Humans have an innate ability to deal with numerosity and other aspects of magnitude. This ability is generally honed through education in and experience with mathematics, which necessarily changes the brain structurally and functionally. These changes can be further manipulated through non-invasive electrical brain stimulation. Studying these processes in the case of maths not only constitutes research of great practical impact – given the importance of numerical skills in today's society – but also makes use of maths as a suitable domain in which to study plasticity. In this thesis, I aimed to explore how expertise with numbers shapes brain and behaviour, and also the degree to which processing numbers is similar to other domains in terms of the necessity of healthy brain regions believed to underlie normal processing within and across these domains. In Study 1, behavioural and structural brain differences were found cross-sectionally between mathematicians and non-mathematicians. A double dissociation between those groups was found between grey matter density in the frontal lobe and behavioural performance: their correlation was positive for mathematicians but negative for controls. These effects may have been caused by years of experience, by congenital predispositions, or, plausibly, by both of these factors, whose disambiguation is non-trivial. Study 2 used transcranial random noise stimulation (tRNS) to assist arithmetic learning. A novel montage was used to enhance brain function during the stage when it is believed to be most involved. Real as compared to sham tRNS enhanced reaction times (RTs) and learning rate on a calculation-based task, but not on a retrieval-based task. The effects were only observed in conditions of high task difficulty. Study 3 examined structural MRI measures before and after arithmetic training to determine how either frontal or parietal tRNS applied with the task changes the structure of the brain longitudinally as compared to sham. Previous results (including those of Study 2) of behavioural facilitation in terms of enhanced RTs to calculation problems were replicated, and further interpreted. Both frontal and parietal tRNS modulated the changes that occurred, pre-to-post training, in terms of cortical volume and gyrification of frontal, parietal and temporal areas. Study 4 investigated the shared neural and cognitive resources used for processing numerical magnitude and musical pitch, by probing how stimulus-response compatibility (SRC) effects for each of the two dimensions compare in a group of mainly temporoparietal lesion patients with numerical impairments versus controls. A double dissociation was found in that numerically impaired patients did not show the number-based SRC effect but did show the pitch-based one, while control subjects demonstrated the opposite trend. Overall, the results of these studies leave us with three main messages. First, expertise in numbers and mathematics, whether acquired through years of experience (Study 1) or through a few days of tRNS-assisted training (Study 3), appears to be associated with complex changes in the morphology of several brain structures. Some – but not all – of these structures are maths-relevant, and, in the case of tRNS-assisted training, they are distal to the site of the stimulating electrodes. Second, tRNS can improve performance in arithmetic (Studies 2 and 3), although the mechanisms by which this occurs are not yet fully understood, neither neurally nor behaviourally. Third, I found (Study 4) that brain lesions leading to impairment in the number domain do not necessarily affect processing in other domains – such as pitch – that are otherwise linked to number via a putative common code in the parietal lobes.

