Submitted by Johnny Rodrigues (johnnyrodrigues@ufcg.edu.br) on 2018-04-04T21:48:36Z No. of bitstreams: 1 EUGÊNIO DE CARVALHO SARAIVA - DISSERTAÇÃO PPGCC 2014..pdf: 2813039 bytes, checksum: 9b76f48c8df4aee95923a8ce5f0385ce (MD5) / Made available in DSpace on 2018-04-04T21:48:36Z (GMT). No. of bitstreams: 1 EUGÊNIO DE CARVALHO SARAIVA - DISSERTAÇÃO PPGCC 2014..pdf: 2813039 bytes, checksum: 9b76f48c8df4aee95923a8ce5f0385ce (MD5) Previous issue date: 2014-07-30 / Há um número crescente de tecnologias que fazem uso de algoritmos de classificação para a automação de tarefas. Em particular, em Neurociências, algoritmos de classificação foram usados para testar hipóteses sobre o funcionamento do sistema nervoso central. No entanto, a relação entre as classes de padrões de ativação neuronal de áreas específicas do cérebro, como resultado de experiências sensoriais tem recebido pouca atenção. No contexto da Neurociência Computacional, este trabalho apresenta uma análise do nível de similaridade entre classes de padrões de ativação neuronal, com o uso das abordagens de aprendizagem não supervisionada e semi-supervisionada, em áreas específicas do cérebro de ratos em contato com objetos, obtidos durante um experimento envolvendo exploração livre de objetos pelos animais. As classes foram definidas de acordo com determinados tratamentos construídos com níveis específicos de um conjunto de 8 fatores (Animal, Região do Cérebro, Objeto ou Par de Objeto, Algoritmo de Agrupamento, Métrica, Bin, Janela e Intervalo de Contato). No total foram analisados 327.680 tratamentos. Foram definidas hipóteses quanto à relação de cada um dos fatores para com o nível de similaridade existente entre os tratamentos. As hipóteses foram verificadas por meio de testes estatísticos entre as distribuições que representavam cada uma das classes. Foram realizados testes de normalidade (Shapiro-Wilk, QQ-plot), análise de variância e um teste para diferenças entre tendência central (Kruskal-Wallis). Com base nos resultados encontrados nos estudos utilizando abordagem não supervisionada, foi inferido que os processos de aquisição e de definição dos padrões de ativação por um observador foram sujeitos a uma quantidade não significativa de ruídos causados por motivos não controláveis. Pela abordagem semisupervisionada, foi observado que nem todos os graus de similaridade entre pares de classes de objetos são iguais a um dado tratamento, o que indicou que a similaridade entre classes de padrões de ativação neuronal é sensível a todos os fatores analisados e fornece evidências da complexidade na codificação neuronal. / There are a growing number of technologies that make use of classification algorithms for automating tasks. In particular, in Neuroscience, classification algorithms were used to test hypotheses about the functioning of the central nervous system. However, the relationship between the classes of patterns of neuronal activation in specific brain areas as a result of sensorial experience has received little attention. In the context of Computational Neuroscience , this paper presents an analysis of the level of similarity between classes of patterns of neuronal activation with the use of learning approaches unsupervised and semi - supervised in specific areas of rat brain in contact with objects , obtained during an experiment involving free exploration of objects by animals. The classes were defined according to certain treatments constructed with specific levels with set of 8 factors (Animal, Brain Region, Object or Pair of Objects, Clustering Algorithm, Metric, Bin, Window and Interval Contact). In total 327.680 treatments were analyzed. Hypotheses regarding the relationship of each of the factors with the existing level of similarity between treatments were defined. The hypotheses were tested through between statistical distributions representing each class tests. The tests applied where the tests for normality (Shapiro-Wilk, QQ–plot), analysis of variance and a test for differences in central tendency (Kruskal-Wallis) were performed. Based on the results found in studies using an unsupervised approach, it was inferred that the process of acquisition and definition of patterns of activation by an observer was not subject to a significant amount of noise caused by uncontrollable reasons. For the semi-supervised approach, it was observed that not all degrees of similarity between pairs of classes of objects are equal to a given treatment, which indicated that the similarity between classes of patterns of neuronal activation is sensitive to all the factors analyzed and provides evidence about the complexity of neuronal coding.

