Hippocampal place-cell sequences observed during awake immobility often represent previous experience, suggesting a role in memory processes. However, recent reports of goals being overrepresented in sequential activity suggest a role in short-term planning, although a detailed understanding of the origins of hippocampal sequential activity and of its functional role is still lacking. In particular, it is unknown which mechanism could support efficient planning by generating place-cell sequences biased toward known goal locations, in an adaptive and constructive fashion. To address these questions, we propose a model of spatial learning and sequence generation as interdependent processes, integrating cortical contextual coding, synaptic plasticity and neuromodulatory mechanisms into a map-based approach. Following goal learning, sequential activity emerges from continuous attractor network dynamics biased by goal memory inputs. We apply Bayesian decoding on the resulting spike trains, allowing a direct comparison with experimental data. Simulations show that this model (1) explains the generation of never-experienced sequence trajectories in familiar environments, without requiring virtual self-motion signals, (2) accounts for the bias in place-cell sequences toward goal locations, (3) highlights their utility in flexible route planning, and (4) provides specific testable predictions.

While in traditional computer game playing agents were designed solely for the purpose of playing one single game, General Game Playing is concerned with agents capable of playing classes of games. Given the game's rules and a few minutes time, the agent is supposed to play any game of the class and eventually win it. Since the game is unknown beforehand, previously optimized data structures or human-provided features are not applicable. Instead, the agent must derive a strategy on its own. One approach to obtain such a strategy is to analyze the game rules and create a state evaluation function that can be subsequently used to direct the agent to promising states in the match. In this thesis we will discuss existing methods and present a general approach on how to construct such an evaluation function. Each topic is discussed in a modular fashion and evaluated along the lines of quality and efficiency, resulting in a strong agent.:Introduction Game Playing Evaluation Functions I - Aggregation Evaluation Functions II - Features General Evaluation Related Work Discussion

Ce mémoire rassemble des éléments d'optimisation, d'apprentissage profond et de contrôle optimal afin de répondre aux problématiques d'apprentissage et de planification dans le contexte des systèmes dynamiques en temps continu. Deux approches générales sont explorées. D'abord, une approche basée sur la méthode du maximum de vraisemblance est présentée. Ici, les trajectoires ``d'entrainement'' sont échantillonnées depuis la dynamique réelle, et à partir de celles-ci un modèle de prédiction des états observés est appris. Une fois que l'apprentissage est terminé, le modèle est utilisé pour la planification, en utilisant la dynamique de l'environnement et une fonction de coût pour construire un programme non linéaire, qui est par la suite résolu pour trouver une séquence de contrôle optimal. Ensuite, une approche de bout en bout est proposée, dans laquelle la tâche d'apprentissage de modèle dynamique et celle de planification se déroulent simultanément. Ceci est illustré dans le cadre d'un problème d'apprentissage par imitation, où le modèle est mis à jour en rétropropageant le signal de perte à travers l'algorithme de planification. Grâce au fait que l'entrainement est effectué de bout en bout, cette technique pourrait constituer un sous-module de réseau de neurones de plus grande taille, et pourrait être utilisée pour fournir un biais inductif en faveur des comportements optimaux dans le contexte de systèmes dynamiques en temps continu. Ces méthodes sont toutes les deux conçues pour fonctionner avec des modèles d'équations différentielles ordinaires paramétriques et neuronaux. Également, inspiré par des applications réelles pertinentes, un large recueil de systèmes dynamiques et d'optimiseurs de trajectoire, nommé Myriad, est implémenté; les algorithmes sont testés et comparés sur une variété de domaines de la suite Myriad. / This thesis brings together elements of optimization, deep learning and optimal control to study the challenge of learning and planning in continuous-time dynamical systems. Two general approaches are explored. First, a maximum likelihood approach is presented, in which training trajectories are sampled from the true dynamics, and a model is learned to accurately predict the state observations. After training is completed, the learned model is then used for planning, by using the dynamics and cost function to construct a nonlinear program, which can be solved to find a sequence of optimal controls. Second, a fully end-to-end approach is proposed, in which the tasks of model learning and planning are performed simultaneously. This is demonstrated in an imitation learning setting, in which the model is updated by backpropagating the loss signal through the planning algorithm itself. Importantly, because it can be trained in an end-to-end fashion, this technique can be included as a sub-module of a larger neural network, and used to provide an inductive bias towards behaving optimally in a continuous-time dynamical system. Both the maximum likelihood and end-to-end methods are designed to work with parametric and neural ordinary differential equation models. Inspired by relevant real-world applications, a large repository of dynamical systems and trajectory optimizers, named Myriad, is also implemented. The algorithms are tested and compared on a variety of domains within the Myriad suite.

