Assessment of Severe Apnoea through Voice Analysis, Automatic Speech, and Speaker Recognition Techniques

This study is part of an ongoing collaborative effort between the medical and the signal processing communities to promote research on applying standard Automatic Speech Recognition (ASR) techniques for the automatic diagnosis of patients with severe obstructive sleep apnoea (OSA). Early detection of severe apnoea cases is important so that patients can receive early treatment. Effective ASR-based detection could dramatically cut medical testing time. Working with a carefully designed speech database of healthy and apnoea subjects, we describe an acoustic search for distinctive apnoea voice characteristics. We also study abnormal nasalization in OSA patients by modelling vowels in nasal and nonnasal phonetic contexts using Gaussian Mixture Model (GMM) pattern recognition on speech spectra. Finally, we present experimental findings regarding the discriminative power of GMMs applied to severe apnoea detection. We have achieved an 81% correct classification rate, which is very promising and underpins the interest in this line of inquiry.

Alterations of specific biomarkers of metabolic pathways in vascular tree from patients with Type 2 diabetes.

The aims of this study were to check whether different biomarkers of inflammatory, apoptotic, immunological or lipid pathways had altered their expression in the occluded popliteal artery (OPA) compared with the internal mammary artery (IMA) and femoral vein (FV) and to examine whether glycemic control influenced the expression of these genes. The study included 20 patients with advanced atherosclerosis and type 2 diabetes mellitus, 15 of whom had peripheral arterial occlusive disease (PAOD), from whom samples of OPA and FV were collected. PAOD patients were classified based on their HbA1c as well (HbA1c ≤ 6.5) or poorly (HbA1c > 6.5) controlled patients. Controls for arteries without atherosclerosis comprised 5 IMA from patients with ischemic cardiomyopathy (ICM). mRNA, protein expression and histological studies were analyzed in IMA, OPA and FV. After analyzing 46 genes, OPA showed higher expression levels than IMA or FV for genes involved in thrombosis (F3), apoptosis (MMP2, MMP9, TIMP1 and TIM3), lipid metabolism (LRP1 and NDUFA), immune response (TLR2) and monocytes adhesion (CD83). Remarkably, MMP-9 expression was lower in OPA from well-controlled patients. In FV from diabetic patients with HbA1c ≤6.5, gene expression levels of BCL2, CDKN1A, COX2, NDUFA and SREBP2 were higher than in FV from those with HbA1c >6.5. The atherosclerotic process in OPA from diabetic patients was associated with high expression levels of inflammatory, lipid metabolism and apoptotic biomarkers. The degree of glycemic control was associated with gene expression markers of apoptosis, lipid metabolism and antioxidants in FV. However, the effect of glycemic control on pro-atherosclerotic gene expression was very low in arteries with established atherosclerosis.

Microtubule-associated protein-2 immunoreactivity: a useful tool in the differential diagnosis of low-grade neuroepithelial tumors.

Complex and variable morphological phenotypes pose a major challenge to the histopathological classification of neuroepithelial tumors. This applies in particular for low-grade gliomas and glio-neuronal tumors. Recently, we and others have identified microtubule-associated protein-2 (MAP2) as an immunohistochemical marker expressed in the majority of glial tumors. Characteristic cell morphologies can be recognized by MAP2 immunoreactivity in different glioma entities, i.e., process sparse oligodendroglial versus densely ramified astrocytic elements. Here, we describe MAP2-immunoreactivity patterns in a large series of various neuroepithelial tumors and related neoplasms (n = 960). Immunohistochemical analysis led to the following conclusions: (1) specific pattern of MAP2-positive tumor cells can be identified in 95% of glial neoplasms; (2) ependymal tumors do not express MAP2 in their rosette-forming cell component; (3) tumors of the pineal gland as well as malignant embryonic tumors are also characterized by abundant MAP2 immunoreactivity; (4) virtually no MAP2 expression can be observed in the neoplastic glial component of glio-neuronal tumors, i.e. gangliogliomas; (5) malignant glial tumor variants (WHO grade III or IV) exhibit different and less specific MAP2 staining patterns compared to their benign counterparts (WHO grade I or II); (6) with the exception of melanomas and small cell lung cancers, MAP2 expression is very rare in metastatic and non-neuroepithelial tumors; (7) glial MAP2 expression was not detected in 56 non-neoplastic lesions. These data point towards MAP2 as valuable diagnostic tool for pattern recognition and differential diagnosis of low-grade neuroepithelial tumors.