Les processus de type “action-conséquence” (orienté vers un but) et stimulus-réponse sont deux composants importants du comportement. Le premier évalue le bénéfice d’une action pour choisir la meilleure parmi celles disponibles (sélection d’action) alors que le deuxième est responsable du comportement automatique, suscitant une réponse dès qu’un stimulus connu est présent. De telles habitudes sont généralement associées (et surtout opposées) aux actions orientées vers un but qui nécessitent un processus délibératif pour évaluer la meilleure option à prendre pour atteindre un objectif donné. En utilisant un modèle computationnel, nous avons étudié l’hypothèse classique de la formation et de l’expression des habitudes au niveau des ganglions de la base et nous avons formulé une nouvelle hypothèse quant aux rôles respectifs des ganglions de la base et du cortex. Inspiré par les travaux théoriques et expérimentaux de Leblois et al. (2006) et Guthrie et al. (2013), nous avons conçu un modèle computationnel des ganglions de la base, du thalamus et du cortex qui utilise des boucles distinctes (moteur, cognitif et associatif) ce qui nous a permis de poser l’hypothèse selon laquelle les ganglions de la base ne sont nécessaires que pour l’acquisition d’habitudes alors que l’expression de telles habitudes peut être faite par le cortex seul. En outre, ce modèle a permis de prédire l’existence d’un apprentissage latent dans les ganglions de la base lorsque leurs sorties (GPi) sont inhibées. En utilisant une tâche de bandit manchot à 2 choix, cette hypothèse a été expérimentalement testée et confirmée chez le singe; suggérant au final de rejeter l’idée classique selon laquelle l’automatisme est un trait subcortical. / Action-outcome and stimulus-response processes are two important components of behavior. The former evaluates the benefit of an action in order to choose the best action among those available (action selection) while the latter is responsible for automatic behavior, eliciting a response as soon as a known stimulus is present. Such habits are generally associated (and mostly opposed) to goal-directed actions that require a deliberative process to evaluate the best option to take in order to reach a given goal. Using a computational model, we investigated the classic hypothesis of habits formation and expression in the basal ganglia and proposed a new hypothesis concerning the respective role for both the basal ganglia and the cortex. Inspired by previous theoretical and experimental works (Leblois et al., 2006; Guthrie et al., 2013), we designed a computational model of the basal ganglia-thalamus-cortex that uses segregated loops (motor, cognitive and associative) and makes the hypothesis that basal ganglia are only necessary for the acquisition of habits while the expression of such habits can be mediated through the cortex. Furthermore, this model predicts the existence of covert learning within the basal ganglia ganglia when their output is inhibited. Using a two-armed bandit task, this hypothesis has been experimentally tested and confirmed in monkey. Finally, this works suggest to revise the classical idea that automatism is a subcortical feature.