La trisomie 21 (T21), le Syndrome X Fragile (SXF), la Sclérose tubéreuse de Bourneville (STB) et les mutations SYNGAP1 sont causés par des dysfonctionnements des voies moléculaires qui entraînent notamment un déséquilibre dans l’excitation et l’inhibition de l’activité neuronale qui aurait des impacts sur le développement et le fonctionnement du cerveau. Toutefois, il est difficile de faire le pont entre les déséquilibres moléculaires observés dans les modèles animaux et les particularités structurelles, fonctionnelles et cognitives observées dans ces syndromes chez l’humain. À notre connaissance, peu d’études ont comparé différents syndromes génétiques sur les processus sensoriels, l’apprentissage de base ou encore leurs caractéristiques comportementales en utilisant des paradigmes similaires et translationnels, permettant de mieux comprendre leurs particularités. Le premier volet de cette thèse vise à identifier si l’activité électroencéphalographique serait un biomarqueur adéquat représentant les altérations neurobiologiques tant des processus sensoriels que d’apprentissage chez les humains présentant ces syndromes. L’étude #1 avait comme objectif de décrire le traitement sensoriel auditif, comme il s’agit d’un processus élémentaire, et ce, chez les mutations SYNGAP1 qui représentent une condition génétique encore peu étudiée chez l’humain. Les résultats ont d’ailleurs permis d’identifier une diminution de la synchronisation de phase et une augmentation de la puissance dans la bande gamma qui distinguent cette condition génétique tant des participants sans DI que de la T21. Toujours dans l’esprit d’identifier des biomarqueurs électroencéphalographiques, mais cette fois au niveau d’un processus cognitif de base, l’étude #2 avait pour objectif de comparer tous ces syndromes dans un paradigme de suppression neuronale (SN) afin de vérifier la présence de SN et de comparer l’apprentissage de base chez ces populations. Les résultats ont identifiés que la T21 et le SXF présentaient tous les deux un patron de SN et que le SXF présentait relativement une plus forte habituation indiquant des particularités spécifiques selon les syndromes. Le deuxième volet, davantage clinique, permet de comparer les profils comportementaux associés au fonctionnement adaptatif entre les syndromes et à décrire les pistes d’intervention existantes. L’étude #3 a notamment mis en évidence que le QI et les symptômes de TDAH sont associés au fonctionnement adaptatif auprès de ces différents syndromes dont le SXF et la STB. Cet article a aussi permis de décrire les profils comportementaux de ces mêmes conditions en révélant davantage de difficultés rapportées chez les individus présentant un SXF, alors que la T21 présentait moins de particularités cliniques au niveau comportemental. Enfin, l’article #4 a mis en lumière diverses interventions utilisées auprès de la population présentant une DI notamment des stratégies cognitivo-comportementales et compensatoires. Cette thèse permet donc de dresser un portrait spécifique de ces syndromes génétiques concernant leur signature électrophysiologique lors du traitement sensoriel et de l’apprentissage ainsi que sur le plan des comorbidités comportementales et de leur relation avec le fonctionnement adaptatif, pour ensuite aborder les interventions actuelles en DI. Les diverses particularités identifiées à plusieurs niveaux ont permis de générer des suggestions pouvant guider certaines interventions futures. / Down syndrome (DS), Fragile X syndrome (FXS), Tuberous sclerosis complex (TSC) and SYNGAP1 mutations are caused by dysfunctions of the molecular pathways which lead among others to an imbalance in excitation and inhibition of the neuronal activity that would impact the brain development and its functioning. However, it is difficult to directly bridge the gap between the molecular imbalances observed in animal models with the structural, functional and cognitive characteristics observed in human with these syndromes. To our knowledge, few studies have compared those different genetic syndromes on sensory processing, basic learning or on their behavioural issues using similar and translational paradigms then allowing a better understanding of their specificities. The first part of this thesis aims to identify whether electroencephalographic activity would be an adequate biomarker representing neurobiological alterations both in sensory processing and learning in humans with these syndromes. The goal of study #1 was to describe auditory sensory processing, as a very first basic process, in SYNGAP1 mutations being a genetic condition still little studied in humans. Results showed a decrease in phase synchronization and an increase in the power of gamma band which distinguish this genetic condition both from participants without ID and from DS. Still in order to identify electroencephalographic biomarkers, but this time at a basic cognitive level, study #2 aimed to compare all these syndromes in a repetition suppression (RS) paradigm in order to observe the presence of RS and compare basic learning in these populations. The results identified a RS pattern in both DS and FXS. FXS also exhibited relatively higher habituation then indicating specific features according to the syndrome. The second part, addressing clinical aspects, permits to compare the behavioural profiles associated with adaptive functioning between syndromes and to describe existing interventions on ID population. Study #3 notably highlighted that IQ and ADHD symptoms are associated with adaptive functioning especially in FXS and TSC. This article also made it possible to describe the behavioural profiles of these syndromes, revealing more difficulties reported in individuals with FXS, while DS presented fewer behavioural issues. Finally, article #4 highlighted various interventions used with ID population, notably cognitive-behavioural and compensatory strategies. This thesis therefore makes it possible to gain a better understanding of these genetic syndromes concerning their electrophysiological signature during sensory processing and learning as well as in terms of behavioural comorbidities and their relationship with adaptive functioning, to then address current ID interventions. These different syndromic particularities identified at several levels made it possible to generate suggestions that could guide future interventions in this field.

