übertragung, welche in allen von mir untersuchten Modellen diskutiert wird. Diese Arbeit hebt die Gewichtigkeit von Nichtlinearitäten in der neuronalen Informationsfilterung hervor.
Neurons transmit information about time-dependent input signals via highly non-linear responses, so-called action potentials or spikes. This type of information transmission can be frequency-dependent and allows for preferences for certain stimulus components. A single neuron can transmit either slow components (low pass filter), fast components (high pass filter), or intermediate components (band pass filter) of a time-dependent input signal. Using methods developed in theoretical physics (statistical physics) within the framework of information theory, in this thesis, cell-intrinsic mechanisms are being investigated that can lead to frequency selectivity on the level of information transmission. Various stochastic single neuron models are examined numerically and, if tractable analytically. Ranging from simple spiking models to complex conductance-based models with and without nonlinearities, these models include integrator as well as resonator dynamics. First, spectral information filtering characteristics of different types of stochastic current-based integrator neuron models are being studied. Subsequently, the simple deterministic PIF model is being extended with a stochastic spiking rule, leading to positive correlations between successive interspike intervals (ISIs). Thereafter, models are being examined which show subthreshold resonances (so-called resonator models) and their effects on the spectral information filtering characteristics are being investigated. Finally, the spectral information filtering properties of stochastic linearnonlinear cascade neuron models are being researched by employing different static nonlinearities (SNLs). The trade-off between frequency-dependent signal transmission and the total amount of transmitted information will be demonstrated in all models and constitutes a direct consequence of the nonlinear formulation of the models.

Nous considérons les réseaux de neurones en tant que modèles physiques de systèmes naturels de traitement de l'information. Pour avancer dans leur modélisation, nous cherchons à inclure progressivement plus d'éléments physiques des neurones en réseaux, à analyser les propriétés résultantes, et à considérer les conséquences pour le traitement de l'information. Dans le chapitre 1, nous examinons et mettons en cohérence les éléments de base de la modélisation des réseaux de neurones, à différents niveaux de description. Nous nous efforçons aussi d'utiliser des grandeurs physiques munies d'unités et de valeurs réalistes, pour une interprétation qualitative et quantitative des résultats. Puis, nous déduisons de cette démarche un modèle non linéaire de réseau de neurones, qui décrit l'évolution des conductances membranaires gouvernant l'émission des potentiels d'action (PA). Dans le chapitre 2, nous exploitons d'abord ce modèle pour étudier les états d'activité autonomes d'un réseau. Nous montrons que leur apparition peut ici se caractériser comme une transition de phase, qu'ils peuvent se décrire par une théorie de champ moyen, et comment ils peuvent servir à mémoriser de l'information. Nous appliquons ensuite le modèle pour la comparaison des dynamiques neuronales utilisant les PA, avec celles basées directement sur les fréquences de décharge. Nous montrons les limites d'une utilisation de ces dernières à la place des PA. Dans le chapitre 3, nous étudions la résonance stochastique, dans laquelle un bruit devient utile à la transmission de signaux par certains systèmes non linéaires. Nous montrons que c'est une propriété possible pour des neurones naturellement monostables, soumis soit à une stimulation analogique, soit plus généralement à des trains de PA.

The objective of the thesis is to get familiar with the method of psychophysiological interactions and its common inplementation. It is explaining the usual methods of removing disruptive signals from the data processed in correlation analysis and presents the possibility of their implementation. In the practical part it is focused on cerating suggested program and its testing on the real data sets.

This PhD thesis focuses on the development of deep learning based methods for accurate segmentation of the sub-cortical brain structures from MRI. First, we have proposed a 2.5D CNN architecture that combines convolutional and 2/2 spatial features. Second, we proposed a supervised domain adaptation technique to improve the robustness and consistency of deep learning model. Third, an unsupervised domain adaptation method was proposed to eliminate the requirement of manual intervention to train a deep learning model that is robust to differences in the MRI images from multi-centre and multi-scanner datasets. The experimental results for all the proposals demonstrated the effectiveness of our approaches in accurately segmenting the sub-cortical brain structures and has shown state-of-the-art performance on well-known publicly available datasets
Esta tesis doctoral se centra en el desarrollo de métodos basados en el aprendizaje profundo para la segmentación precisa de las estructuras cerebrales subcorticales a partir de la resonancia magnética. En primer lugar, hemos propuesto una arquitectura 2.5D CNN que combina características convolucionales y espaciales. En segundo lugar, hemos propuesto una técnica de adaptación de dominio supervisada para mejorar la robustez y la consistencia del modelo de aprendizaje profundo. En tercer lugar, hemos propuesto un método de adaptación de dominio no supervisado para eliminar el requisito de intervención manual para entrenar un modelo de aprendizaje profundo que sea robusto a las diferencias en las imágenes de la resonancia magnética de los conjuntos de datos multicéntricos y multiescáner. Los resultados experimentales de todas las propuestas demostraron la eficacia de nuestros enfoques para segmentar con precisión las estructuras cerebrales subcorticales y han mostrado un rendimiento de vanguardia en los conocidos conjuntos de datos de acceso público