Les astrocytes sont des cellules gliales du système nerveux central, essentielles à la formation des synapses, à la barrière hémato-encéphalique ainsi qu’au maintien de l'homéostasie. Récemment, les astrocytes ont été identifiés comme éléments clés du traitement de l'information dans le système nerveux central. Les astrocytes peuvent communiquer avec les neurones au niveau des synapses et moduler la communication neuronale en libérant des gliotransmetteurs et en absorbant des neurotransmetteurs. L’utilisation de nouvelles techniques comme la microscopie à super-résolution et les indicateurs calciques encodés génétiquement a permis de révéler une grande diversité spatio-temporelle des signaux calciques astrocytaires. La majorité de ces signaux sont observés au sein de leurs prolongements cellulaires, qui sont le site de communication entre neurones et astrocytes. Ces prolongements sont trop fins pour être observés en microscopie optique conventionnelle, de sorte que la microscopie à super-résolution et la modélisation informatique sont les seules méthodes adaptées à leur étude. Les travaux présentés dans cette thèse ont pour but d’étudier l'effet des propriétés spatiales (telles que la géométrie cellulaire, les distributions moléculaires et la diffusion) sur les signaux calciques dans les prolongements astrocytaires. Historiquement, les signaux calciques ont été modélisés à l'aide d'approches déterministes non spatiales. Ces modèles ont permis l'étude des signaux calciques à l’échelle de la cellule entière voire à l’échelle du réseau de cellules. Ces méthodes ne prennent cependant pas en compte la stochasticité inhérente aux interactions moléculaires ainsi que les effets de diffusion, qui jouent un rôle important dans les petits volumes. Cette thèse présente un modèle stochastique et spatial qui a été développé dans le but d’étudier les signaux calciques dans les prolongements fins astrocytaires. Ce travail a été réalisé en collaboration avec des expérimentateurs, qui nous ont fourni des données de microscopie électronique et à super-résolution. Ces données ont permis de valider le modèle. Les simulations du modèle suggèrent que (1) la diffusion moléculaire, fortement influencée par la concentration et la cinétique des buffers calciques endogènes et exogènes, (2) l'organisation spatiale intracellulaire des molécules, notamment le co-clustering des canaux calciques, (3) la géométrie du reticulum endoplasmique et sa localisation dans la cellule, (4) la géométrie cellulaire influencent fortement les signaux calciques et pourraient être responsables de leur grande diversité spatio-temporelle. Ces travaux contribuent à une meilleure compréhension du traitement de l’information par les astrocytes, un prérequis pour une meilleure compréhension de la communication entre les neurones et les astrocytes ainsi que de son influence sur le fonctionnement du cerveau. / Astrocytes are predominant glial cells in the central nervous system, which are essential for the formation of synapses, participate to the blood-brain barrier and maintain the metabolic, ionic and neurotransmitter homeostasis. Recently, astrocytes have emerged as key elements of information processing in the central nervous system. Astrocytes can contact neurons at synapses and modulate neuronal communication via the release of gliotransmitters and the uptake of neurotransmitters. The use of super-resolution microscopy and highly sensitive genetically encoded Ca2+ indicators (GECIs) has revealed a striking spatiotemporal diversity of Ca2+ signals in astrocytes. Most astrocytic signals occur in processes, which are the sites of neuron-astrocyte communication. Those processes are too fine to be resolved by conventional light microscopy so that super-resolution microscopy and computational modeling remain the only methodologies to study those compartments. The work presented in this thesis aims at investigating the effect of spatial properties (as e.g cellular geometry, molecular distributions and diffusion) on Ca2+ signals in those processes, which are deemed essential in such small volumes. Historically, Ca2+ signals were modeled with deterministic well-mixed approaches, which enabled the study of Ca2+ signals in astrocytic networks or whole-cell events. Those methods however ignore the stochasticity inherent to molecular interactions as well as diffusion effects, which both play important roles in small volumes. In this thesis, we present the spatially-extended stochastic model that we have developed in order to investigate Ca2+ signals in fine astrocytic processes. This work was performed in collaboration with experimentalists that performed electron as well as super-resolution microscopy. The model was validated against experimental data. Simulations of the model suggest that (1) molecular diffusion, strongly influenced by the concentration and kinetics of endogenous and exogenous buffers, (2) intracellular spatial organization of molecules, notably the co-clustering of Ca2+ channels, (3) ER geometry and localization within the cell, (4) cellular geometry strongly influence Ca2+ dynamics and can be responsible for the striking diversity of astrocytic Ca2+ signals. This work contributes to a better understanding of astrocyte Ca2+ signals, a prerequisite for understanding neuron-astrocyte communication and its influence on brain function.