Quantitative proteomics reveals subset-specific viral recognition in dendritic cells.

Dendritic cell (DC) populations consist of multiple subsets that are essential orchestrators of the immune system. Technological limitations have so far prevented systems-wide accurate proteome comparison of rare cell populations in vivo. Here, we used high-resolution mass spectrometry-based proteomics, combined with label-free quantitation algorithms, to determine the proteome of mouse splenic conventional and plasmacytoid DC subsets to a depth of 5,780 and 6,664 proteins, respectively. We found mutually exclusive expression of pattern recognition pathways not previously known to be different among conventional DC subsets. Our experiments assigned key viral recognition functions to be exclusively expressed in CD4(+) and double-negative DCs. The CD8alpha(+) DCs largely lack the receptors required to sense certain viruses in the cytoplasm. By avoiding activation via cytoplasmic receptors, including retinoic acid-inducible gene I, CD8alpha(+) DCs likely gain a window of opportunity to process and present viral antigens before activation-induced shutdown of antigen presentation pathways occurs.

Adaptive model selection using empirical complexities

Given $n$ independent replicates of a jointly distributed pair $(X,Y)\in {\cal R}^d \times {\cal R}$, we wish to select from a fixed sequence of model classes ${\cal F}_1, {\cal F}_2, \ldots$ a deterministic prediction rule $f: {\cal R}^d \to {\cal R}$ whose risk is small. We investigate the possibility of empirically assessingthe {\em complexity} of each model class, that is, the actual difficulty of the estimation problem within each class. The estimated complexities are in turn used to define an adaptive model selection procedure, which is based on complexity penalized empirical risk.The available data are divided into two parts. The first is used to form an empirical cover of each model class, and the second is used to select a candidate rule from each cover based on empirical risk. The covering radii are determined empirically to optimize a tight upper bound on the estimation error. An estimate is chosen from the list of candidates in order to minimize the sum of class complexity and empirical risk. A distinguishing feature of the approach is that the complexity of each model class is assessed empirically, based on the size of its empirical cover.Finite sample performance bounds are established for the estimates, and these bounds are applied to several non-parametric estimation problems. The estimates are shown to achieve a favorable tradeoff between approximation and estimation error, and to perform as well as if the distribution-dependent complexities of the model classes were known beforehand. In addition, it is shown that the estimate can be consistent,and even possess near optimal rates of convergence, when each model class has an infinite VC or pseudo dimension.For regression estimation with squared loss we modify our estimate to achieve a faster rate of convergence.

A Small World Perspective on Urban Systems

The theory of small-world networks as initiated by Watts and Strogatz (1998) has drawn new insights in spatial analysis as well as systems theory. The theoryâeuro?s concepts and methods are particularly relevant to geography, where spatial interaction is mainstream and where interactions can be described and studied using large numbers of exchanges or similarity matrices. Networks are organized through direct links or by indirect paths, inducing topological proximities that simultaneously involve spatial, social, cultural or organizational dimensions. Network synergies build over similarities and are fed by complementarities between or inside cities, with the two effects potentially amplifying each other according to the âeurooepreferential attachmentâeuro hypothesis that has been explored in a number of different scientific fields (Barabási, Albert 1999; Barabási A-L 2002; Newman M, Watts D, Barabàsi A-L). In fact, according to Barabási and Albert (1999), the high level of hierarchy observed in âeurooescale-free networksâeuro results from âeurooepreferential attachmentâeuro, which characterizes the development of networks: new connections appear preferentially close to nodes that already have the largest number of connections because in this way, the improvement in the network accessibility of the new connection will likely be greater. However, at the same time, network regions gathering dense and numerous weak links (Granovetter, 1985) or network entities acting as bridges between several components (Burt 2005) offer a higher capacity for urban communities to benefit from opportunities and create future synergies. Several methodologies have been suggested to identify such denser and more coherent regions (also called communities or clusters) in terms of links (Watts, Strogatz 1998; Watts 1999; Barabási, Albert 1999; Barabási 2002; Auber 2003; Newman 2006). These communities not only possess a high level of dependency among their member entities but also show a low level of âeurooevulnerabilityâeuro, allowing for numerous redundancies (Burt 2000; Burt 2005). The SPANGEO project 2005âeuro"2008 (SPAtial Networks in GEOgraphy), gathering a team of geographers and computer scientists, has included empirical studies to survey concepts and measures developed in other related fields, such as physics, sociology and communication science. The relevancy and potential interpretation of weighted or non-weighted measures on edges and nodes were examined and analyzed at different scales (intra-urban, inter-urban or both). New classification and clustering schemes based on the relative local density of subgraphs were developed. The present article describes how these notions and methods contribute on a conceptual level, in terms of measures, delineations, explanatory analyses and visualization of geographical phenomena.