O dito popular afirma que quando uma pessoa perde um dos sentidos há uma compensação por parte dos outros sentidos para suprir a perda. Através de três experimentos psicofísicos baseados no modelo de Posner (Inibição de Retorno) e técnicas de Ressonância Magnética Funcional, surdos congênitos foram comparados a pessoas ouvintes com o objetivo de verificar se os surdos possuem processos atencionais diferentes dos ouvintes, e se as mesmas áreas corticais como a área de Wernicke, Broca e Córtex auditivo - eram ativadas em ambos os grupos. A tarefa consistia em pressionar um botão todas as vezes que os sujeitos detectassem a presença de um quadrado maior (alvo) apresentado em uma tela, enquanto também eram apresentados quadrados menores (pista) ora do mesmo lado, ora do lado oposto ao alvo. Através do Experimento I se pôde verificar que ambos os grupos apresentaram os fenômenos clássicos do Paradigma de Posner: Facilitação ou Inibição de Retorno, o que denotou a possibilidade de mecanismos atencionais semelhantes para ambos os grupos. Foi observado, porém, que os ouvintes eram mais rápidos que os surdos para responder à tarefa quando o intervalo temporal entre pista e alvo era longo (800ms), comparado ao tempo que levavam para responder quando o intervalo entre pista e alvo era curto (100 ms). O Experimento I suscitou a hipótese de que os surdos possivelmente apresentassem uma diferença de processamento temporal. No Experimento I todas as condições eram apresentadas de forma randômica. O Experimento II foi elaborado com o objetivo de por em evidência a 22 diferença dos TRM para intervalos curtos e longos, portanto os intervalos entre pista e alvo passaram a ser apresentados de forma fixa. Ao comparar os resultados do Experimento I com os do Experimento II (Intervalos Temporais Fixos), se pode verificar que os ouvintes apresentaram Tempos de Reação Manual mais lentos, enquanto os surdos apresentaram as mesmas médias a despeito da vantagem temporal, o que levou a sugerir a hipótese de que os surdos apresentem um déficit no processamento temporal. O experimento III consistiu na utilização do Paradigma de Posner enquanto os sujeitos eram submetidos ao exame de Ressonância Magnética Funcional com o objetivo de investigar se as regiões corticais ativadas poderiam ser semelhantes nos dois grupos. As imagens por Ressonância Magnética Funcional (RMF) demonstraram ativações nas áreas de Wernickie, Broca, e córtex auditivo em ambos os grupos enquanto executavam a tarefa, que embora não possuísse nenhum contexto semântico explícito, possuía o tempo como o principal parâmetro físico no qual os sujeitos pudessem se basear para melhorar o desempenho na tarefa. O tempo é um dos parâmetros físicos primários da língua oral, diferente da língua de sinais que possui o parâmetro visual e espacial como primário. Os resultados sugerem que as ativações corticais nos centros de audição e fala podem indicar uma plasticidade cruzada no grupo de surdos. Ainda, a participação do córtex auditivo no processamento da elaboração de estratégias para responder a uma tarefa que não contenha um contexto semântico explicito possivelmente indica sua participação no processamento de linguagem.
It is popularly said that when a person loses one sense, there is a compensation by the other remaining senses to suppress the loss. Throughout three Phsycophysic Experiments based on Inhibition of Return Posners Paradigm and Functional Magnetic Resonance (fMRI) Techniques, congenital deaf people were compared to normal hearing people in order to check if deaf people possess different attentional pattern compared to normal hearing people, and if the same cortical areas Wernicke and Brocas area and Hearing Cortex were activated in both groups. Experiment I consisted on pressing a button every time the presence of a big square (target) was detected by subjects while non-predictive small squares (cue) were also presented at the same or opposite side of the target. At Experiment I it was observed that both groups presented Posners Paradigm classical phenomena: Facilitation or Inhibition of Return, what suggested the possibility that attentional pattern may be similar to both groups. Therefore, it was observed that normal hearing people were faster than deaf people to respond to the task when time interval between cue and target was long (800 ms) when compared to the time they spent to respond when time interval between cue and target as short (100 ms). 24 Experiment I raised the hypotheses that possibly deaf people may present a temporal processing difference. At Experiment I every condition was randomly presented. Experiment II was elaborated to highlight MRT differences between short and long time intervals, so every time interval was presented on a fixed order. Comparison of Experiment I and II (Fixed Time Intervals) showed that normal hearing people presented shorter Manual Reaction Times (MRT), while deaf people kept the same averages despite the temporal advantage, what suggested that deaf people may present a deficit on temporal processing. Experiment III used Posners Paradigm while subjects were submitted to fMRI scanning in order to check if activated cortical regions could be similar in both groups. fMRI images demonstrate Wernicke and Brocas area and hearing cortex activations in both groups while executing the task, which, although did not have any explicit semantic content, had time as the main physical parameter on which subjects could be based to increase performance to respond to the task. Time is one of the oral language primary physical parameter, different of signed language which has visual and spatial parameters as primaries. Results suggest that cortical audition center activations may indicate a cross-modal plasticity at the deaf group. Yet, participation of hearing cortex on strategy elaboration to respond to a task which does not have any explicit semantic content possibly indicates the participation of hearing cortex on language processing.

Real neurons, and their networks, are far too complex to be described exactly by simple deterministic equations. Any description of their dynamics must therefore incorporate noise to some degree. It is my thesis that the nervous system is organized in such a way that its performance is optimal, subject to this constraint. I further contend that neuronal dynamics may even be enhanced by noise, when compared with their deterministic counter-parts. To support my thesis I will present and analyze three case studies. I will show how noise might (i) extend the dynamic range of mammalian cold-receptors and other cells that exhibit a temperature-dependent discharge; (ii) feature in the perception of ambiguous figures such as the Necker cube; (iii) alter the discharge pattern of single cells.

The purpose of this thesis is threefold: 1) A behavioral examination of the role of diacritics in Arabic, 2) A functional magnetic resonance imaging (fMRI) investigative study of diacritics in Arabic, and 3) An optimization algorithm for brain mapping and detecting brain function. Firstly, the role of diacritics in Arabic was examined behaviorally. The stimulus was a lexical decision task (LDT) that constituted of low, mid, and high frequency words and nonwords; with and without diacritics. Results showed that the presence of vowel diacritics slowed reaction time but did not affect word recognition accuracy. The longer reaction times for words with diacritics versus without diacritics suggest that the diacritics may contribute to differences in word recognition strategies. Secondly, an Event-related fMRI experiment of lexical decisions associated with real words with versus without diacritics in Arabic readers was done. Real words with no diacritics yielded shorter response times and stronger activation than with real words with diacritics in the hippocampus and middle temporal gyrus possibly reflecting a search from among multiple meanings associated with these words in a semantic store. In contrast, real words with diacritics had longer response times than real words without diacritics and activated the insula and frontal areas suggestive of phonological and semantic mediation in lexical retrieval. Both the behavioral and fMRI results in this study appear to support a role for diacritics in reading in Arabic. The third research work in this thesis is an optimization algorithm for fMRI data analysis. Current data-driven approaches for fMRI data analysis, such as independent component analysis (ICA), rely on algorithms that may have low computational expense, but are much more prone to suboptimal results. In this work, a genetic algorithm (GA) based on a clustering technique was designed, developed, and implemented for fMRI ICA data analysis. Results for the algorithm, GAICA, showed that although it might be computationally expensive; it provides global optimum convergence and results. Therefore, GAICA can be used as a complimentary or supplementary technique for brain mapping and detecting brain function by way of fMRI.