Normalization is important for a large range of phenomena in biological neural systems such as light adaptation in the retina, context dependent decision making and probabilistic inference. In a normalizing circuit the activity of one neuron/-group of neurons is divisively rescaled in relation to the activity of other neurons/­­groups. This creates neural responses invariant to certain stimulus dimensions and dynamically adapts the range over which a neural system can respond discriminatively on stimuli. This thesis examines whether a biologically realistic normalizing circuit can be implemented by a spiking neural network model based on the columnar structure found in cortex. This was done by constructing and evaluating a highly structured spiking neural network model, modelling layer 2/3 of a cortical hypercolumn using a group of neurons as the basic computational unit. The results show that the structure of this hypercolumn module does not per se create a normalizing network. For most model versions the modulatory effect is better described as subtractive inhibition. However three mechanisms that shift the modulatory effect towards normalization were found: An increase in membrane variance for increased modulatory inputs; variability in neuron excitability and connections; and short-term depression on the driving synapses. Moreover it is shown that by combining those mechanisms it is possible to create a spiking neural network that implements approximate normalization over at least ten times increase in input magnitude. These results point towards possible normalizing mechanisms in a cortical hypercolumn; however more studies are needed to assess whether any of those could in fact be a viable explanation for normalization in the biological nervous system. / Normalisering är viktigt för en lång rad fenomen i biologiska nervsystem såsom näthinnans ljusanpassning, kontextberoende beslutsfattande och probabilistisk inferens. I en normaliserande krets skalas aktiviteten hos en nervcell/grupp av nervceller om i relation till aktiviteten hos andra nervceller/grupper. Detta ger neurala svar som är invarianta i förhållande till vissa dimensioner hos stimuli, och anpassar dynamiskt för vilka inputmagnituder ett system kan särskilja mellan stimuli. Den här uppsatsen undersöker huruvida en biologiskt realistisk normal­iserande krets kan implementeras av ett spikande neuronnätverk konstruerat med utgångspunkt från kolumnstrukturen i kortex. Detta gjordes genom att konstruera och utvärdera ett hårt strukturerat rekurrent spikande neuronnätverk, som modellerar lager 2/3 av en kortikal hyperkolumn med en grupp av neuroner som grundläggande beräkningsenhet. Resultaten visar att strukturen i hyperkolumn­modulen inte i sig skapar ett normaliserande nätverk. För de flesta nätverks­versioner implementerar nätverket en modulerande effekt som bättre beskrivs som subtraktiv inhibition. Dock hittades tre mekanismer som skapar ett mer normaliserande nätverk: Ökad membranvarians för större modulerande inputs; variabilitet i excitabilitet och inkommande kopplingar; och korttidsdepression på drivande synapser. Det visas också att genom att kombinera dessa mekanismer är det möjligt att skapa ett spikande neuronnät som approximerar normalisering över ett en åtminstone tio gångers ökning av storleken på input. Detta pekar på möjliga normaliserande mekanismer i en kortikal hyperkolumn, men ytterligare studier är nödvändiga för att avgöra om en eller flera av dessa kan vara en förklaring till hur normalisering är implementerat i biologiska nervsystem.

<div>The innovations in Magnetic Resonance Imaging (MRI) in the recent decades have given rise to large open-source datasets. MRI affords researchers the ability to look at both structure and function of the human brain. This dissertation will make use of one of these large open-source datasets, the Human Connectome Project (HCP), to study the structural and functional connectivity in the brain.</div><div>Communication processes within the human brain at different cognitive states are neither well understood nor completely characterized. We assess communication processes in the human connectome using ant colony-inspired cooperative learning algorithm, starting from a source with no <i>a priori</i> information about the network topology, and cooperatively searching for the target through a pheromone-inspired model. This framework relies on two parameters, namely <i>pheromone</i> and <i>edge perception</i>, to define the cognizance and subsequent behaviour of the ants on the network and the communication processes happening between source and target. Simulations with different configurations allow the identification of path-ensembles that are involved in the communication between node pairs. In order to assess the different communication regimes displayed on the simulations and their associations with functional connectivity, we introduce two network measurements, effective path-length and arrival rate. These measurements are tested as individual and combined descriptors of functional connectivity during different tasks. Finally, different communication regimes are found in different specialized functional networks. This framework may be used as a test-bed for different communication regimes on top of an underlying topology.