Cette thèse porte sur les problèmes de déploiement des Réseaux de Capteurs sans Fil (RCsF). Elle suit trois directions principales : (1) le déploiement aléatoire, (2) le déploiement déterministe, et (3) l'auto-déploiement. En premier lieu, nous présentons une étude sur le placement aléatoire des capteurs dans les RCsF et nous élaborons une stratégie pratique de déploiement aléatoire. Ensuite, dans le cadre d'un déploiement déterministe, nous analysons le problème de gestion des imperfections liées à la collecte des données par les capteurs. Nous discutons les modèles de couverture et les algorithmes de placement existants et nous exploitons la théorie de l'évidence pour concevoir des stratégies de déploiement plus efficaces. Enfin, nous explorons les stratégies d'auto-déploiement existantes et nous élaborons un protocole en deux phases, léger et complet, pour assurer une couverture optimisée de la zone contrôlée en utilisant un RCsF mobile. Les résultats obtenus montrent l'efficacité des approches proposée qui ont été étudiées à la fois sur des données synthétiques que sur un test expérimental / This dissertation is an in-depth investigation of the Wireless Sensor Networks (WSNs) deployment problems that follows three general directions: (1) random deployment, (2) deterministic deployment, and (3) self-deployment. First, we present a survey and taxonomy of random node placement in WSNs and we devise a practical random deployment strategy. Second, we analyze the uncertainty-aware deterministic WSNs deployment problem where sensors may not always provide reliable information. We discuss sensor coverage models and placement algorithms found in the literature and we investigate the evidence theory to design better deployment strategies. We devise evidence-based sensor coverage models and we propose several polynomial-time uncertainty-aware deployment algorithms. Third, we explore the published self-deployment strategies and we devise a lightweight and comprehensive two-phase protocol, for ensuring area coverage employing a mobile WSN. Experimental results based on synthetic data sets, data traces collected in a real deployment, and an experimental test, show that the proposed approaches outperform the state-of-the-art deployment strategies

Le contexte lié au changement global et aux risques encourus par la biodiversité ont conduit les gestionnaires à développer des outils afin de préserver notre environnement et la biodiversité De nombreuses questions ont alors été soulevées notamment au sujet de la connectivité entre populations à préserver et de la définition des aires protégées. En milieu marin cette question soulève le problème des capacités de dispersion et du lien entre durée de la phase larvaire et flux de gènes. Dans ce contexte nous avons choisi d'étudier la connectivité entre populations de deux espèces à valeur commerciale et patrimoniale que sont l'oursin commun Paracentrotus lividus et le sar commun Diplodus sargus. Dans le cas de P. lividus, l'étude combinant plusieurs jeux de données utilisant des marqueurs moléculaires différents, à la fois produits pendant ce travail de thèse et issus de la littérature, a mis en évidence une différenciation génétique significative insoupçonnée à l'échelle régionale (40km) pour une espèce présentant une phase larvaire pélagique pouvant atteindre un mois. Cette étude a également permis de caractériser l'intensité et la direction des flux de gènes entre les bassins océaniques du sud-est de l'Atlantique Nord, de la Méditerranée et de l'Adriatique. / The context related to global change and risks of biodiversity loss have led managers to develop tools to help preserve our environment and the associated biodiversity. Many questions have been raised especially concerning the connectivity between populations and the definition of marine protected areas. These questions raise the problem of dispersal abilities and of the relationship between pelagic larval duration and gene flow. In this context we have chosen to study the connectivity between populations of two species of high commercial and patrimonial values: the edible sea urchin Paracentrotus lividus and the white seabream Diplodus sargus. In the case of P. lividus, the study combining several data sets and using different molecular markers, produced during this work or obtained from the literature, showed unexpected significant genetic differentiation at the regional scale (40km) for a species with a pelagic larval duration of up to one month. This study also allowed us to characterize the intensity and direction of gene flow between ocean basins of the southeastern North Atlantic, Mediterranean Sea and Adriatic Sea. Indeed gene flow was asymmetric and oriented from the Atlantic to the Mediterranean Sea and from the Adriatic Sea and to the Mediterranean Sea. These findings raise important questions about the origin of these different levels of differentiation and led me to make several assumptions on the impact of phenomena such as local adaptation, selective effects coupled with endogenous barriers, hydrological factors and / or biogeochemical. Several research perspectives are proposed to test these hypotheses.

