s. Text / s. Text

Prosopagnosia is defined as a profound deficit in facial identification which can be either acquired due to brain damage or is present from birth, i.e. congenital. Normally, faces and objects are processed in different parts of the inferotemporal cortex by distinct cortical systems for face vs. object recognition, an association of function and location. Accordingly, in acquired prosopagnosia locally restricted damage can lead to specific deficits in face recognition. However, in congenital prosopagnosia faces and objects are also processed in spatially separated areas. Accordingly, the face recognition deficit in congenital prosopagnosia can not be solely explained by the association of function and location. Rather, this observation raises the question why and how such an association evolves at all. So far, no quantitative or computational model of congenital prosopagnosia has been proposed and models of acquired prosopagnosia have focused on changes in the information processing taking place after in icting some kind of \damage" to the system. To model congenital prosopagnosia, it is thus necessary to understand how face processing in congenital prosopagnosia differs from normal face processing, how differences in neuroanatomical development can give rise to differences in processing and last but not least why facial identification requires a specialized cortical processing system in the first place. In this work, a computational model of congenital prosopagnosia is derived from formal considerations, implemented in artificial neural network models of facial information encoding, and tested in experiments with prosopagnosic subjects. The main hypothesis is that the deficit in congenital prosopagnosia is caused by a failure to obtain adequate descriptions of individual faces: A predisposition towards a reduced structural connectivity in visual cortical areas enforces descriptions of visual stimuli that lack the amount of detail necessary to distinguish a specific exemplar from its population, i.e. achieve a successful identification. Formally recognition tasks can be divided into identification tasks (separating a single individual from its sampling population) and classification tasks (partitioning the full object space into distinct classes). It is shown that a high-dimensionality in the sensory representation facilitates individuation (\blessing of dimensionality"), but complicates estimation of object class representations (\curse of dimensionality"). The dimensionality of representations is then studied explicitly in a neural network model of facial encoding. Whereas optimal encoding entails a \holistic" (high-dimensional) representation, a constraint on the network connectivity induces a decomposition of faces into localized, \featural" (low-dimensional) parts. In an experimental validation, the perceptual deficit in congenital prosopagnosia was limited to holistic face manipulations and didn't extend to featural manipulations. Finally, an extensive and detailed investigation of face and object recognition in congenital prosopagnosia enabled a better behavioral characterization and the identification of subtypes of the deficit. In contrast to previous models of prosopagnosia, here the developmental aspect of congenital prosopagnosia is incorporated explicitly into the model, quantitative arguments for a deficit that is task specific (identification) - and not necessarily domain specific (faces) - are provided for synthetic as well as real data (face images), and the model is validated empirically in experiments with prosopagnosic subjects.

Sensorimotor processing is a critical function of the human brain with multiple cortical areas specialized for sensory recognition or motor execution. Although there has been considerable research into sensorimotor control in humans, the steps between sensory recognition and motor execution are not fully understood. This thesis investigates different aspects of sensorimotor processing in the human brain and proposes approaches to cognitive architectures. Here, I describe a series of six studies: an examination of sensorimotor processing in the human brain, evaluation of new mobile EEG systems for modern sensorimotor paradigms, investigation of a balance between memorization and active sampling of visual information in a sensorimotor task, and three studies on cognitive architectures for spatial reasoning and navigation in 2D environments.

