With increasing amounts of data being collected in various fields of neuroscience, there is a growing need for robust techniques for the analysis of this information. This thesis focuses on the evaluation and development of quantitative frameworks for the analysis and classification of neuronal data from a variety of contexts. Firstly, I investigate methods for analysing spontaneous neuronal network activity recorded on microelectrode arrays (MEAs). I perform an unbiased evaluation of the existing techniques for detecting ‘bursts’ of neuronal activity in these types of recordings, and provide recommendations for the robust analysis of bursting activity in a range of contexts using both existing and adapted burst detection methods. These techniques are then used to analyse bursting activity in novel recordings of human induced pluripotent stem cell-derived neuronal networks. Results from this review of burst analysis methods are then used to inform the development of a framework for characterising the activity of neuronal networks recorded on MEAs, using properties of bursting as well as other common features of spontaneous activity. Using this framework, I examine the ontogeny of spontaneous network activity in in vitro neuronal networks from various brain regions, recorded on both single and multi-well MEAs. I also develop a framework for classifying these recordings according to their network type, based on quantitative features of their activity patterns. Next, I take a multi-view approach to classifying neuronal cell types using both the morphological and electrophysiological features of cells. I show that a number of multi-view clustering algorithms can more reliably differentiate between neuronal cell types in two existing data sets, compared to single-view clustering techniques applied to either the morphological or electrophysiological ‘view’ of the data, or a concatenation of the two views. To close, I examine the properties of the cell types identified by these methods.

Os interneurônios do bulbo olfatório são elementos chave para o entendimento do processamento de odores. O papel funcional desses neurônios ainda não é bem compreendido, em especial o papel da célula periglomerular (PG). O presente trabalho consiste em construir um modelo biologicamente plausível da célula PG e investigar os efeitos dessa célula em conjunto com modelos da célula mitral e da célula granular. Esses modelos são acoplados através de conexões sinápticas inspiradas nas conexões existentes no bulbo olfatório, formando uma pequena rede simplificada. A rede é usada para analisar o efeito da inibição inicial da célula mitral por parte da célula PG e os mecanismos que podem influenciar o padrão de atividade da célula mitral. Através deste estudo, verifica-se que a célula PG pode influenciar na frequência, no tempo de disparo e gerar atrasos na propagação do potencial da célula mitral, agindo como um mecanismo de controle nas camadas iniciais do processamento de odores do bulbo olfatório. / Interneurons of the olfactory bulb are key elements for understanding odor processing. The functional role of these cells are not yet well understood, in particular the role of periglomerular cell (PG). This work aims at constructing a biologically plausible model of the PG cell to study effects of the coupling of this cell with model of mitral and granule cells of the olfactory bulb. Single cell models of these three cell types coupled by synaptic connections inspired on existing connections in the olfactory bulb, constituting a small and simple network. This network is used to investigate the effect of early lateral inhibition of the mitral cell by PG cell and the mechanisms witch can influence the output activity pattern of mitral cell. The study shows that the PG cell may influence the spike frequency and the spike timing of the mitral cell, as well as provoke delays in the propagation of action potential along this cell. Therefore, the PG cell may act as a control mechanism in the early odor processing stages in the olfactory bulb.