Balance of motor network activity between the two brain hemispheres after stroke is crucial for functional recovery. Several studies have extensively studied the role of the affected brain hemisphere to better understand changes in motor network activity following stroke. Very few studies have examined the role of the unaffected brain hemisphere and confirmed the testretest reliability of connectivity measures on unaffected hemisphere. We recorded blood oxygenation level dependent functional magnetic resonance imaging (fMRI) signals from nine stroke survivors with hemiparesis of the left or right hand. Participants performed a motor execution task with affected hand, unaffected hand, and both hands simultaneously. Participants returned for a repeat fMRI scan 1 week later. Using dynamic causal modeling (DCM), we evaluated effective connectivity among three motor areas: the primary motor area (M1), the premotor cortex (PMC) and the supplementary motor area for the affected and unaffected hemispheres separately. Five participants manual motor ability was assessed by Fugl-Meyer Motor Assessment scores and root-mean square error of participants tracking ability during a robot-assisted game. We found (i) that the task performance with the affected hand resulted in strengthening of the connectivity pattern for unaffected hemisphere, (ii) an identical network of the unaffected hemisphere when participants performed the task with their unaffected hand, and (iii) the pattern of directional connectivity observed in the affected hemisphere was identical for tasks using the affected hand only or both hands. Furthermore, paired t-test comparison found no significant differences in connectivity strength for any path when compared with one-week follow-up. Brain-behavior linear correlation analysis showed that the connectivity patterns in the unaffected hemisphere more accurately reflected the behavioral conditions than the connectivity patterns in the affected hemisphere. Above findings enrich our knowledge of unaffected brain hemisphere following stroke, which further strengthens our neurobiological understanding of stroke-affected brain and can help to effectively identify and apply stroke-treatments.

EUROPEAN AIR HUBS IN THE CONTEXT OF NETWORK AND ITS RESISTANCE AGAINST DISTURBANCES Abstract The submitted master thesis is addressing the theme of connectivity of European Air Transport Network, its properties and resistance against negative influences based on data from flight schedules for winter season 2018. The main objective of the thesis is to analyse European Air Transport Network from the point of connectivity in order to find out the most important airport hubs according to their geographic conditions, community structure and resistance of the whole network. Used methods are based on the Graph Theory and the centrality measures as indicators of connectivity. The empiric part of the thesis is divided into three parts. The aim of the first part is to find out, what airports are the most important in the European Air Transport Network. In the second part are explored properties and structure of the network. The last part is aiming to analyse the resistance of the European Air Transport Network from the view of robustness and resilience. Among the most important air hubs in Europe belong airports, that are serving world cities and tourist attractive localities. There is a strong dominance of the Schiphol airport in Amsterdam, the El Prat airport in Barcelona and the Frankfurt Airport. The European Air...

Landscape genetic approaches allow to study effects of landscape to population microevolution. Landscape can influence gene flow even in large carnivores with good dispersal ability. Understanding the influence of landscape to the gene flow between populations is crucial for species conservation, especially in the species with low population densities. Aim of the study was to describe genetic structure of the grey wolf (Canis lupus) and snow leopard (Panthera uncia) in selected areas and to determine the influence of the landscape features on observed structure. Non-invasive genetic samples of snow leopard from Nepal were analysed, as well as invasive and non-invasive samples of grey wolf from Central Europe. Population structure was determined a posteriori using Bayesian clustering approaches that integrate genetic and geographical data, and compared to landscape connectivity models. Population structure of snow leopards is mostly influenced by human presence and presence of frequented roads, which represent a substantial dispersal barrier. Habitat suitable for this species is greatly restricted by altitude, however, during dispersal they are able to overcome areas with higher elevation than what is optimal for them. Pronounced genetic difference was found between central European and Carpathian...