O traumatismo craniencefálico (TCE) fechado é, classicamente, um modelo de lesão neuronal e axonal monofásica, onde a destruição do parênquima, incluindo neurônios e células gliais, ocorre principalmente no momento do trauma, seguida pela degeneração Walleriana anterógrada e retrógrada nos dias subseqüentes. Há evidências de progressão da perda neuronal e axonal na fase tardia após o trauma, observada principalmente pela evolução da atrofia cerebral, secundária a vários fatores, incluindo a apoptose neuronal. Com o objetivo de testar a hipótese de que as técnicas quantitativas em ressonância magnética (RM) permitem identificar, de modo não invasivo, as variáveis biológicas que estimam a perda neuronal e axonal no cérebro relacionadas ao TCE moderado e grave e à lesão axonal difusa, na fase tardia, foram avaliados 9 pacientes, sendo 5 do sexo masculino e 4 do sexo feminino, com idades variando de 11 a 28 anos (média de 21,1 anos), que foram vítimas de TCE moderado ou grave (Escala de Coma de Glasgow menor que 12 na admissão hospitalar após o TCE) e que tiveram boa recuperação. O tempo médio entre o trauma e o exame de ressonância magnética foi de 3,1 anos (± 0,5 anos). Foram utilizados os índices ventrículo-cerebrais bifrontal (IVCF) e bicaudado (IVCC), a medida do tempo de relaxação T2, o índice de transferência de magnetização (MTR), o coeficiente de difusão aparente (ADC) e a espectroscopia de prótons multi-voxel, com o cálculo dos índices metabólicos N-acetilaspartato/creatina (NAA/Cre) e colina/creatina (Cho/Cre). Foram estudados a substância branca (SB) frontal e parietal bilateralmente, o joelho e o esplênio do corpo caloso (CC) e a substância cinzenta (SC). As médias dos valores medidos foram comparadas às de um grupo controle formado por 9 pessoas sadias, pareadas pela idade e sexo. Observou-se aumento estatisticamente significativo (p ≤ 0,05) do IVCF e do IVCC nos pacientes, devido ao aumento ventricular secundário à atrofia subcortical; aumento no tempo de relaxação T2 na SB e no CC, que reflete o aumento da concentração de água por provável perda axonal e gliose; aumento do ADC e redução do MTR na SB e no CC, que demonstram lesão das fibras axonais mielinizadas, e redução do índice NAA/Cre no CC, indicando perda axonal. Não houve diferença estatisticamente significativa nas medidas realizadas na SC e nem no índice Cho/Cre (p › 0,05). Os resultados encontrados mostram que as técnicas quantitativas em RM foram capazes de detectar, de modo não invasivo, o dano neuronal e axonal na substância branca e no corpo caloso de cérebros humanos, secundário ao TCE moderado e grave.
Closed traumatic brain injury (TBI) is a classic model of monophasic neuronal and axonal injury, where tissue damage mainly occurs at the moment of trauma, followed by anterograde and retrograde Wallerian degeneration in the subsequent days. There are some evidences of delayed progression of the neuronal and axonal loss after TBI, mainly shown by gradual development of cerebral atrophy, due to many factors, including neuronal apoptosis. For the purpose of testing the hypothesis that quantitative magnetic resonance techniques are able to assess the biological variables which estimate neuronal and axonal loss in brain, related to moderate or severe TBI and diffuse axonal injury, nine patients (age range 11 – 28 years; mean age 21,1 years; 5 male and 4 female), who sustained a moderate or severe TBI (initial Glasgow Coma Scale less than 12), with good recovery, were evaluated in a mean of 3,1 years after trauma (± 0,5 year). The following techniques were applied: bicaudate (CVIC) and bifrontal (CVIF) cerebroventricular indexes; T2 relaxation time measurement (T2 relaxometry); magnetization transfer ratio (MTR); apparent diffusion coefficient (ADC); multivoxel proton magnetic resonance spectroscopy, using N-acetylaspartate/creatine (NAA/Cre) and choline/creatine (Cho/Cre) ratios; measured in the frontal and parietal white matter (WM) of both cerebral hemispheres, in the genu and splenium of the corpus callosum (CC) and in the gray matter (GM). The results were compared with those of a control group constituted by 9 healthy volunteers with a matched age and sex distribution. The CVIC and CVIF mean values were significantly increased (p ≤ 0,05) in patients due to ventricular enlargement secondary to subcortical atrophy; an increase in T2 relaxation time was observed in the WM and CC, reflecting an enhancement in water concentration, probably secondary to axonal loss and gliosis; increased ADC mean values and reduced MTR mean values were found in the WM and CC, showing damage in the myelinated axonal fibers; and decreased NAA/Cre ratio mean values in the CC, indicating axonal loss. No significant differences were observed in the mean values measured at the GM or in the Cho/Cre ratio mean values (p › 0,05). These quantitative magnetic resonance techniques were able to non-invasively demonstrate the neuronal and axonal damage in the WM and CC of human brains, secondary to moderate or severe TBI.

Abstract The purpose of the present study was to gain more insight into the blood oxygen level-dependent (BOLD)-contrast functional MRI (fMRI) in the brain and its connection to EEG, both in global and local scales of their temporal and spatial relations. BOLD signal changes were studied during hyperventilation (HV) induced EEG reactivity of intermittent rhythmic delta activity (IRDA). The BOLD signal in gray matter decreased 30% more in subjects with IRDA (N = 4) than in controls (N = 4), during the first two minutes of HV. This difference disappeared during IRDA in EEG. BOLD signal changes may provide additional information about dynamic hemodynamic changes relative to HV induced EEG reactivity. BOLD signal changes were investigated during sudden deepening of thiopental anesthesia into EEG burst-suppression level in pigs (N = 5). Positive (6–8%) or negative (-3– -8%) group average BOLD signal changes correlated to the thiopental bolus injection were seen. Positive and negative responses covered 1.6% and 2.3% of the brain voxels, respectively. BOLD signal changes in brain are associated with sudden deepening of thiopental anesthesia into EEG burst-suppression level, but they are spatially inconsistent and scarce. Somatosensory BOLD response was studied in brain before and after globally induced methotrexate (MTX) exposition in pigs (N = 4). After the MTX exposure, reduced (from 2–4% to 0–1%) or negative (-2% to -3%) BOLD responses were detected. Somatosensory BOLD-contrast response shows a slight difference in brain before and after globally induced MTX exposition. An experimental epilepsy model for development of simultaneous EEG and BOLD-contrast fMRI in the localization of epilepsy was developed and tested. Dynamic penicillin induced local epilepsy was applied in deep isoflurane anesthesia in pigs (N = 6). Relatively high (10–20%) and localized BOLD signal increase was found. The dynamic penicillin induced focal epilepsy model in deep isoflurane anesthesia with simultaneous EEG and BOLD-contrast fMRI is feasible for the development of these methods for localization of epileptic focus or foci. In conclusion, with careful experimental design and analysis, BOLD-contrast fMRI with EEG provides a potential tool for monitoring and localising functional changes in the brain.