Ongoing research to fight cancer, one of the dominant diseases of the 21st century has led to big progress especially when it comes to understanding the tumor growth and metastasis. This includes the discovery of the molecular mechanisms of tumor vascularization, which is critically required for establishment of tumor metastasis. Formation of new blood vessels is the first step in tumor vascularization. Therefore, understanding the molecular and cellular basis of tumor vascularization attracted a significant effort studying in biomedical research. The blood vessels for supplying tumor can be formed by sprouting from pre-existing vessels, a process called angiogenesis, or by vasculogenesis, that is de novo formation of blood vessels from not fully differentiated progenitor cell populations. Vasculogenic endothelial progenitor cells (EPCs) can either be activated from populations in the bone marrow reaching the pathological region via the circulation or they can be recruited from local reservoirs. Neovessel formation influences tumor progression, hence therapeutic response model systems of angiogenesis/vasculogenesis are necessary to study the underlying mechanisms. Although, initially the research in this area focused more on angiogenesis, it is now well understood that both angiogenesis and postnatal vasculogenesis contribute to neovessel formation in adult under both most pathological as well as physiological conditions. Studies in the last two decades demonstrate that in addition to the intimal layer of fully differentiated mature endothelial cells (ECs) and various smaller supplying vessels (vasa vasorum) that can serve as a source for new vessels by angiogenesis, especially the adventitia of large and medium size blood vessels harbors various vascular wall-resident stem and progenitor cells (VW-SPCs) populations that serve as a source for new vessels by postnatal vasculogenesis. However, little is known about the potential role of VW-SPCs in tumor vascularization. To this end, the present work started first to establish a modified aortic ring assay (ARA) using mouse aorta in order to study the contribution of vascular adventitia-resident VW-SPCs to neovascularization in general and in presence of tumor cells. ARA is already established an ex vivo model for neovascularization allows to study the morphogenetic events of complex new vessel formation that includes all layers of mature blood vessels, a significant advantage over the assays that employ monolayer endothelial cell cultures. Moreover, in contrast to assays employing endothelial cells monocultures, both angiogenic and vasculogenic events take place during new vessel formation in ARA although the exact contribution of these two processes to new vessel formation cannot be easily distinguished in conventional ARA. Thus, in this study, a modified protocol for the ARA (mdARA) was established by either removing or keeping the aortic adventitia in place. The mdARA allows to distinguish the role of VW-SPCs from those of other aortic layers. The present data show that angiogenic sprouting from mature aortic endothelium was markedly delayed when the adventitial layer was removed. Furthermore, the network between the capillary-like sprouts was significantly reduced in absence of aortic adventitia. Moreover, the stabilization of new sprouts by assembling the NG2+ pericyte-like cells that enwrapped the endothelial sprouts from the outside was improved when the adventitial layer remained in place. Next, mimicking the tumor-vessel adventitia-interaction, multicellular tumor spheroids (MCTS) and aortic rings (ARs) with or without adventitia of C57BL/6-Tg (UBC-GFP) mice were confronted within the collagen gel and cultured ex vivo. This 3D model enabled analysis of the mobilization, migration and capillary-like sprouts formation by VW-SPCs within tumor-vessel wall-interface in comparison to tumor-free side of the ARs. Interestingly, while MCTS preferred the uptake of single vascular adventitia-derived cells, neural spheroids were directly penetrated by capillary-like structures that were sprouted from the aortic adventitia. In summary, the model established in this work allows to study new vessel formation by both postnatal vasculogenesis and angiogenesis under same conditions. It can be applied in various mouse models including reporter mouse models, e.g. Cxcr1 CreER+/mTmG+/- mice, in which GFP-marked macrophages of the vessel wall were directly observed as they mobilized from their niche and migrated into collagen gel. Another benefit of the model is that it can be used for testing different factors such as small molecules, growth factors, cytokines, and drugs with both pro- and anti-angiogenic/vasculogenic effects. / Die Forschungsarbeiten der letzten Jahrzehnte zur Bekämpfung der Krebserkrankung, einer der dominierenden Krankheiten des 21. Jahrhunderts, haben zu großen Fortschritten, insbesondere im Verständnis bezüglich der Tumormetastasierung geführt. Dies schließt die Prozesse der Tumorvaskularisierung als einen der initialen Schritte der Metastasierung mit ein, die nach wie vor nicht ausreichend geklärt sind. Die Blutgefäßbildung zur Versorgung des Tumorgewebes kann durch die