Le traçage rétrograde des voies corticales chez le singe a permis d’étudier deux sujets liés. En premier lieu, des injections dans 40 aires d'un atlas cortical de 91 aires ont permis de constituer une base de données cohérente sur la connectivité corticale à l’échelle de l’hémisphère. Les structures sous-corticales favorisant la communication corticale via la formation de boucles cortico-sous-cortico-corticales ont été examinées. Nous montrons que la force des projections du claustrum (considéré comme ayant une affiliation étroite avec le cortex) vers chaque aire explorée est exceptionnelle. De plus, un chevauchement des neurones marqués dans le claustrum a été observé suite à des paires d'injections dans des aires largement éloignées, y compris dépourvues de connexions cortico-corticales directes. A l’aide d’outils de la théorie des graphes, nous avons examiné la centralité des 40 aires et du claustrum dans le réseau cortical. En particulier, le claustrum est le meilleur exemple d’une aire pouvant prétendre au statut de « hub ». Ces résultats soulignent l'importance d'étudier les principes organisationnels du cortex via l'analyse de la topologie de son réseau. En second lieu, nous avons étudié le développement de la voie corticospinale par laquelle le cortex influence la planification, l'exécution et le contrôle de la motricité fine. Nous montrons que la topologie des projections corticospinales chez l’adulte émerge suite à un processus développemental de raffinement des projections ipsi- et controlatérale étendues. Ces résultats suggèrent que le développement de la connectivité corticale pourrait être régulé de manière dynamique et spécifique aux primates / The retrograde tracing experiments in macaque cortex in this thesis had two related objectives. Firstly, injections in 40 cortical areas (from a 91-area atlas) allowed the construction of a hemisphere-wide consistent database of cortical connectivity. We examined which subcortical structures promote cortical communication via the formation of cortico-subcortical-cortical loops. The claustrum, which we argue has a tight affiliation with the cortex, showed uniquely strong outputs to every cortical area. Widely separated injection pairs led to overlapping labelled neurons in the claustrum including those pairs lacking direct cortico-cortical connections. Using graph theoretic tools, we examined how central the 40 areas and claustrum are in the cortical network, specifically with respect to hub status. This showed that the claustrum is, beyond doubt, the prime hub of the cortex. These findings emphasise the importance of studying the organizational principles of the cortex via the analysis of its network topology. Secondly, we investigated the development of the corticospinal pathway, a route over which the cortex directly influences the planning, execution and control of fine voluntary movements. We show that the adult pattern of corticospinal projections emerges via a developmental process from a widespread ipsi- and contralateral distribution. These findings suggest that the developmental refinement of cortical connectivity might be dynamically regulated and primate specific

INTRODUÇÃO: Transtornos psiquiátricos com sintomas psicóticos trazem prejuízos ocupacionais e sociais significativos aos seus portadores, com aumento da mortalidade e morbidade. Estes transtornos têm sido estudados como alterações do padrão de conectividade funcional nas redes cerebrais do estado de repouso. Entretanto, tais relatos são mais frequentes em pacientes crônicos, com literatura escassa sobre pacientes em primeiro episódio psicótico, principalmente não realizando comparações entre as redes. OBJETIVO: Este foi um estudo exploratório com objetivo de caracterizar a conectividade funcional cerebral das redes do estado de repouso em pacientes de primeiro episódio psicótico através da ressonância nuclear magnética funcional comparando-os a irmãos e a controles de base populacional. MÉTODOS: A amostra foi composta por por 38 pacientes em primeiro episódio psicótico, 13 irmãos e 41 controles de base populacional. Foram coletadas imagens por ressonância magnética funcional e a conectividade funcional foi obtida através do coeficiente de correlação de Pearson da série temporal do sinal BOLD de 264 regiões de interesse. A comparação da conectividade funcional entre os grupos de participantes foi feita pelo método Partial Least Square. Também foi utilizado o método Behavior Partial Least Square para buscar um padrão de conexões alteradas que estivesse associado a gravidade dos sintomas psicóticos, ao tempo de tratamento e a duração da psicose não tratada. A análise estatística contou com 10.000 permutações e um método de reamostragem e foram considerados significativos valores de p<0.05. RESULTADOS: As alterações nas redes, em sua maioria, foram devido à mudança de correlação positiva para correlação negativa nos pacientes em relação aos controles. As redes com maior número de conexões alteradas entre pacientes e controles foram a rede sensóriomotor mão, Default Mode Network (DMN) e rede visual. As conexões estiveram mais alteradas no lobo