</div><div>The assessment of brain <i>fingerprints</i> has emerged in the recent years as an important tool to study individual differences. Studies so far have mainly focused on connectivity fingerprints between different brain scans of the same individual. We extend the concept of brain connectivity fingerprints beyond test/retest and assess <i>fingerprint gradients</i> in young adults by developing an extension of the differential identifiability framework. To do so, we look at the similarity between not only the multiple scans of an individual (<i>subject fingerprint</i>), but also between the scans of monozygotic and dizygotic twins (<i>twin fingerprint</i>). We have carried out this analysis on the 8 fMRI conditions present in the Human Connectome Project -- Young Adult dataset, which we processed into functional connectomes (FCs) and time series parcellated according to the Schaefer Atlas scheme, which has multiple levels of resolution. Our differential identifiability results show that the fingerprint gradients based on genetic and environmental similarities are indeed present when comparing FCs for all parcellations and fMRI conditions. Importantly, only when assessing optimally reconstructed FCs, we fully uncover fingerprints present in higher resolution atlases. We also study the effect of scanning length on subject fingerprint of resting-state FCs to analyze the effect of scanning length and parcellation. In the pursuit of open science, we have also made available the processed and parcellated FCs and time series for all conditions for ~1200 subjects part of the HCP-YA dataset to the scientific community.</div><div>Lastly, we have estimated the effect of genetics and environment on the original and optimally reconstructed FC with an ACE model.</div>

No description available.

Neuronal membrane disruption and mechanoporation are nanoscale damage mechanisms that critically affect brain cell viability during traumatic brain injury (TBI). These nanoscale cellular impairments are elusive in experiments and necessitate in silico approaches such as molecular dynamics (MD) simulations. Implementing MD, this research aims to investigate the effects of different key factors related to membrane deformation and damage, including force field resolutions, lipid compositions, and loading conditions. To examine the impact of force field resolution, MD deformation simulations were conducted on 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphatidylcholine (POPC) lipid bilayer membranes, using all-atom (AA), united-atom (UA), and coarse-grained Martini (CG-M) force fields. The mechanical responses of the three models progressively changed based on the coarse-graining level. The coarser systems exhibited lower yield stresses and failure strains, and higher mechanoporation damage. To study the influence of lipid components, tensile deformation was applied on seven lipid bilayers, each of which contained a different lipid type commonly found in human brain membrane. Larger headgroup structure, greater degree of unsaturation, and tail-length asymmetry decreased lipid packing, increased the area per lipid (APL), and decreased the failure strain of membrane. Lastly, the deformation behavior of a complex multicomponent MD bilayer (realistically representing human neuronal plasma membrane) under different strain rates and strain states was inspected. The yield stress increased with increasing strain rates and more equibiaxial strain states. Meanwhile, lower strain rates resulted in fewer but larger pores, as well as lower strain and APL at failure. Besides, more equibiaxial strain states exhibited more and larger pores, and lower failure strain. Similar failure APL was obtained regardless of strain states, suggesting that the membrane failed when reaching a critical APL value. In addition, the inclusion of cholesterol was shown to decrease the critical APL. The strain-state dependence results were then used to update the Membrane Failure Limit Diagram (MFLD) that indicates the planar strains for potential membrane failure. Overall, the study provides a non-invasive approach that aids in the current understanding of nanoscale neuronal damage dynamics and essential aspects affecting membrane mechanical responses, and furthermore lays the groundwork for future studies on brain injury biomechanics under various TBI scenarios.