Background: In recent years it has become recognized that schizophrenia is associated with cognitive impairment, which affects particularly executive function and memory. Although cognitive impairment implies brain damage or dysfunction, little is known about the relationship of cognitive impairment in schizophrenia to the structural and functional brain abnormalities that characterize the disorder. Aims and hypothesis: This study aimed to identify structural and/or functional brain abnormalities associated with cognitive impairment in schizophrenia. According to the literature reviewed, the general hypothesis is that the cognitive deficits of schizophrenic patients are reflected in both structural and functional brain changes. Accordingly, we expect that patients with cognitive impairment will have more GM reductions and more dysfunctional patterns of brain activity than patents without such deficits. Method. structural MRI and voxel-based morphometry was carried out in 26 cognitively impaired and 23 cognitively preserved schizophrenia patients, plus 39 matched controls. Nineteen cognitively impaired patients, 18 cognitively preserved patients and 34 controls also underwent fMRI during performance of a working memory task. Some subjects, partly corresponding to the ones participating in the neuroimaging study, were also administered a battery of tests of different cognitive functions. Altogether, 25 participants with schizophrenia and relatively preserved cognitive function, 29 with impaired functions and 22 healthy participants were included in this part of the study. Results: The participants with cognitive impairment in executive functioning and memory also perform worse on visuospatial and language tasks when compared to other patients intact in memory and executive function and controls. These differences are independent of estimated premorbid IQ. No differences were found between the cognitively intact and cognitively impaired groups in lateral ventricular volume or whole brain volume. Voxel-based morphometry also failed to reveal clusters of significant difference in either GM or WM volume between the two patient groups. However, during performance of the n-back task, the cases with schizophrenia and impaired cognitition showed hypoactivation compared to the cognitively intact patients in DLPFC among other brain regions. Conclusions: This study provides evidence that structural brain abnormality in schizophrenia is a function of having the disorder, not the cognitive impairment that accompanies it. In contrast, a substantial part of the task-related functional imaging abnormality appears to reflect cognitive impairment.
MARCO: Los déficits cognitivos son una característica conocida de la esquizofrenia. Sin embargo, poco se sabe sobre su relación con las anormalidades cerebrales tanto estructurales como funcionales que caracterizan dicha enfermedad. OBJETIVOS: Identificar las alteraciones estructurales y/o funcionales asociadas al déficit cognitivo en la esquizofrenia. MÉTODO: Se adquirió imágenes de resonancia magnética funcional (RMf) y morfometría basada en el vóxel (VBM) en 26 participantes que tenían alteraciones cognitivas y 23 que presentaban una cognición relativamente preservada, todos con esquizofrenia, así como en 39 controles apareados. Diecinueve de quienes presentaban déficits cognitivos y 18 de los preservados cognitivamente y 24 controles también ejecutaron una tarea n-back de memoria de trabajo durante la adquisición de imágenes de RMf. RESULTADOS: No se encontró diferencias entre los participantes con cognición preservada y quienes presentaban déficits cognitivos en el volumen de los ventrículos laterales ni tampoco en el volumen cerebral total. La VBM tampoco mostró clústers con diferencias significativas entre ambos grupos en el volumen de sustancia blanca y gris. Sin embargo, durante la realización de la tarea n-back de memoria de trabajo, los participantes con alteración cognitiva presentaron hipoactivación en la corteza prefrontal dorsolateral, entre otras regiones, en relación a quienes mostraban una cognición preservada. CONCLUSIONES: No se encontró evidencia de que los déficits cognitivos de la esquizofrenia sean una función de anormalidades cerebrales estructurales que acompañan a la enfermedad sino que se asociarían con un funcionamiento cerebral alterado.