Anagiogenese, die als Einsprießen neuer Gefäße aus den bereits vorhandenen Blutgefäßen definiert wird, oder durch die Vaskulogenese, die als Gefäßneubildung aus Stamm- und Vorläuferzellen beschrieben wird, sichergestellt werden. Noch nicht vollständig ausdifferenzierte endotheliale Vorläuferzellen (EPCs) werden dabei nach dem bisherigen Kenntnistand aus dem Knochenmark rekrutiert und erreichen die Regionen der Gefäßneubildung über die Blutzirkulation und können dort zur de novo Formierung neuer Blutgefäße auch beim Erwachsenen beitragen. Untersuchungen der letzten zwei Dekaden haben gezeigt, dass solche Vorläuferzellen auch aus lokalen Reservoiren, wie z.B. aus der Gefäßwandadventitia der bereits existierenden Blutgefäße mobilisiert werden. Da die Bildung neuer Gefäße einen direkten Einfluss auf die Tumorprogression hat, sind entsprechende Modellsysteme notwendig, um die zugrundeliegenden Mechanismen möglichst präzise zu untersuchen. Obwohl sich die Forschung der Tumorvaskularisierung zunächst auf Prozesse der Angiogenese konzentrierte, ist es mittlerweile ausreichend belegt, dass auch die Vaskulogenese zur Tumorvaskularisierung beiträgt und somit die Tumorprogression beeinflusst. Anhand einer Vielzahl an Studien der letzten zwei Jahrzehnte konnte demonstriert werden, dass sich, neben der Intimaschicht, die vollständig differenzierte Endothelzellen (ECs) enthält und kleineren Gefäßwand-versorgenden Blutgefäßen, der sogenannten Vasa vasorum in der Adventitia der großen Gefäße, auch Populationen gefäßwandresidenter Stamm- und Vorläuferzellen (VW-SPCs) in der äußeren Gefäßwandschicht, nämlich der Adventitia fast aller Gefäßabschnitte nachweisen lassen. Obwohl diese als Quelle neuer Gefäße in der postnatalen Vaskulogenese beschrieben sind, ist nach wie vor wenig über die potenzielle Rolle von VW-SPCs in der Tumorvaskularisation bekannt. Daher wurde im Rahmen dieser Arbeit zuerst ein modifiziertes Aortic Ringassay (ARA) unter Verwendung der Mausaorta etabliert, um den Beitrag der VW-SPCs zur Neovaskularisierung im Allgemeinen und zur Tumorvaskularisierung im Speziellen ex vivo untersuchen zu können. ARA ist ein bereits seit einigen Jahrzehnten etabliertes ex vivo Modell zur Untersuchung der Gefäßneubildung durch Angiogenese. Mittels ARA kann die Bildung komplexer vaskulärer Strukturen in Präsenz aller Wandschichten reifer Blutgefäße untersucht werden, was einen wesentlichen Vorteil gegenüber der Verwendung von Endothelzellkulturen in Monolayer bedeutet. Hierbei ist jedoch anzumerken, dass in ARA sowohl angiogene als auch vaskulogene Prozesse zur Gefäßbildung beitragen, aber der genaue Beitrag beider Prozesse schwer oder kaum voneinander unterscheidbar ist. Daher wurde im Rahmen dieser Arbeit ein modifiziertes ARA-Protokoll etabliert (mdARA), in welchem die aortale Adventitia vor dem Beginn des ARA entweder entfernt oder belassen wurde und somit die Rolle der VW-SPCs bei der Gefäßsprossung von den Zellen anderer Aortenwandschichten differenziert studiert werden konnte. Die dabei generierten Daten zeigen, dass sich die angiogene Aussprossung aus dem reifen Aortenendothel nach Entfernung der adventitialen Schicht deutlich verzögerte. Darüber hinaus war das Netzwerk zwischen den kapillarartigen Sprossen in Abwesenheit der Aortenadventitia signifikant reduziert. Mehr noch, das Belassen der adventitialen Wandstruktur führte zu einer verbesserten Stabilisierung neuer Gefäßsprossen. Als sichtbares Korrelat hierfür zeigte sich eine stärkere und bessere Anlagerung der NG2+ Perizyten-ähnlichen Zellen zu den endothelialen Kapillar-ähnlichen Aussprossungen von außen, wie Perizyten an der Kapillarwand in situ. Als nächstes wurden die Aortenringe (ARs) von C57BL/6-Tg (UBC-GFP)-Mäusen mit multizellulären Tumor-Sphäroiden (MCTS) in Kollagengel ko-kultiviert, um die Interaktion zwischen Tumor und Gefäßwand-Adventitia ex vivo nachzuahmen. Dieses 3D Modell ermöglichte die Analyse der Mobilisierung und Migration der VW-SPCs von der aortalen Adventitia sowohl zu der Tumorseite in den Tumor-Gefäßwand-Interfaces als auch zu der tumorfreien Seite der Aortenringe. Interessanterweise wurde die Kapillarsprossung im Tumor-Gefäßwand-Interface an der Grenze zum MCTS gestoppt und die VW-SPCs als Einzelzellen in die MCTS aufgenommen. Demgegenüber wurde auf der tumor-freien Seite der Aortenringe eine deutlich längere Kapillaraussprossung beobachtet. Im Gegensatz zu MCTS resultierte die Ko-Kultivierung der ARs mit neuronalen Spheroiden darin, dass die aus der aortalen Adventitia aussprossenden Kapillar-ähnlichen Strukturen direkt in die neuronalen Spheroide penetrierten. Zusammenfassend berücksichtigt dieses neuartige in vitro 3D-Modell sowohl Angiogenese als auch Vaskulogenese und bietet vielfältige Vorteile, wie zum Beispiel die Kompatibilität zu verschiedenen Mausmodellen einschließlich der Reporter-Mausmodelle, wie z.B. die in dieser Arbeit gezeigte Verwendung der Aorta von Cxcr1 CreER+/mTmG+/- um die GFP-markierten Makrophagen aus der Gefäßwand bei der Gefäßaussprossung studieren zu können. Des Weiteren ist dieses Model auch für Testung unterschiedlicher Faktoren und Therapeutika einschließlich der anti-angiogenen und -vasculogenen Substanzen unter ex vivo Bedingungen geeignet.