From the initial computing machines, Colossus of 1943 and ENIAC of 1945, to modern high-performance data centers and Internet of Things (IOTs), four design goals, i.e., high-performance, energy-efficiency, resource utilization, and ease of programmability, have remained a beacon of development for the computing industry. During this period, the computing industry has exploited the advantages of technology scaling and microarchitectural enhancements to achieve these goals. However, with the end of Dennard scaling, these techniques have diminishing energy and performance advantages. Therefore, it is necessary to explore alternative techniques for satisfying the computational and energy requirements of modern applications. Towards this end, one promising technique is analyzing and surrendering the strict notion of correctness in various layers of the computation stack. Most modern applications across the computing spectrum---from data centers to IoTs---interact and analyze real-world data and take decisions accordingly. These applications are broadly classified as Recognition, Mining, and Synthesis (RMS). Instead of producing a single golden answer, these applications produce several feasible answers. These applications possess an inherent error-resilience to the inexactness of processed data and corresponding operations. Utilizing these applications' inherent error-resilience, the paradigm of Approximate Computing relaxes the strict notion of computation correctness to realize high-performance and energy-efficient systems with acceptable quality outputs. The prior works on circuit-level approximations have mainly focused on Application-specific Integrated Circuits (ASICs). However, ASIC-based solutions suffer from long time-to-market and high-cost developing cycles. These limitations of ASICs can be overcome by utilizing the reconfigurable nature of Field Programmable Gate Arrays (FPGAs). However, due to architectural differences between ASICs and FPGAs, the utilization of ASIC-based approximation techniques for FPGA-based systems does not result in proportional performance and energy gains. Therefore, to exploit the principles of approximate computing for FPGA-based hardware accelerators for error-resilient applications, FPGA-optimized approximation techniques are required. Further, most state-of-the-art approximate arithmetic operators do not have a generic approximation methodology to implement new approximate designs for an application's changing accuracy and performance requirements. These works also lack a methodology where a machine learning model can be used to correlate an approximate operator with its impact on the output quality of an application. This thesis focuses on these research challenges by designing and exploring FPGA-optimized logic-based approximate arithmetic operators. As multiplication operation is one of the computationally complex and most frequently used arithmetic operations in various modern applications, such as Artificial Neural Networks (ANNs), we have, therefore, considered it for most of the proposed approximation techniques in this thesis. The primary focus of the work is to provide a framework for generating FPGA-optimized approximate arithmetic operators and efficient techniques to explore approximate operators for implementing hardware accelerators for error-resilient applications. Towards this end, we first present various designs of resource-optimized, high-performance, and energy-efficient accurate multipliers. Although modern FPGAs host high-performance DSP blocks to perform multiplication and other arithmetic operations, our analysis and results show that the orthogonal approach of having resource-efficient and high-performance multipliers is necessary for implementing high-performance accelerators. Due to the differences in the type of data processed by various applications, the thesis presents individual designs for unsigned, signed, and constant multipliers. Compared to the multiplier IPs provided by the FPGA Synthesis tool, our proposed designs provide significant performance gains. We then explore the designed accurate multipliers and provide a library of approximate unsigned/signed multipliers. The proposed approximations target the reduction in the total utilized resources, critical path delay, and energy consumption of the multipliers. We have explored various statistical error metrics to characterize the approximation-induced accuracy degradation of the approximate multipliers. We have also utilized the designed multipliers in various error-resilient applications to evaluate their impact on applications' output quality and performance. Based on our analysis of the designed approximate multipliers, we identify the need for a framework to design application-specific approximate arithmetic operators. An application-specific approximate arithmetic operator intends to implement only the logic that can satisfy the application's overall output accuracy and performance constraints. Towards this end, we present a generic design methodology for implementing FPGA-based application-specific approximate arithmetic operators from their accurate implementations according to the applications' accuracy and performance requirements. In this regard, we utilize various machine learning models to identify feasible approximate arithmetic configurations for various applications. We also utilize different machine learning models and optimization techniques to efficiently explore the large design space of individual operators and their utilization in various applications. In this thesis, we have used the proposed methodology to design approximate adders and multipliers. This thesis also explores