Every year, 250,000 people worldwide suffer a spinal cord injury (SCI) that leaves them with chronic paraplegia - permanent loss of ability to move their legs. SCI interrupts axons passing along the spinal cord, thereby isolating motor neurons from brain inputs. To date, there are no effective treatments that can reconnect these interrupted axons. In a recent breakthrough, .NeuroRestore developed the STIMO neuroprosthesis that can restore walking after paralyzing SCI using Epidural Electrical Stimulation (EES) of the lumbar spinal cord. Yet, the calibration of EES requires highly trained personnel and a vast amount of time, and the mechanism by which EES restores movement is not fully understood. In this master thesis, we propose to address this issue using modeling combined with Artificial Neural Networks (ANNs). To do so, we introduce the CyberSpine model to predict EES-induced motor response. The implementation of the model relies on the construction of a multipolar basis of solution of the Poisson equation which is then coupled to an ANN trained against actual data of an implanted STIMO user. Furthermore, we show that our CyberSpine model is particularly well adapted to extract biologically relevant information regarding the efficient connectivity of the patient’s spine. Finally, a user-friendly interactive visualization software is built. / Varje år drabbas 250 000 människor i hela världen av en ryggmärgsskada som ger dem kronisk paraplegi - permanent förlust av förmågan att röra benen. Vid en ryggmärgsskada bryts axonerna som passerar längs ryggmärgen, vilket isolerar de motoriska neuronpoolerna från hjärnans ingångar. Hittills finns det inga effektiva behandlingar som kan återansluta dessa avbrutna axoner. NeuroRestore utvecklade nyligen neuroprotesen STIMO som kan återställa gångförmågan efter förlamande ryggmärgsskada med hjälp av epidural elektrisk stimulering (EES) av ländryggmärgen. Kalibreringen av EES-stimuleringar kräver dock högutbildad personal och mycket tid, och den mekanism genom vilken EES återställer rörelse är inte helt klarlagd. I denna masteruppsats föreslår vi att vi tar itu med denna fråga med hjälp av modellering i kombination med artificiell intelligens. För att göra detta introducerar vi CyberSpine-modellen, en modell som kan förutsäga EES-inducerad motorisk respons. Implementeringen av modellen bygger på konstruktionen av en multipolär bas för lösning av Poisson-ekvationen som sedan kopplas till ett artificiellt neuralt nätverk som tränas mot faktiska data från en implanterad STIMO-deltagare. Dessutom visar vi att vår CyberSpine-modell är särskilt väl anpassad för att extrahera biologiskt relevant information om den effektiva anslutningen av patientens ryggrad. Slutligen bygger vi en användarvänlig interaktiv visualiseringsprogramvara.

This research investigated the concept of a sparsity method inspired by the principles of structural plasticity in the brain in order to create a sparse model of the Bayesian Confidence Propagation Neural Networks (BCPNN) during the training phase. This was done by extending the structural plasticity in the implementation of the BCPNN. While the initial algorithm presented two synaptic states (Active and Silent), this research extended it to three synaptic states (Active, Silent and Absent) with the aim to enhance sparsity configurability and emulate a more brain-like algorithm, drawing parallels with synaptic states observed in the brain. Benchmarking was conducted using the MNIST and Fashion-MNIST dataset, where the proposed threestate model was compared against the previous two-state model in terms of representational learning. The findings suggest that the three-state model not only provides added configurability but also, in certain low-sparsity settings, showcases similar representational learning abilities as the two-state model. Moreover, in high-sparsity settings, the three-state model demonstrates a commendable balance between accuracy and sparsity trade-off. / Denna forskning undersökte en konceptuell metod för gleshet inspirerad av principerna för strukturell plasticitet i hjärnan för att skapa glesa BCPNN. Forskningen utvidgade strukturell plasticitet i en implementering av BCPNN. Medan den ursprungliga algoritmen presenterade två synaptiska tillstånd (Aktiv och Tyst), utvidgade denna forskning den till tre synaptiska tillstånd (Aktiv, Tyst och Frånvarande) med målet att öka konfigurerbarheten av sparsitet och efterlikna en mer hjärnliknande algoritm, med paralleller till synaptiska tillstånd observerade i hjärnan. Jämförelse gjordes med hjälp av MNIST och Fashion-MNIST datasetet, där det föreslagna tre-tillståndsmodellen jämfördes med den tidigare tvåtillståndsmodellen med avseende på representationslärande. Resultaten tyder på att tre-tillståndsmodellen inte bara ger ökad konfigurerbarhet utan också, i vissa lågt glesa inställningar, visar samma inlärningsförmåga som två-tillståndsmodellen. Dessutom visar den tre-tillståndsmodellen i högsparsamma inställningar en anmärkningsvärd balans mellan noggrannhet och avvägningen mellan sparsitet.

No description available.