Aquesta tesi es planteja com a objectiu principal identificar, en pacients amb trastorn depressiu major (TDM), característiques neuroanatòmiques, tant a nivell neuroestructural com neurofuncional, que es puguin relacionar amb variables clíniques, d'evolució i de resposta a tractament i per tant, constituir fenotips de neuroimatge. Amb aquesta finalitat hem estudiat la relació entre les dades obtingudes mitjançant ressonància magnètica (RM) i variables com el subtipus depressiu (depressió amb malenconia i depressió comòrbida a trastorn obsessiu-compulsiu (TOC)), el nombre de recaigudes, el temps fins a la remissió o la resposta a tractament. Aquesta tesis doctoral consta de tres estudis principals: PRIMER ESTUDI: Es va estudiar un grup de pacients amb TDM que apareixia en el decurs d'un TOC, comparant-los amb un grup de pacients amb TOC pur. Es va deteminar que els primers presentaven una major reducció de l'escorça orbito-frontal medial (EOFm) i un patró de correlacions amb altres estructures que els diferenciava clarament dels subjectes amb TOC pur. Publicació: N Cardoner, C Soriano-Mas, J Pujol, P Alonso, B Harrison, J Deus, R Hernández-Ribas, JM Menchón, J Vallejo. Brain structural correlates of depressive comorbidity in obsessive-compulsive disorder. Neuroimage 2007; 38: 413-421 Factor d’impacte = 5,937. JCR 2010 = 1r quartil, 1r decil, número 1/14 (Neuroimatge). SEGON ESTUDI: Es va estudiar de forma transversal un grup de 70 pacients amb depressió malenconiosa, comparant-los amb una mostra de controls sans. Es van trobar reduccions de substància gris (SG) en l'ínsula esquerra i increment de la substància blanca (SB) en l'àrea troncoencefálica superior. Posteriorment es va realitzar el seguiment longitudinal d'aquests pacients, obtenint-se, en 30 dels 70 pacients inclosos en l'estudi transversal, una segona RM al cap de 7 anys. La SB de l'àrea troncoencefálica superior va mostrar una evolució temporal diferencial entre els pacients depressius i els controls sans, mostrant una disminució de volum en els primers i un increment en els segons. El decrement de SG en ambdues ínsules es va relacionar amb el nombre de recaigudes durant el seguiment. Publicació: C Soriano-Mas, R Hernández-Ribas, J Pujol, M Urretavizcaya, J Deus, B Harrison, H Ortiz, M López-Solà, JM Menchón, N Cardoner. Cross-Sectional and Longitudinal Assessment of Structural Brain Alterations in Melancholic Depression. Biological Psychiatry 2011; 69:318-325 Factor d’impacte = 8,674. JCR 2010 = 1r quartil, 1r decil, número 4/128 (Psiquiatria). TERCER ESTUDI: Es va estudiar mitjançant RM funcional (RMf) una mostra de pacients amb TDM amb la finalitat de determinar els correlats neurofuncionals de la resposta al tractament actiu mitjançant estimulació magnètica transcranial repetitiva (EMTr), comparant-lo amb un grup de subjectes que rebien EMTr placebo. Es va relacionar la resposta clínica a EMTr amb l'activitat en l'escorça cingulada anterior (CCA), l’EOFm esquerra, l'escorça frontal medial (EFm) dreta i el putamen ventro-caudal esquerre. Publicació: R Hernández-Ribas, J Deus, J Pujol, C Segalàs, J Vallejo, JM Menchón, N Cardoner, C Soriano-Mas. Identifying Brain Imaging Correlates of Clinical Response to Repetitive Transcranial Magnetic Stimulation (rTMS) in Major Depression. Brain Stimulation (en premsa) Factor de impacto = 4,964. JCR 2010 = 1r quartil, número 41/239 (Neurociències). VALORACIÓ GLOBAL: En concordança amb les hipòtesis i els objectius del present treball, s’han descrit un conjunt d’alteracions neuroestructurals i neurofuncionals que es poden considerar correlats neurals d’aspectes clínics com el subtipus depressiu, d’aspectes evolutius com el nombre de recurrències clíniques, o d’aspectes relacionats amb la resposta al tractament antidepressiu. Aquests correlats es podrien considerar, segons el nostre criteri, com a fenotips de neuroimatge, ja que s’associen a característiques concretes del trastorn depressiu major i ens permeten diferenciar entre grups/subgrups de pacients.
MAGNETIC RESONANCE IMAGING STUDY OF NEURAL BASIS OF MAJOR DEPRESSIVE DISORDER: FOCUS ON SYMPTOMS, LONGITUDINAL COURSE AND RESPONSE TO TREATMENT SUMMARY: Main objective: To identify brain imaging correlates of clinical, outcome and treatment response variables in patients with major depressive disorder (MDD). FIRST STUDY 72 obsessive-compulsive patients (OCD) with (n=33) or without (n=39) MDD, and 72 control subjects (CS) were assessed. OCD with MDD showed a larger reduction of medial orbitofrontal cortex (mOFC) gray matter (GM) volume. Both OCD groups showed distinct correlations of mOFC GM volume with other brain regions. Publication: N Cardoner, C Soriano-Mas, J Pujol, P Alonso, B Harrison, J Deus, R Hernández-Ribas, JM Menchón, J Vallejo. Brain structural correlates of depressive comorbidity in obsessive-compulsive disorder. Neuroimage 2007; 38: 413-421 SECOND STUDY Whole-brain magnetic resonance images (MRI) were collected from 70 melancholic inpatients and 40 CS. A subset of 30 patients and 20 CS were rescanned after a 7-year period. In cross-sectional assessment, melancholic patients showed GM reductions in the left insula and white matter (WM) increases in the upper brainstem tegmentum. In longitudinal assessment, WM of the upper brainstem tegmentum showed a different temporal evolution between groups. Bilateral GM insulae reductions were associated with the number of relapses. Publication: C Soriano-Mas, R Hernández-Ribas, J Pujol, M Urretavizcaya, J Deus, B Harrison, H Ortiz, M López-Solà, JM Menchón, N Cardoner. Cross-Sectional and Longitudinal Assessment of Structural Brain Alterations in Melancholic Depression. Biological Psychiatry 2011; 69:318-325 THIRD STUDY Twenty-one MDD patients received active or sham repetitive Transcranial Magnetic Stimulation (rTMS) on the left dorsolateral prefrontal cortex. In the active rTMS group, clinical response was correlated to smaller pre-treatment deactivations in the anterior cingulate, the left medial orbitofrontal and the right middle frontal cortices, and to larger activations in the left ventral-caudal putamen. Publication: R Hernández-Ribas, J Deus, J Pujol, C Segalàs, J Vallejo, JM Menchón, N Cardoner, C Soriano-Mas. Identifying Brain Imaging Correlates of Clinical Response to Repetitive Transcranial Magnetic Stimulation (rTMS) in Major Depression. Brain Stimulation (in press) COMMENT In our research, we found brain imaging changes related to variables such as depressive subtype, number of clinical relapses or response to treatment. These imaging correlates could be considered as neuroimaging phenotypes.

The quantification of cerebral atrophy is fundamental in neuroinformatics since it permits diagnosing brain diseases, assessing their progression, and determining the effectiveness of novel treatments to counteract them. However, this is still an open and challenging problem since the performance 2/2 of traditional methods depends on imaging protocols and quality, data harmonisation errors, and brain abnormalities. In this doctoral thesis, we question whether deep learning methods can be used for better estimating cerebral atrophy from magnetic resonance images. Our work shows that deep learning can lead to a state-of-the-art performance in cross-sectional assessments and compete and surpass traditional longitudinal atrophy quantification methods. We believe that the proposed cross-sectional and longitudinal methods can be beneficial for the research and clinical community
La cuantificación de la atrofia cerebral es fundamental en la neuroinformática ya que permite diagnosticar enfermedades cerebrales, evaluar su progresión y determinar la eficacia de los nuevos tratamientos para contrarrestarlas. Sin embargo, éste sigue siendo un problema abierto y difícil, ya que el rendimiento de los métodos tradicionales depende de los protocolos y la calidad de las imágenes, los errores de armonización de los datos y las anomalías del cerebro. En esta tesis doctoral, cuestionamos si los métodos de aprendizaje profundo pueden ser utilizados para estimar mejor la atrofia cerebral a partir de imágenes de resonancia magnética. Nuestro trabajo muestra que el aprendizaje profundo puede conducir a un rendimiento de vanguardia en las evaluaciones transversales y competir y superar los métodos tradicionales de cuantificación de la atrofia longitudinal. Creemos que los métodos transversales y longitudinales propuestos pueden ser beneficiosos para la comunidad investigadora y clínica