other layers of the computation stack (cross-layer) for possible approximations to satisfy an application's accuracy and performance requirements. Towards this end, we first present a low bit-width and highly accurate quantization scheme for pre-trained Deep Neural Networks (DNNs). The proposed quantization scheme does not require re-training (fine-tuning the parameters) after quantization. We also present a resource-efficient FPGA-based multiplier that utilizes our proposed quantization scheme. Finally, we present a framework to allow the intelligent exploration and highly accurate identification of the feasible design points in the large design space enabled by cross-layer approximations. The proposed framework utilizes a novel Polynomial Regression (PR)-based method to model approximate arithmetic operators. The PR-based representation enables machine learning models to better correlate an approximate operator's coefficients with their impact on an application's output quality.:1. Introduction 1.1 Inherent Error Resilience of Applications 1.2 Approximate Computing Paradigm 1.2.1 Software Layer Approximation 1.2.2 Architecture Layer Approximation 1.2.3 Circuit Layer Approximation 1.3 Problem Statement 1.4 Focus of the Thesis 1.5 Key Contributions and Thesis Overview 2. Preliminaries 2.1 Xilinx FPGA Slice Structure 2.2 Multiplication Algorithms 2.2.1 Baugh-Wooley’s Multiplication Algorithm 2.2.2 Booth’s Multiplication Algorithm 2.2.3 Sign Extension for Booth’s Multiplier 2.3 Statistical Error Metrics 2.4 Design Space Exploration and Optimization Techniques 2.4.1 Genetic Algorithm 2.4.2 Bayesian Optimization 2.5 Artificial Neural Networks 3. Accurate Multipliers 3.1 Introduction 3.2 Related Work 3.3 Unsigned Multiplier Architecture 3.4 Motivation for Signed Multipliers 3.5 Baugh-Wooley’s Multiplier 3.6 Booth’s Algorithm-based Signed Multipliers 3.6.1 Booth-Mult Design 3.6.2 Booth-Opt Design 3.6.3 Booth-Par Design 3.7 Constant Multipliers 3.8 Results and Discussion 3.8.1 Experimental Setup and Tool Flow 3.8.2 Performance comparison of the proposed accurate unsigned multiplier 3.8.3 Performance comparison of the proposed accurate signed multiplier with the state-of-the-art accurate multipliers 3.8.4 Performance comparison of the proposed constant multiplier with the state-of-the-art accurate multipliers 3.9 Conclusion 4. Approximate Multipliers 4.1 Introduction 4.2 Related Work 4.3 Unsigned Approximate Multipliers 4.3.1 Approximate 4 × 4 Multiplier (Approx-1) 4.3.2 Approximate 4 × 4 Multiplier (Approx-2) 4.3.3 Approximate 4 × 4 Multiplier (Approx-3) 4.4 Designing Higher Order Approximate Unsigned Multipliers 4.4.1 Accurate Adders for Implementing 8 × 8 Approximate Multipliers from 4 × 4 Approximate Multipliers 4.4.2 Approximate Adders for Implementing Higher-order Approximate Multipliers 4.5 Approximate Signed Multipliers (Booth-Approx) 4.6 Results and Discussion 4.6.1 Experimental Setup and Tool Flow 4.6.2 Evaluation of the Proposed Approximate Unsigned Multipliers 4.6.3 Evaluation of the Proposed Approximate Signed Multiplier 4.7 Conclusion 5. Designing Application-specific Approximate Operators 5.1 Introduction 5.2 Related Work 5.3 Modeling Approximate Arithmetic Operators 5.3.1 Accurate Multiplier Design 5.3.2 Approximation Methodology 5.3.3 Approximate Adders 5.4 DSE for FPGA-based Approximate Operators Synthesis 5.4.1 DSE using Bayesian Optimization 5.4.2 MOEA-based Optimization 5.4.3 Machine Learning Models for DSE 5.5 Results and Discussion 5.5.1 Experimental Setup and Tool Flow 5.5.2 Accuracy-Performance Analysis of Approximate Adders 5.5.3 Accuracy-Performance Analysis of Approximate Multipliers 5.5.4 AppAxO MBO 5.5.5 ML Modeling 5.5.6 DSE using ML Models 5.5.7 Proposed Approximate Operators 5.6 Conclusion 6. Quantization of Pre-trained Deep Neural Networks 6.1 Introduction 6.2 Related Work 6.2.1 Commonly Used Quantization Techniques 6.3 Proposed Quantization Techniques 6.3.1 L2L: Log_2_Lead Quantization 6.3.2 ALigN: Adaptive Log_2_Lead Quantization 6.3.3 Quantitative Analysis of the Proposed Quantization Schemes 6.3.4 Proposed Quantization Technique-based Multiplier 6.4 Results and Discussion 6.4.1 Experimental Setup and Tool Flow 6.4.2 Image Classification 6.4.3 Semantic Segmentation 6.4.4 Hardware Implementation Results 6.5 Conclusion 7. A Framework for Cross-layer Approximations 7.1 Introduction 7.2 Related Work 7.3 Error-analysis of approximate arithmetic units 7.3.1 Application Independent Error-analysis of Approximate Multipliers 7.3.2 Application Specific Error Analysis 7.4 Accelerator Performance Estimation 7.5 DSE Methodology 7.6 Results and Discussion 7.6.1 Experimental Setup and Tool Flow 7.6.2 Behavioral Analysis 7.6.3 Accelerator Performance Estimation 7.6.4 DSE Performance 7.7 Conclusion 8. Conclusions and Future Work

Neuronen sind bei der Informationsübertragung des zentralen Nervensystems von entscheidender Bedeutung. Ihre Aktivität liegt der Signalverarbeitung und höheren kognitiven Prozessen zugrunde. Neuronen sind in den extrazellulären Raum eingebettet, der mehrere Teilchen, darunter auch Ionen, enthält. Ionenkonzentrationen sind nicht statisch. Intensive neuronale Aktivität kann intrazelluläre und extrazelluläre Ionenkonzentrationen verändern. In dieser Arbeit untersuche ich das Wechselspiel zwischen neuronaler Aktivität und der Dynamik der Ionenkonzentrationen. Dabei konzentriere ich mich hauptsächlich auf extrazelluläre Kalium- und intrazelluläre Natriumkonzentrationen. Mit Hilfe der Theorie dynamischer Systeme zeige ich, wie moderate Änderungen dieser Ionenkonzentrationen die neuronale Aktivität qualitativ verändern können, wodurch sich möglicherweise die Signalverarbeitung verändert. Dann