frontal direito. Não foi encontrada associação entre o padrão de conectividade funcional dos pacientes e a duração de psicose não tratada, o tempo de tratamento farmacológico e a gravidade da psicose. Na comparação entre pacientes e irmãos, foi encontrada uma tendência à significância de um padrão de conexões alteradas. Não foi encontrada diferença significativa entre o grupo de irmãos e o grupo controle. DISCUSSÃO: Pacientes em primeiro episódio psicótico apresentaram maior segregação das redes do estado de repouso comparados a controles de base populacional, corroborando a hipótese etiológica da Esquizofrenia de uma desconectividade funcional do cérebro. A rede sensório-motor mão surpreendentemente foi a rede com maior número de alterações, apontando a necessidade de mais estudos sobre a mesma. Os irmãos não apresentaram um padrão de conexões do repouso diferente dos controles, não corroborando as hipóteses de que tal grupo apresentaria um padrão intermediário entre pacientes e controles. CONCLUSÃO: Os achados apontam para um uma topologia cerebral amplamente prejudicada já no início da psicose, com uma maior segregação entre as redes do estado de repouso em pacientes de primeiro episódio psicótico. / INTRODUCTION: Psychiatric disorders with psychotic symptoms bring significant occupational and social harm to their patients, with increased mortality and morbidity. These disorders have been studies as changes in the functional connectivity patterns in resting state brain networks. However, reports are more frequently made in chronic patients, with a scarce literature from first episode psychosis patients, mostly not making intra-networks comparison. OBJECTIVE: This was an exploratory study that had the objective of characterize the brain functional connectivity of resting networks in first episode psychosis patients through functional magnetic resonance imaging compared to siblings and to population based controls. METHODS: Final sample consisted of 38 first episode psychosis, 13 siblings and 41 population-based controls. Functional magnetic resonance images were collected in first episode psychosis, siblings and population based controls. Functional connectivity was obtained through the Pearson Correlation Coefficient of 264 regions of interest BOLD signal time series´. Comparison of functional connectivity among groups of participants was made using Partial Least Square method. Behavior Partial Least Square was performed to seek for a pattern associated with illness severity, pharmacological treatment time and duration of untreated psychosis. The statistical analysis was conducted with 10,000 permutations and bootstrap considering significant values of p<0.05. RESULTS: Aberrant network connections were mostrly due to changes of positive correlation to negative correlation in patients compared to controle. The majority of altered connections were found in sensory-motor network, DMN and visual network. The areas most affected were right frontal lobe. It was not found a functional connectivity pattern associated with illness severity, treatment time and duration of untreated psychosis. A tendency difference was found in the connectivity pattern between siblings and patients. No different connectivity pattern was found between siblings and controls. DISCUSSION: First episode psychosis presented more segregated resting state networks than controls, reinforcing the disconnectivity etiology hypothesis for schizophrenia. An unexpected result was sensory-motor hand network being the network with more altered connections, pointing to the need of more studies to comprehend it. The sibling group did not differ from the control group, not corroborating the hypotheses that such a group would present an intermediate pattern between patients and controls. CONCLUSION: The findings point to a largely impaired brain topology already at the beginning of the psychosis, with greater segregation between resting state networks in patients with first episode psychosis.

A quantidade de habitat remanescente, a conectividade e o tamanho dos fragmentos são sugeridos como determinantes para a manutenção de espécies em paisagens fragmentadas, e dessa maneira é essencial entendermos e modelarmos tais relações para o embasamento de políticas de conservação. Para analisarmos tais efeitos na riqueza e abundância de aves de sub-bosque e na abundância de 29 espécies de aves da Mata Atlântica, construímos modelos com variáveis representando o tamanho do fragmento e graus diversos de conectividade funcional em três paisagens com diferentes proporções de habitat (14, 31 e 45% de habitat). Posteriormente, confrontamos tais modelos com dados provenientes de amostragens com redes de neblina (4.818 indivíduos) com um esforço amostral total de quase 34.000 horas-rede em 53 fragmentos de diferentes tamanhos (2 a 159 ha) e graus de conectividades (considerando conexões por corredor ou por curtas distâncias pela matriz). Em seguida, hierarquizamos tais modelos através do critério de Informação de Akaike. Um total de 117 espécies foi capturado, e a paisagem com maior riqueza foi aquela de maior proporção de