INTRODUÇÃO: Alterações em circuito neural têm sido associadas com o transtorno depressivo maior. Entretanto, até o momento não se tem estudos investigando a associação entre a lesão do acidente vascular cerebral neste circuito e a incidência do episódio depressivo maior após o acidente vascular cerebral. Este estudo teve como objetivo principal investigar de modo prospectivo a associação entre o volume da lesão no circuito neural límbico-córtico-estriado-pálido-talâmico no hemisfério esquerdo e a incidência de episódio depressivo maior nos primeiros quatro meses posteriores ao acidente vascular cerebral isquêmico. Como objetivo secundário, visou investigar a associação entre o volume da lesão em regiões específicas do circuito e a incidência do episódio depressivo maior após o acidente vascular cerebral isquêmico. MÉTODOS: Neste estudo foram triados de modo consecutivo 326 pacientes admitidos na enfermaria de neurologia do Hospital das Clinicas de São Paulo. Destes, foram elegíveis 68 pacientes e foram acompanhados prospectivamente. A avaliação psiquiátrica consistiu na aplicação da entrevista clinica estruturada para diagnóstico pelo DSM-IV e no manual estruturado para entrevista da Escala de Hamilton para Depressão; o grau de comprometimento nas atividades de vida diária foi medido pelo Índice de Barthel; o grau de gravidade do acidente vascular cerebral foi mensurado pela escala para acidente vascular cerebral do National Institutes of Health e, a capacidade cognitiva foi avaliada pelo Miniexame do estado mental. As avaliações ocorreram em 3 momentos sendo a primeira em média 12,4 dias (+ 4,5) após o acidente vascular cerebral, a segunda em média 37 dias (+ 6) e, a terceira em média e 91,6 dias (+ 5,4) após o acidente vascular cerebral. As imagens por ressonância magnética foram realizadas dentro dos 15 dias posteriores ao acidente vascular cerebral em um aparelho de 1.5 Tesla (GE-Horizon LX) com protocolo específico. A localização do acidente vascular cerebral e o volume da lesão foram processados usando um método semi-automatizado baseado no Atlas Citoarquitetônico de Brodmann. Os grupos de pacientes com episódio depressivo maior e sem depressão foram comparados quanto aos dados clínicos e neurorradiológicos. RESULTADOS: Nos primeiros 4 meses posteriores ao acidente vascular cerebral 21 (31%) pacientes apresentaram episódio depressivo maior. Não houve diferenças entre os grupos com respeito à idade, distribuição por sexo, estado civil, condição empregatícia, lateralização hemisférica da lesão isquêmica, gravidade do infarto, comprometimento das atividades de vida diária e capacidade cognitiva. O volume da lesão no circuito límbico-córtico-estriado-pálido-talâmico no hemisfério esquerdo apresentou associação com a incidência do episódio depressivo maior (P=0,004) assim como, volume da lesão nas seguintes áreas específicas: córtex ventral anterior do cíngulo (AB24; P=0,032), córtex dorsal anterior do cíngulo (AB32; P=0,043), córtex subgenual (AB25; P=0,038), subículo (AB28/AB36; P=0,032) e amígdala (AB34; P=0,010). Este estudo tem algumas limitações como não avaliação de lesões isquêmicas subcorticais e outros possíveis fatores de risco para a depressão posterior ao acidente vascular cerebral. CONCLUSÕES: Os resultados fornecem evidências da existência de associação entre o volume da lesão do acidente vascular cerebral no hemisfério esquerdo no circuito límbico-córticoestriado- pálido-talâmico e a incidência do episódio depressivo maior posterior ao acidente vascular cerebral.
BACKGROUND: Dysfunction in the neural circuit has been etiologically related to major depressive disorder. However no study has investigated the role of lesion in these circuits and post-stroke major depression. The objective of this study was to prospectively investigate the association between stroke volume in left limbiccortical- striatal-pallidal-thalamic neural circuit and incidence of major depressive episode after stroke, and secondary to investigate the association between stroke volume in specific areas of the neural circuit and the incidence of major depressive episode after stroke. METHODS: From 326 consecutively screened patients admitted in the neuroclinical unit of Clinics Hospital, São Paulo, 68 were eligible and followed. The Structured Clinical Interview for DSM-IV and Hamilton Depression Scale were applied in the psychiatry evaluations. The stroke severity was evaluated using the National Institutes of Health Stroke Scale and the activities of daily living limitations were measured using Barthel Index. Cognitive capacity was measured using Mini Mental State Examination. The evaluations were done in three timepoints the first in mean of 12.4 (+ 4.5) days after stroke, the second in 37 (+ 6) days and, the third, 91.6 (+ 5.4) days after stroke. Magnetic resonance scans were performed within 2 weeks after stroke in a 1.5 Tesla (GE-Horizon LX) scanner. Stroke localization and volume quantification were performed using a semi-automated method based on the Brodmann Cytoarchitectonic Atlas. The depressed and non depressed patients were compared. RESULTS: Twenty-one patients (31%) experienced a new onset of major depressive episode within a four-month period after stroke. No differences were found between depressed and non depressed patients regarding age, gender distribution, marital status, employment status, ischemic lesion hemispheric lateralization, stroke severity, level of limitations in activities of daily living and cognitive capacity. Stroke lesion volume in the left limbic-cortical-striatal-pallidal-thalamic circuit was associated with incidence of major depressive episode (P=0.004). Complementary analyses revealed association the incidence of major depressive episode with larger stroke lesion volume in specific areas of this circuit, including the ventral anterior cingulate cortex (BA24; P=0.032), dorsal anterior cingulate cortex (BA32; P=0.043), subgenual cortex (BA25; P=0.038), hippocampal subiculum (BA28/BA36; P=0.032) and amygdala (BA34; P=0.010). There are some study limitations, as no subcortical ischemic stroke analyses and other possible risk factors for depression after stroke. CONCLUSIONS: The results of this study have shown evidence of an association between stroke lesion size in left hemisphere in the limbic-cortical-striatal-pallidalthalamic circuit and the incidence of major depressive episode after stroke.