modelliere ich ein leitfähigkeitsbasiertes neuronales Netzwerk mit Spikes. Das Modell sagt voraus, dass eine moderate Änderung der Konzentrationen, die einen Mikroschaltkreis von Neuronen umgeben, die Leistungsspektraldichte der Populationsaktivität verändern könnte. Insgesamt unterstreicht diese Arbeit die Bedeutung der Dynamik der Ionenkonzentrationen für das Verständnis neuronaler Aktivität auf langen Zeitskalen und liefert technische Erkenntnisse darüber, wie das Zusammenspiel zwischen ihnen modelliert und analysiert werden kann. / Neurons are essential in the information transfer mechanisms of the central nervous system. Their activity underlies both basic signal processing, and higher cognitive processes. Neurons are embedded in the extracellular space, which contains multiple particles, including ions which are vital to their functioning. Ionic concentrations are not static, intense neuronal activity alters the intracellular and extracellular ionic concentrations which in turn affect neuronal functioning. In this thesis, I study the interplay between neuronal activity and ionic concentration dynamics. I focus specifically on the extracellular potassium and intracellular sodium concentrations. Using dynamical systems theory, I illustrate how moderate changes in these ionic concentrations can qualitatively change neuronal activity, potentially altering signal processing. I then model a conductance-based spiking neural network. The model predicts that a moderate change in the concentrations surrounding a microcircuit of neurons could modify the power spectral density of the population activity. Altogether, this work highlights the need to consider ionic concentration dynamics to understand neuronal activity on long time scales and provides technical insights on how to model and analyze the interplay between them.

In the context of evolutionary robotics and neurorobotics, artificial neural networks, used as controllers for animats, are examined to identify principles of neuro-control, network organization, the interaction between body and control, and other likewise properties. Before such an examination can take place, suitable neuro-controllers have to be identified. A promising and widely used technique to search for such networks are evolutionary algorithms specifically adapted for neural networks. These allow the search for neuro-controllers with various network topologies directly on physically grounded (simulated) animats. This neuro-evolution approach works well for small neuro-controllers and has lead to interesting results. However, due to the exponentially increasing search space with respect to the number of involved neurons, this approach does not scale well with larger networks. This scaling problem makes it difficult to find non-trivial, larger networks, that show interesting properties. In the context of this thesis, networks of this class are called mid-scale networks, having between 50 and 500 neurons. Searching for networks of this class involves very large search spaces, including all possible synaptic connections between the neurons, the bias terms of the neurons and (optionally) parameters of the neuron model, such as the transfer function, activation function or parameters of learning rules. In this domain, most evolutionary algorithms are not able to find suitable, non-trivial neuro-controllers in feasible time. To cope with this problem and to shift the frontier for evolvable network topologies a bit further, a novel evolutionary method has been developed in this thesis: the Interactively Constrained Neuro-Evolution method (ICONE). A way to approach the problem of increasing search spaces is the introduction of measures that reduce and restrict the search space back to a feasible domain. With ICONE, this restriction is realized with a unified, extensible and highly adaptable concept: Instead of evolving networks freely, networks are evolved within specifically designed constraint masks, that define mandatory properties of the evolving networks. These constraint masks are defined primarily using so called functional constraints, that actively modify a neural network to enforce the adherence of all required limitations and assumptions. Consequently, independently of the mutations taking place during evolution, the constraint masks repair and readjust the networks so that constraint violations are not able to evolve. Such functional constraints can be very specific and can enforce various network properties, such as symmetries, structure reuse, connectivity patterns, connectivity density heuristics, synaptic pathways, local processing assemblies, and much more. Constraint masks therefore describe a narrow, user defined subset of the parameter space -- based on domain knowledge and user experience -- that focuses the search on a smaller search space leading to a higher success rate for the evolution. Due to the involved domain knowledge, such evolutions are strongly biased towards specific classes of networks, because only networks within the defined search space can evolve. This, surely, can also be actively used to lead the evolution towards specific solution approaches, allowing the experimenter not only to search for any upcoming solution, but also to confirm assumptions about possible solutions. This makes it easier to investigate specific neuro-control principles, because the experimenter can systematically search for networks implementing the desired principles, simply by using suitable constraints to enforce them. Constraint masks in ICONE are built up by functional constraints working on so called neuro-modules. These modules are used to structure the networks, to define the scope for constraints