A ocorrência de status epileticus (SE) desencadeia algumas alterações no sistema nervoso central. O giro denteado (GD) do hipocampo sofre com modificações na expressão gênica dos canais iônicos das células granulares (CGs) e essas células sofrem alterações morfológicas. Essas alterações se manifestam com o brotamento de fibras musgosas, redução no número de espinhas dendríticas, encurtamento e estreitamento da arborização dendrítica. As modificações na expressão gênica dos canais iônicos afetam suas densidades máximas de condutância. Este estudo utilizou 40 modelos computacionais realistas para simular alterações nas condutâncias de canais iônicos e seus efeitos sobre dois grupos de CGs do GD. Os modelos foram construídos com base em reconstruções tridimensionais de 20 CGS com morfologia alterada após SE induzido por pilocarpina (CG-PILO) e 20 de morfologia normal (CG-controle). Foram dotados dos canais iônicos de sódio rápido (Na), canal de potássio de retificação tardia rápido (fKdr), canal de potássio de retificação tardia lento (fKdr), canal de potássio de tipo A (KA), canal de potássio dependente de cálcio e de voltagem de alta condutância (BK), canal de potássio dependente de cálcio de baixa condutância (SK) e canais de cálcio dos tipos T, N e L. As simulações foram realizadas no software Neuron. Foram realizados test t para detectar se ocorre diferenças significativas entre os grupos CG-controle e CG-PILO As alterações nas densidades máximas de condutância provocaram mudanças nos parâmetros de excitabilidade dos grupos CG-PILO e CG- controle, alterando valores de frequência de disparos, reobase e cronaxia. Os grupos apresentam respostas significativamente diferentes para as médias de reobase para a maioria dos valores de densidade máxima de condutância,, porém para cronaxia a maioria dos grupo não apresentou diferenças significativas. O grupo CG-controle apresentou médias maiores de frequência de disparos que o CG-PILO e o grupo CG-PILO apresentou valores de reobase maior para as alterações de densidade de condutância da maioria dos canais, sendo essas diferenças significativas. / The occurrence of status epilepticus (SE) triggers some changes in the central nervous system. The dentate gyrus (DG) of the hippocampus suffers from changes in gene expression of ion channels of granule cells (GCs) and these cells undergo morphological changes. These changes manifest themselves with mossy fiber sprouting, reduction in the number of dendritic spines, shortening and narrowing of dendritic branching. Changes in gene expression of ion channels affect their maximum densities of conductance. This study used 40 realistic computer models to simulate changes in conductance of ion channels and its effect on two groups of CGs of the GD. The models were built based on three-dimensional reconstructions of 20 CGS with morphology changed after pilocarpine-induced SE (CG-PILO) and 20 normal morphology (CG-control). The models were equipped with the ion channels of fast sodium (Na), fast delayed rectifying potassium channel (fKDR), slow delayed rectifying potassium channel (fKdr), potassium channel type A (KA), potassium channel dependent calcium and high voltage conductance (BK), potassium channel dependent calcium low conductance (SK) and the calcium channel types T, N and L. The simulations were performed at Neuron software.T tests were performed to p-values <0.05 for detecting significant differences between the GC-control group and GC-PILO. Changes in maximum densities conductance caused changes in excitability parameters CG-PILO and GC- control groups, by changing frequency values of spikes, rheobase and chronaxie. The groups have significantly different responses to the averages for the most rheobase maximum density values of conductance, but these differences were shortly found for chronaxie values. The CG-control group had higher average frequency of spikes than the CG-PILO group. The CG-PILO group had rheobase values higher for conductance density changes the most channels. These differences are significant.

Desde as descobertas das propriedades dos campos receptivos dos neurônios corticais no córtex visual primário, sua organização tem sido estudada com o auxílio de diversos métodos como eletrofisiologia, imageamento cortical e neurociência computacional. Poucos modelos mostram-se capazes de apresentar dominância ocular e seletividade à orientação simultaneamente. Um modelo em larga escala do sistema visual primário de mamíferos foi construído usando o GENESIS 2.2. O modelo contém aproximadamente 10.000 neurônios biologicamente plausíveis em oito matrizes representando setores das duas retinas, duas lâminas do núcleo geniculado lateral dorsal e duas lâminas representando o córtex visual (cada lâmina composta por uma matriz de células excitatórias e uma matriz de células inibitórias). As propriedades fisiológicas e estruturais do modelo foram determinadas com base em dados experimentais do sistema visual primário de mamíferos. Os neurônios apresentaram respostas binoculares e seletividade à orientação em boa concordância com os resultados experimentais. Apesar de neurônios corticais terem mostrado grande heterogeneidade em seus níveis de seletividade, a latência da resposta manteve-se constante e em boa concordância com resultados experimentais. / Since the discovery of the receptive field properties of cortical neurons in the primary visual cortex, their organization has been studied with many methods ranging from electrophysiology and optical imaging to computational neuroscience. Few models have been capable of showing ocularity and orientation selectivity simultaneously. A large-scale computational model of the mammalian primary visual pathway was constructed using GENESIS 2.2. The model consists of ~10,000 biologically plausible neurons organized in eight arrays to represent sectors of two retinas, two laminae of the dorsal lateral geniculate nucleus and two laminae of the visual cortex (each cortical lamina composed of a matrix of excitatory neurons and a matrix of inhibitory neurons). The physiological and architecture properties of the model were derived from experimental data for the mammalian primary visual pathway. Neurons have shown ocular and orientation selectivity dependent responses in good agreement with data. Though neurons in the cortex have shown markedly heterogeneity in the tuning responses, the latency of response was uniform and in good agreement with reported data.

Tinnitus is the perception of sound in the absence of an actual sound stimulus. Recent developments have shifted the focus to the central nervous system and the neural correlate of tinnitus. Broadly, tinnitus involves cortical map rearrangement, pathological neural synchrony, and increased spontaneous firing rates. Various cortical regions, such as Heschl’s gyrus in the auditory cortex, have been found to be associated with different aspects of tinnitus, such as perception and loudness. I propose a cortical stimulation mapping study of Heschl’s gyrus using a depth and subdural electrode montage to conduct electrocorticography. This study would provide high-resolution data on abnormal frequency band oscillations characteristic of tinnitus and pinpoint regions where they occur. The validity of the neural synchrony model would also be tested in this study.