Garantir a qualidade da carne bovina produzida no Brasil tem sido uma preocupação dos produtores, pois contribui para aumentar a exportação e o consumo interno do produto. Por isso, tem-se pesquisado novos métodos que analisam e garantam a qualidade da carne, de forma rápida, eficiente e não destrutiva. A ressonância magnética nuclear (RMN) tem se destacado como uma das técnicas de controle de qualidade de carne. Neste trabalho as redes neurais artificiais estão sendo utilizadas para o reconhecimento de padrões dos dados de ressonância magnética nuclear oriundos de carne bovina. Mais especificamente, os respectivos dados têm sido utilizados por uma rede perceptron multicamadas para a extração de características da carne bovina, possibilitando a classificação do grupo genético e do sexo dos animais a partir de uma amostra da referida carne. Os resultados dos experimentos são também apresentados para ilustrar o desempenho da abordagem proposta.
Guaranteeing the quality of the bovine meat produced in Brazil has been a concern of the producers because it contributes to increase the export and the domestic consumption of the product. Therefore, new methods have been researched that analyze and guarantee the quality of the meat in a fast, efficient and non destructive way. Nuclear magnetic resonance (NMR) has been highlighted as one of the techniques of meat quality control. In this work study artificial neural networks are being used for pattern recognition from data obtained by the resonance equipment, originating from bovine meat. More specifically, the respective data have been used by a multilayer perceptron network for extraction of bovine meat characteristics, making possible the classification of both genetic group and animal sex starting from a single meat sample. Several results of experimental tests are also presented to illustrate the performance of the proposed approach.