and to simplify the reuse of (evolved) neural structures. The concept of functional, constrained neuro-modules allows a simple and flexible way to construct constraint masks and to inherit constraints when neuro-modules are reused or shared. A final cornerstone of the ICONE method is the interactive control of the evolution process, that allows the adaptation of the evolution parameters and the constraint masks to guide evolution towards promising domains and to counteract undesired developments. Due to the constraint masks, this interactive guidance is more effective than the adaptation of the evolution parameters alone, so that the identification of promising search space regions becomes easier. This thesis describes the ICONE method in detail and shows several applications of the method and the involved features. The examples demonstrate that the method can be used effectively for problems in the domain of mid-scale networks. Hereby, as effects of the constraint masks and the herewith reduced complexity of the networks, the results are -- despite their size -- often easy to comprehend, well analyzable and easy to reuse. Another benefit of constraint masks is the ability to deliberately search for very specific network configurations, which allows the effective and systematic exploration of distinct variations for an evolution experiment, simply by changing the constraint masks over the course of multiple evolution runs. The ICONE method therefore is a promising novel evolution method to tackle the problem of evolving mid-scale networks, pushing the frontier of evolvable networks a bit further. This allows for novel evolution experiments in the domain of neurorobotics and evolutionary robotics and may possibly lead to new insights into neuro-dynamical principles of animat control.

Menschliche Hirnsignale von der Kopfhaut bieten einen Einblick in die neuronalen Prozesse, denen Wahrnehmung, Denken und Verhalten zugrunde liegen. Rhythmen, die historisch den Grundstein für die Erforschung großflächiger Hirnsignale legten, sind ein häufiges Zeichen neuronaler Koordination, und damit von weitem Interesse für die kognitiven, systemischen und komputationalen Neurowissenschaften. Typischen Messungen von Rhythmizität fehlt es jedoch an Details, z. B. wann und wie lange Rhythmen auftreten. Darüber hinaus weisen neuronale Zeitreihen zahlreiche dynamische Muster auf, von denen nur einige rhythmisch erscheinen. Obwohl aperiodischen Beiträgen traditionell der Status irrelevanten „Rauschens“ zugeschrieben wird, attestieren neuere Erkenntnisse ihnen ebenfalls eine Signalrolle in Bezug auf latente Hirndynamik. Diese kumulative Dissertation fasst Projekte zusammen, die darauf abzielen, rhythmische und aperiodische Beiträge zum menschlichen Elektroenzephalogramm (EEG) methodisch zu dissoziieren, und ihre Relevanz für die flexible Wahrnehmung zu untersuchen. Projekt 1 ermittelt insbesondere die Notwendigkeit und Durchführbarkeit der Trennung rhythmischer von aperiodischer Aktivität in kontinuierlichen Signalen. Projekt 2 kehrt diese Perspektive um und prüft Multiscale Entropy als Index für die Unregelmäßigkeit von Zeitreihen. Diese Arbeit weist auf methodische Probleme in der klassischen Messung zeitlicher Unregelmäßigkeit hin, und schlägt Lösungen für zukünftige Anwendungen vor. Abschließend untersucht Projekt 3 die neurokognitive Relevanz rhythmischer und aperiodischer Zustände. Anhand eines parallelen multimodalen EEG-fMRT-Designs mit gleichzeitiger Pupillenmessung liefert dieses Projekt erste Hinweise dafür, dass erhöhte kognitive Anforderungen Hirnsignale von einem rhythmischen zu einem unregelmäßigen Regime verschieben und impliziert gleichzeitige Neuromodulation und thalamische Aktivierung in diesem Regimewechsel. / Non-invasive signals recorded from the human scalp provide a window on the neural dynamics that shape perception, cognition and action. Historically motivating the assessment of large-scale network dynamics, rhythms are a ubiquitous sign of neural coordination, and a major signal of interest in the cognitive, systems, and computational neurosciences. However, typical descriptions of rhythmicity lack detail, e.g., failing to indicate when and for how long rhythms occur. Moreover, neural times series exhibit a wealth of dynamic patterns, only some of which appear rhythmic. While aperiodic contributions are traditionally relegated to the status of irrelevant ‘noise’, they may be informative of latent processing regimes in their own right. This cumulative dissertation summarizes and discusses work that (a) aims to methodologically dissociate rhythmic and aperiodic contributions to human electroencephalogram (EEG) signals, and (b) probes their relevance for flexible cognition. Specifically, Project 1 highlights the necessity, feasibility and limitations of dissociating rhythmic from aperiodic activity at the single-trial level. Project 2 inverts this perspective, and examines the utility of multi-scale entropy as an index for the irregularity of brain dynamics, with a focus on the relation to rhythmic and aperiodic descriptions. By highlighting prior biases and proposing solutions, this work indicates future directions for measurements of temporal irregularity. Finally, Project 3 examines the neurocognitive relevance of rhythmic and aperiodic regimes with regard to the neurophysiological context in which they may be engaged. Using a parallel multi-modal EEG-fMRI design with concurrent pupillometry, this project provides initial evidence that elevated demands shift cortical dynamics from a rhythmic to an irregular regime; and implicates concurrent phasic neuromodulation and subcortical thalamic engagement in these regime shifts.