Alzheimer’s Disease (AD) is the only condition in the top ten leading causes of death for which we do not have an effective treatment that prevents, slows, or stops its progression. Our ability to design useful interventions relies on (a) increasing our understanding of the pathological process of AD and (b) improving our ability for its early detection. These goals are impeded by our current reliance on the clinical symptoms of AD for its diagnosis. This characterizations of AD often falsely assumes a unified, underlying AD-specific pathology for similar presentations of dementia that leads to inconsistent diagnoses. It also hinges on postmortem verification, and so is not a helpful method for identifying patients and research subjects in the beginning phases of the pathophysiological process. Instead, a new biomarker-based approach provides a more biological understanding of the disease and can detect pathological changes up to 20 years before the clinical symptoms emerge. Subjects are assigned a profile according to their biomarker measures of amyloidosis (A), tauopathy (T) and neurodegeneration (N) that reflects their underlying pathology in vivo. AD is confirmed as the underlying pathology when subjects have abnormal values of both amyloid and tauopathy biomarkers, and so have a biomarker profile of A+T+(N)- or A+T+(N)+. This new biomarker based characterization of AD can be combined with machine learning techniques in multimodal classification studies to shed light on the elements of the AD pathological process and develop early detection paradigms. A guiding research framework is proposed for the development of reliable, biologically-valid and interpretable multimodal classification models.

We present a novel scheme ("Categorical Basis Functions", CBF) for object class representation in the brain and contrast it to the "Chorus of Prototypes" scheme recently proposed by Edelman. The power and flexibility of CBF is demonstrated in two examples. CBF is then applied to investigate the phenomenon of Categorical Perception, in particular the finding by Bulthoff et al. (1998) of categorization of faces by gender without corresponding Categorical Perception. Here, CBF makes predictions that can be tested in a psychophysical experiment. Finally, experiments are suggested to further test CBF.

The co-existence of bursting activity and silence is a common property of various neuronal models. We describe a novel mechanism explaining the co-existence of and the transition between these two regimes. It is based on the specific homoclinic and Andronov-Hopf bifurcations of the hyper- and depolarized steady states that determine the co-existence domain in the parameter space of the leech heart interneuron models: canonical and simplified. We found that a sub-critical Andronov-Hopf bifurcation of the hyperpolarized steady state gives rise to small amplitude sub-threshold oscillations terminating through the secondary homoclinic bifurcation. Near the corresponding boundary the system can exhibit long transition from bursting oscillations into silence, as well as the bi-stability where the observed regime is determined by the initial state of the neuron. The mechanism found is shown to be generic for the simplified 4D and the original 14D leech heart interneuron models.

The study of neurons is of fundamental importance in biology and medicine. Neurons are the most basic unit of information processing in the nervous system of humans and all other vertebrates and in complex invertebrates. In addition, networks of neurons (the human brain) are the most sophisticated computational devices known, and the study of neurons individually and working in concert is seen as a step toward understanding consciousness and cognition. In the 1950's Hodgkin and Huxley developed a system of nonlinear ordinary differential equations to describe the behavior of a neuron found in the squid. Equations of this form have since been used to model the behavior of a multitude of neurons across a broad spectrum of species. Hodgkin-Huxley type neuron models helped lay the foundation for computational neuroscience, and they remain widely used in the study of neuron behavior almost sixty years after their development. Hodgkin-Huxley type models accept a set of parameters as input and generate data describing the electrical activity of the neuron as a function of time. We develop inversion algorithms to predict a set of input parameter values from the voltage trace data generated by the model. We test our algorithm on data from the Hodgkin-Huxley equations, and we extend the algorithm to solve the inverse problem associated with a more complex Hodgkin-Huxley type model for a lobster stomatogastric neuron. We find strong empirical evidence that the algorithms produce parameter values that generate a good fit to the target voltage trace, and we prove that under certain conditions the inversion algorithm for the Hodgkin-Huxley equations converges to a perfect match. To our knowledge this is the first parameter optimization procedure for which convergence has been shown theoretically. Understanding the relationship between the parameters of a neuron model and its output has implications for designing effective neuron models and for explaining the mechanisms by which neurons regulate their behavior. Inversion algorithms for Hodgkin-Huxley type neuron models are an important theoretical and practical step toward understanding the relationship between model parameters and model behavior, and toward the larger problem of inferring neuronal parameters from behavior observed experimentally.

A multifunctional central pattern generator (CPG) can produce bursting polyrhythms that determine locomotive activity in an animal: for example, swimming and crawling in a leech. Each rhythm corresponds to a specific attractor of the CPG. We employ a Hodgkin-Huxley type model of a bursting leech heart interneuron, and connect three such neurons by fast inhibitory synapses to form a ring. This network motif exhibits multistable co-existing bursting rhythms. The problem of determining rhythmic outcomes is reduced to an analysis of fixed points of Poincare mappings and their attractor basins, in a phase plane defined by the interneurons' phase differences along bursting orbits. Using computer assisted analysis, we examine stability, bifurcations of attractors, and transformations of their basins in the phase plane. These structures determine the global bursting rhythms emitted by the CPG. By varying the coupling synaptic strength, we examine the dynamics and patterns produced by inhibitory networks.