Racional: A epilepsia do lobo temporal está comumente associada à farmacorresistência e tem a esclerose hipocampal como achado neuropatológico em mais da metade dos casos. Histologicamente, a esclerose hipocampal está associada à perda neuronal diferencial e gliose, além de alterações nos níveis de moléculas associadas à homeostase da água tecidual, como a aquaporina 4 e a molécula de matriz sulfato de condroitina. Em imagens de ressonância nuclear magnética, a esclerose é caracterizada por redução de volume em sequências ponderadas em T1, aumento de sinal e tempo de relaxamento em sequências ponderadas em T2 e redução na transferência de magnetização. Justificativa e Objetivos: Uma vez que tanto o sinal T2 quando a transferência de magnetização são dependentes da água tecidual, nosso objetivo é avaliar, na formação hipocampal de pacientes com epilepsia do lobo temporal, as correlações entre populações celulares e moléculas ligadas à homeostase da água e as imagens ponderadas em T2 e transferência de magnetização. Visamos ainda definir, na formação hipocampal de indivíduos sem alterações neuropatológicas, o volume de cada um dos subcampos hipocampais. Metodologia: Pacientes com epilepsia do lobo temporal farmacorresistente (ELT, n = 43), bem como voluntários sadios (controle radiológico, CH, n = 20), foram submetidos a exames de ressonância magnética em máquina de 3T para mensuração da volumetria hipocampal, tempo de relaxamento T2 e transferência de magnetização hipocampal (exames in vivo). Após o tratamento cirúrgico para o controle das crises, os hipocampos dos pacientes com ELT foram fixados por 8 dias e submetidos aos exames ex vivo em máquina de 3T para cálculo do tempo de relaxamento T2 de cada subcampo hipocampal. Hipocampos controle (Controle historadiológico, CHR, n = 14), foram obtidos de autópsias de pacientes sem histórico ante-mortem de doença neurológica ou presença de patologia no exame do encéfalo pos mortem. Ambos os grupos controle foram pareados para idade em relação ao grupo ELT. Alguns dos casos CHR (n = 6) foram também submetidos à imagem 3D T2 em máquina de 4,7T para cálculo de volumetria dos subcampos hipocampais. Após emblocamento em parafina, secções coronais hipocampais dos casos CHR e ELT foram submetidas às técnicas de histoquímica básica Hematoxilina e Eosina e Luxol Fast Blue, e às imuno-histoquímicas para avaliação das populações neuronais (NeuN), astrócitos reativos (GFAP), micróglias ativadas (HLA-DR) e para a expressão de aquaporina 4 (AQP4) e níveis de sulfato de condroitina (CS-56). Para a comparação entre os grupos, foram realizados testes t para dados paramétricos e Mann-Whitney para dados não-paramétricos. Testes de correlação foram empregados para análise da associação entre as avaliações histológicas e os exames de ressonância magnética. Resultados: Pacientes com ELT apresentaram menor volume hipocampal, maior tempo de relaxamento T2 e menor transferência de magnetização no exame in vivo, quando comparados com o CR. O exame ex vivo para a volumetria dos subcampos hipocampais em casos do grupo CHR indicou que a fascia dentata, a região CA1 e o subículo correspondem à 85 % do volume hipocampal total. Quanto ao tempo de relaxamento T2 ex vivo, foi observado aumento em todos os subcampos hipocampais do grupo ELT, à exceção da fascia dentata, quando comparados ao CHR. A avaliação da densidade neuronal indicou redução significativa em todos os subcampos dos casos ELT, à exceção do subículo, quando comparados ao CHR. Em relação aos valores do grupo CHR, foi observada astrogliose em quase todos subcampos da formação hipocampal (a exceção da zona subgranular e do hilo) e microgliose em todos os subcampos (exceto pelo subículo) dos casos com ELT. Pacientes com ELT apresentaram redução na expressão de aquaporina 4 perivascular em todos os subcampos do hipocampo, comparados ao CHR. Aumento nos níveis de sulfato de condroitina foi observado em todos os subcampos da formação hipocampal, à exceção da camada granular, nos pacientes com ELT. O volume hipocampal e a transferência de magnetização in vivo dos pacientes com ELT correlacionaram-se tanto com a população neuronal como com os níveis de sulfato de condroitina, enquanto que o tempo de relaxamento in vivo correlacionou-se com a população astroglial e os níveis de sulfato de condroitina. O exame ex vivo corroborou a correlação entre a população glial e o tempo de relaxamento observado nos pacientes com ELT. A diferença entre o tempo de relaxamento in vivo e ex vivo correlacionou-se tanto com a difusibilidade da água no tecido como com os níveis de sulfato de condroitina. Conclusões: Nossos dados indicam correlação entre a patologia hipocampal e as imagens de ressonância nuclear magnética, sendo que a maior qualidade das imagens ex vivo permitiu uma avaliação mais direta entre o sinal de ressonância e a patologia, indicando importância da população celular e matriz extracelular para o volume hipocampal e a transferência de magnetização, e da astrogliose para o tempo de relaxamento T2. Finalmente, nossos dados mostraram que CA1, subículo e fascia dentata tem grande participação no volume hipocampal, sendo que alterações nestas regiões tem um papel mais relevante nas alterações observadas na ressonância magnética, como indicado por nossas correlações.
Rationale: Drug resistant temporal lobe epilepsy is often associated with hippocampal sclerosis. Histological evaluation reveals differential neuronal loss, gliosis and changes in molecules associated with water homeostasis, such as aquaporin 4 and chondroitin sulfate. Magnetic resonance imaging in these cases often reveals hippocampal atrophy, increased T2 signal and T2 relaxation and reduced magnetization transfer ratio in the hippocampus. Aims: Once both T2 signal and magnetization transfer are affected by tissue water, our goal was to evaluate, in the hippocampus of drug-resistant temporal lobe epilepsy patients who underwent surgery for seizure control, the associations between cellular populations, aquaporin 4 and chondroitin sulfate with T2 relaxation time and magnetization transfer. Additionally, we intended to measure the individual volume of each hippocampal subfield in hippocampus from patients without neurological disease. Methods: Patients with drug-resistant temporal lobe epilepsy (TLE, n = 43) and age-matched health volunteers (radiological control, RC, n = 20) were submitted to magnetic resonance in a 3T machine for hippocampal volumetry measure, T2 relaxation and magnetization transfer (in vivo examination). After surgical treatment for seizure control, hippocampi from the TLE patients were fixed in formalin for 8 days and then submitted to ex vivo imaging in 3T for relaxation time of every hippocampal subfield. Control hippocampi were obtained from autopsies of age-matched patients without ante mortem history of neurological disease or post mortem neurological pathology, and underwent the same ex vivo imaging (histo-radiological control, HRC, n = 14). Six cases from the HRC underwent 3D T2 imaging in a 4.7T machine, in order to measure the volumes of the hippocampal subfields. Paraffin embedded hippocampal sections from TLE and HRC were submitted to Hematoxilin-Eosin and Luxol Fast Blue histochemistries, and to immunohistochemistries for the evaluation of neurons (NeuN), reactive astrocytes (GFAP), activated microglia (HLA-DR), for aquaporin 4 (AQP4) and for chondroitin sulfate (CS-56). Students t-test or Mann-Whitneys test were performed for comparison between groups, and correlation tests were performed for the comparison between histological and magnetic resonance measures. Results: Patients with TLE presented reduced hippocampal volume, increased T2 relaxation time and reduced magnetization transfer, when compared to RC. The ex vivo volumetry of the hippocampal subfields revealed that fascia dentata, CA1 and subiculum together correspond to 85 % of the total hippocampal volume. Ex vivo relaxation time, as the in vivo, were increased in the subfields of TLE patients, when compared to HRC. Compared to HRC, TLE patients presented neuron loss and microgliosis in all hippocampal subfields but the subiculum, and astrogliosis in all hippocampal subfields but the subgranule zone and the hilus. Reduced perivascular aquaporin 4 was observed in all hippocampal subfields of TLE patients, and increased chondroitin sulfate was observed in all hippocampal subfields, with the exception of granule cell layer, of TLE patients, when compared to HRC. In TLE, both in vivo hippocampal volume and magnetization transfer correlated with the levels of chondroitin sulfate and the neuronal population, whereas the in vivo relaxation time correlated with the astroglial population and the levels of chondroitin sulfate. Ex vivo relaxation time also correlated with the astroglial population in TLE patients. The difference between in vivo and ex vivo relaxation values correlated with water difusibility and the levels of chondroitin sulfate. Conclusion: Our data indicate the importance of neuron population and extracellular matrix to both hippocampal volume and magnetization transfer, and of the reactive astrocytes for T2 relaxation. Ex vivo relaxation time allowed a more detailed evaluation, and indicated more robust correlations between reactive astrocytes and T2 relaxation. Finally, Our data indicated that CA1, the subiculum and fascia dentata are the major contributors to hippocampal volume, so changes in these subfields most likely will affect magnetic resonance imaging.

Dans cette thèse, le comportement d'un modèle mathématique permettant de transcrire la dynamique neuronale est étudié : le système de FitzHugh-Nagumo. En particulier, nous nous intéressons au caractère aléatoire d'ouverture et de fermeture des canaux ioniques d'un neurone qui reçoit ou non un stimulus. Ce caractère aléatoire de la dynamique neuronale est considéré, dans notre modèle, comme un bruit. Dans un premier temps, le comportement du modèle de FitzHugh-Nagumo a été caractérisé au voisinage de la bifurcation d'Andronov-Hopf qui traduit la transition entre l'état d'activation et l'état de repos du neurone. Classiquement, un neurone positionné à l'état de repos ne produit aucun potentiel d'action. Cependant, il a été montré un phénomène pour lequel une quantité appropriée de bruit permet la production de potentiels d'action des plus réguliers : la résonance cohérente. Le deuxième effet observé lors de simulations numériques permet au neurone d'améliorer la détection et l'encodage d'un signal subliminal : il s'agit de la résonance stochastique. De plus, cette thèse s'inscrit dans un contexte électronique puisqu'en plus de simuler numériquement le système de FitzHugh-Nagumo, les résultats de simulations ont également été confirmés en réalisant un circuit électronique. En effet, nous avons reproduit la dynamique non linéaire du système de FitzHugh-Nagumo à l'aide de ce circuit électronique. Cela a permis de mettre en évidence expérimentalement les deux phénomènes de résonance cohérente et de résonance stochastique pour lesquelles le bruit peut avoir une influence constructive sur le comportement de notre circuit électronique.