Aufgabe der elektrophysiologischen Studie zur Charakterisierung der neuronalen hypothalamischen Plastizität beim Haushuhn war es, den Einfluss des Alters sowie GABAerger Substanzen auf die Feuerrate und die Temperatursensitivität (thermischer Koeffizient: TC) von Hypothalamusneuronen mittels extrazellulärer Ableitungen in Hirnschnitten zu untersuchen. Im Vergleich zu adulten Vögeln und Säugetieren wurde bei juvenilen Hühnern eine hohe neuronale Kältesensitivität nachgewiesen, die offensichtlich eine spezifische Eigenschaft juveniler Vögel ist. Die Ontogenese der neuronalen hypothalamischen Thermosensitivität ist deutlich artspezifisch. Einige Neurone wiesen eine inherente Kältesensitivität auf. Eine mögliche zentrale Rolle kältesensitiver Neurone im Rahmen der Thermoregulation juveniler Hühner wurde postuliert. Muscimol und Baclofen hemmen signifikant die Feuerrate der Hypothalamusneurone, unabhängig von der jeweiligen Thermosensitivität. Demgegenüber bewirken Bicucullin und CGP35348 einem Anstieg der Feuerrate. Nur bei kältesensitiven Neuronen wurde der TC signifikant durch GABAB-Rezeptor-Liganden verändert (signifikant erhöht durch Baclofen und durch CGP35348 gehemmt). Der Effekt von Muscimol und Baclofen auf Feuerrate und TC wurde durch Co-Perfusion mit einer 10-fach höheren Konzentration der entsprechenden Antagonisten Bicucullin und CGP35348 aufgehoben. Der wesentliche GABAerge Einfluss auf thermosensitive und –insensitive Hypothalamusneurone ist mit dem bei Säugetieren nachgewiesenen vergleichbar. Der einzige Unterschied betrifft die GABAB-Rezeptor vermittelte Änderung des TC. Beim Hühnerküken betraf dies die kältesensitiven und beim Säugetier die wärmesensitiven Neurone. Der grundlegende Mechanismus der GABAergen Beeinflussung thermosensitiver und –insensitiver Neurone scheint einen älteren evolutionären Ursprung zu haben. Eine funktionelle Rolle GABAerger Substanzen im Rahmen der zentralen Kontrolle der Körpertemperatur beim Vogel ist möglich. / In the present electrophysiological studies, characterization of neuronal hypothalamic plasticity in the chicken aims to investigate the influence of age during development by extracellular recordings. High neuronal cold sensitivity has been found in juvenile chicken in contrast to adult mammals and birds. High hypothalamic cold sensitivity seems to be a specific characteristic feature in juvenile birds. Between species a species specificity of the early development of neuronal hypothalamic thermosensitivity could be clearly demonstrated. Existence of inherent nature to a certain degree suggests a possible thermoregulatory role of cold-sensitive neurons in chicken. The effects of the GABAergic substances on neuronal tonic activity (firing rate) and temperature sensitivity (temperature coefficient) in hypothalamic neurons have been examined. Muscimol and baclofen in equimolar concentrations significantly inhibited tonic activity, regardless of their type of thermosensitivity. In contrast bicuculline and CGP 35348 increased firing rate. Temperature coefficient was significantly changed by ligands of GABAB receptors, restricted to cold-sensitive neurons. The TC was significantly increased by baclofen and significantly decreased by CGP 35348. Effects of muscimol and baclofen on firing rate and TC were prevented by co-perfusion of appropriate antagonists bicuculline and CGP 35348, respectively in tenfold higher concentration. Thus the main effects of GABA in chicken are similar with that described in mammals. The only difference is in respect of the GABAB receptors mediated change restricted to cold-sensitive neurons in chicken but in mammals only seen in warm-sensitive neurons. However, the results indicate that the fundamental mechanism of GABAergic influence in chicken are conserved during evolution. The response of hypothalamic neurons to temperature changes suggest a possible functional role of GABAergic substances in the control of body temperature in birds.