mata (87), enquanto as outras duas apresentaram riquezas semelhantes e menores (62 e 70), o que sugere um limiar de habitat entre 32 e 44% para a perda de um grande número de espécies. Além disso, os aspectos de configuração foram importantes em todas as paisagens, contudo, diferentes características foram relevantes em cada um dos casos. Enquanto na paisagem com menor proporção de habitat o tamanho dos fragmentos foi extremamente influente na determinação da riqueza e da abundância de espécies, os aspectos relacionados à conectividade foram mais importantes nas paisagens com 31 e 45% de proporção de habitat. Dentre as variáveis de conectividade, as que representam conectividade por corredores foram particularmente importantes, seguidas pelas de pequenos cruzamentos pela matriz (20 m). Modelos que levaram em consideração cruzamentos de 40 m pela matriz foram bem menos plausíveis que os demais. Os modelos para abundância das 29 espécies também resultaram em um padrão similar ao observado para a comunidade. No entanto, a importância relativa do tamanho do fragmento e da conectividade mudou em função da proporção de habitat, de acordo com a espécie abordada. Os resultados aqui apresentados são importantes para o embasamento de políticas de manejo para a conservação, e ressaltam a influência relativa das variáveis de configuração ao longo do gradiente de quantidade de hábitat, e devem ser consideradas quando na seleção de áreas para conservação, restauração ou qualquer forma de manejo voltado para a conservação. / Landscape aspects such as habitat amount, fragment size and connectivity, have been suggested as key features for maintaining species in altered landscapes, and thus are essential to be modeled and comprehended for management and conservation purposes. To test their effects on understory Atlantic forest bird species, we built models with variables representing fragment size and functional connectivity in 3 landscapes with different proportions of forest (14, 31 and 45%), and confronted with data from more than 4,818 individuals captured in 53 fragments of different sizes (2.06 to 158.45 ha) and connectivitys (connected by corridors, or by short distances), and then we ranked under a model selection approach (AIC). A total of 117 species were recorded, and the landscape with higher amount of forest was richer (87) than the two others which presented a similar lower richness (62 and 70), what suggests a threshold amount of habitat between 32 to 44%. Configuration aspects were important in all cases; however, different characteristics influenced species richness and abundance in each landscape. Patch size was important in determining species richness and abundance in the landscape with low amount of habitat, while connectivity aspects were crucial in the landscapes with higher amounts of habitat. Within the connectivity models, particularly the ones that take in to account the connectivity by corridors, but also the ones that consider short movements through the matrix are far better supported, then the ones by longer distances gaps. Moreover, specieslevel analyses yielded results similar to the pattern found for the whole assemblage, and the configuration variable influence varied along the habitat conversion gradient. The results presented here are important for management planning towards conservation, and should be considered in site selection for conservation, restoration or any other kind of management.

Os avanços nas técnicas de neuroimagem, principalmente com o de- senvolvimento da ressonância magnética funcional (RMf), vem possibilitando um melhor compreendimento dos processos e mecanismos cerebrais. Este trabalho tem como objetivo o desenvolvimento de um modelo de conectividade dinâmico entre diversas áreas cerebrais útilzando dados de RMf. A modelagem dinâmica do fluxo de informação é realizada com a estimação dos parâmetros de um modelo autoregressivo multivariado com coeficientes variandos no tempo, baseado na projeçã o de funções em bases de ondaletas. Dessa forma, um método para estimação e a derivação de suas propriedades assintóticas são apresentados. Diversos conjuntos de simulações computacionais são realizados visando a avaliação do desempenho do método proposto. Por fim, são apresentadas aplicações do modelo de conectividade variante no tempo em dados de ressonância magnética funcional. / Advances in neuroimage technologies, mainly with the development of functional magnetic resonance imaging (fMRI), improve the comprehension of brain processes and mechanisms. The main goal of this work is the development of a time-varying connectivity model between many brain areas using fMRI datasets. The dynamic modelling of the information flow is related to the parameters estimation of a time-varying multivariate autoregressive process, based on functions projection in wavelet basis. We propose an estimation procedure and present its asymptotic properties. Computational simulations were performed focusing the evaluation of the proposed approach. Further, applications of these methodologies to real functional magnetic resonance datasets are presented.