Leer

When DNA is sequenced, nucleotide calls are produced along with their individual error probabilities, which are usually reported in the form of a per-base quality score. However, these quality scores have not generally been incorporated into probabilistic models as there is typically a poor correlation between the predicted and observed error rates. Computational tools aimed at sequence analysis have therefore used arbitrary cutoffs on quality scores which often unnecessarily reduce the amount of data that can be analyzed. A different approach involves recalibration of those quality scores using known genomic variants to measure empirical error rates. However, for this heuristic to work, an adequate characterization of the variants present in a population must be available -which means that this approach is not possible for a wide range of species. This thesis develops methods to directly produce error probabilities that are representative of their empirical error rates for raw sequencing data. These can then be incorporated into Bayesian maximum a posteriori algorithms to make highly accurate inferences about the likelihood of the model that gave rise to this observed data. First, an algorithm to produce highly accurate nucleotide basecalls along with calibrated error probabilities is presented. Using the resulting data, individual reads can be robustly as- signed to their samples of origin and ancient DNA fragments can be inferred even at high error rates. For archaic hominin samples, the number of DNA fragments from present-day human contamination can also be accurately quantified. The novel algorithms developed during the course of this thesis provide an alternative approach to working with Illumina sequence data. They also provide a demonstrable improvement over existing computational methods for basecalling, inferring ancient DNA fragments, demultiplexing, and estimating present-day human contamination along with reconstruction of mitochondrial genomes in ancient hominins.

We study an energy functional able to describe low energy configurations of a two dimensional lattice with dislocations in a nonlinear elasticity regime. The main result can be described as follows: configurations of energy comparable to the lattice spacing consist of piecewise constant microrotations with small angle grain boundaries between them. Moreover, we also give bounds to the energy of particular configurations describing a small angle symmetric tilt grain boundary.

No description available.

This thesis studies the visualization of set collections either via or defines as the relations among contours. In the first part, dynamic Euler diagrams are used to communicate and improve semimanually the result of clustering methods which allow clusters to overlap arbitrarily. The contours of the Euler diagram are rendered as implicit surfaces called blobs in computer graphics. The interaction metaphor is the moving of items into or out of these blobs. The utility of the method is demonstrated on data arising from the analysis of gene expressions. The method works well for small datasets of up to one hundred items and few clusters. In the second part, these limitations are mitigated employing a GPU-based rendering of Euler diagrams and mixing textures and colors to resolve overlapping regions better. The GPU-based approach subdivides the screen into triangles on which it performs a contour interpolation, i.e. a fragment shader determines for each pixel which zones of an Euler diagram it belongs to. The rendering speed is thus increased to allow multiple hundred items. The method is applied to an example comparing different document clustering results. The contour tree compactly describes scalar field topology. From the viewpoint of graph drawing, it is a tree with attributes at vertices and optionally on edges. Standard tree drawing algorithms emphasize structural properties of the tree and neglect the attributes. Adapting popular graph drawing approaches to the problem of contour tree drawing it is found that they are unable to convey this information. Five aesthetic criteria for drawing contour trees are proposed and a novel algorithm for drawing contour trees in the plane that satisfies four of these criteria is presented. The implementation is fast and effective for contour tree sizes usually used in interactive systems and also produces readable pictures for larger trees. Dynamical models that explain the formation of spatial structures of RNA molecules have reached a complexity that requires novel visualization methods to analyze these model\''s validity. The fourth part of the thesis focuses on the visualization of so-called folding landscapes of a growing RNA molecule. Folding landscapes describe the energy of a molecule as a function of its spatial configuration; they are huge and high dimensional. Their most salient features are described by their so-called barrier tree -- a contour tree for discrete observation spaces. The changing folding landscapes of a growing RNA chain are visualized as an animation of the corresponding barrier tree sequence. The animation is created as an adaption of the foresight layout with tolerance algorithm for dynamic graph layout. The adaptation requires changes to the concept of supergraph and it layout. The thesis finishes with some thoughts on how these approaches can be combined and how the task the application should support can help inform the choice of visualization modality.

We investigate the truncated moment problem, especially the Carathéodory number, the set of atoms and determinacy.