Die Wirkung von zwei endogen neuroprotektiven Substanzen, Melatonin und 17 beta-Estradiol wurde an drei Caspase-abhängigen, apoptotischen, aber Exzitotoxin-unabhängigen Schadensmodellen an neuronalen Primärkulturen untersucht und mit der bei vorwiegend nekrotischen Schadensmodellen verglichen. Es zeigten sich eine Abhängigkeit des neuroprotektiven Potentials von der Art des Zelluntergangs sowie unterschiedliche Mechanismen der Neuroprotektion. Melatonin wirkte in allen drei apoptischen Modellen nicht neuroprotektiv, sondern verstärkte die Schädigung der Neurone noch, während partiell gegen die OGD-induzierte Nekrose (OGD, engl. Oxygen glucose deprivation, kombinierter Sauerstoff- und Glukoseentzug) kortikaler Neurone Schutz erzielt wurde. Der Einsatz des endogenen neuroprotektiven Faktors Melatonin als Therapeutikum ist möglicherweise nur bei neurodegenerativen Erkrankungen mit exzitotoxischer Schädigung durch Glutamat oder oxidativem Stress wie bei Epilepsie oder dem Schlaganfall durch Ischämie sinnvoll. Die fehlende bzw. potenzierenden Wirkung von Melatonin bei neuronaler Apoptose in vitro, stellt jedoch einen therapeutischen Erfolg bei der Behandlung der mit apoptotischer Schädigung einhergehenden Alzheimer'schen Erkrankung in Frage. Bei klinischer Anwendung ist auch der von uns erhobene Befund zu beachten, dass in vitro native neuronale Zellen durch Melatonin geschädigt werden. 17 beta-Estradiol wirkte sowohl bei nekrotischer als auch bei apoptotischer Zellschädigung. Dabei zeigten sich wesentliche Unterschiede in den Mechanismen der Neuroprotektion und in der Ansprechbarkeit verschiedener Regionen des Gehirns. Schutz vor Apoptose konnte nur durch eine Langzeitvorbehandlung (20 h) in septalen und hippokampalen Kulturen, nicht jedoch in kortikalen Kulturen beobachtet werden. Dieser Effekt liess sich durch Rezeptorantagonisten, Proteinsynthesehemmung sowie durch Hemmung der Phosphoinositol-3-Kinase blockieren. Eine Kurzzeitbehandlung war gegen Apoptose nicht wirksam, zeigte gegen OGD und Glutamattoxizität jedoch neuroprotektives Potential. Dieser Effekt liess sich nicht antagonisieren, so dass hier ein direkter antioxidativer Mechanismus wahrscheinlich erscheint. Die antiapoptotische Wirkung in septalen und hippokampalen Kulturen korrelierte mit einer höheren Dichte des Estrogenrezeptors-alpha und einer erhöhten Expression antiapoptotischer Proteine in diesen Regionen. Da bei der Alzheimer'schen Erkrankung der Kortex betroffen ist, könnte der fehlende Effekt von 17 beta-Estradiol in kortikalen Neuronen sowohl auf die neuronale Apoptose als auch auf die Proteinexpression von Bcl-2 und Bcl-xL möglicherweise auf experimenteller Basis erklären, warum eine langfristige Estrogentherapie bei Frauen mit milder bis moderater Alzeimer'scher Erkrankung den Progress der Erkrankung nicht aufhalten konnte (Mulnard et al. 2000). / The neuroprotective effect of melatonin and 17 beta-estradiol has been evaluated in several in vitro models of neuronal apoptosis and necrosis. Melatonin was not neuroprotective in three models of apoptosis but showed a pro-apoptotic effect in primary cortical neurons. Melatonin revealed to damage naïve neurons, too. Partial protection was observed against necrotic neurodegeneration after oxygen-glucose deprivation (OGD). The use of melatonin as a therapeutic agent might be of interest in neurodegenerative diseases with excitotoxic damage like epilepsia or ischemia, but is questioned in case of apoptotic neurodegeneration. 17 beta-estradiol was neuroprotectiv in both necrotic and apoptotic neurodegeneration. Differences in the mechanism of neuroprotetion and in the efficacy in different regions of the brain were observed. A neuroprotective effect was visible only in hippocampal and septal cultures if 17 beta-estradiol was applied 20 h prior (long term pre-treatment) but not in cortical neurons. This effect correlates with an increased density of estrogen receptor-alpha and an increased expression of anti-apoptotic proteins like Bcl-2 and Bcl-xL in these regions. These effect could be blocked with receptor antagonists, protein synthesis inhibitors and an inhibitor of the phosphatidylinositol 3-kinase. A short term pre-treatment revealed a receptor independent neuroprotective potential against OGD and glutamate toxicity. The failure of 17 beta-estradiol to protect cortical neurons against apoptosis could be an experimental basis to understand, why a long lasting treatment with estrogens of women with mild to moderate Alzheimer´s disease failed to inhibit the progress of the illness (Mulnard et al., 2000)