An active contour-based atlas registration model applied to automatic subthalamic nucleus targeting on MRI: method and validation.

This paper presents a new non parametric atlas registration framework, derived from the optical flow model and the active contour theory, applied to automatic subthalamic nucleus (STN) targeting in deep brain stimulation (DBS) surgery. In a previous work, we demonstrated that the STN position can be predicted based on the position of surrounding visible structures, namely the lateral and third ventricles. A STN targeting process can thus be obtained by registering these structures of interest between a brain atlas and the patient image. Here we aim to improve the results of the state of the art targeting methods and at the same time to reduce the computational time. Our simultaneous segmentation and registration model shows mean STN localization errors statistically similar to the most performing registration algorithms tested so far and to the targeting expert's variability. Moreover, the computational time of our registration method is much lower, which is a worthwhile improvement from a clinical point of view.

Efficient total variation algorithm for fetal brain MRI reconstruction.

Fetal MRI reconstruction aims at finding a high-resolution image given a small set of low-resolution images. It is usually modeled as an inverse problem where the regularization term plays a central role in the reconstruction quality. Literature has considered several regularization terms s.a. Dirichlet/Laplacian energy, Total Variation (TV)- based energies and more recently non-local means. Although TV energies are quite attractive because of their ability in edge preservation, standard explicit steepest gradient techniques have been applied to optimize fetal-based TV energies. The main contribution of this work lies in the introduction of a well-posed TV algorithm from the point of view of convex optimization. Specifically, our proposed TV optimization algorithm or fetal reconstruction is optimal w.r.t. the asymptotic and iterative convergence speeds O(1/n2) and O(1/√ε), while existing techniques are in O(1/n2) and O(1/√ε). We apply our algorithm to (1) clinical newborn data, considered as ground truth, and (2) clinical fetal acquisitions. Our algorithm compares favorably with the literature in terms of speed and accuracy.

Surface assembly of poly(I:C) on PEGylated microspheres to shield from adverse interactions with fibroblasts.

By expressing an array of pattern recognition receptors (PRRs), fibroblasts play an important role in stimulating and modulating the response of the innate immune system. The TLR3 ligand polyriboinosinic acid-polyribocytidylic acid, poly(I:C), a mimic of viral dsRNA, is a vaccine adjuvant candidate to activate professional antigen presenting cells (APCs). However, owing to its ligation with extracellular TLR3 on fibroblasts, subcutaneously administered poly(I:C) bears danger towards autoimmunity. It is thus in the interest of its clinical safety to deliver poly(I:C) in such a way that its activation of professional APCs is as efficacious as possible, whereas its interference with non-immune cells such as fibroblasts is controlled or even avoided. Complementary to our previous work with monocyte-derived dendritic cells (MoDCs), here we sought to control the delivery of poly(I:C) surface-assembled on microspheres to human foreskin fibroblasts (HFFs). Negatively charged polystyrene (PS) microspheres were equipped with a poly(ethylene glycol) (PEG) corona through electrostatically driven coatings with a series of polycationic poly(L-lysine)-graft-poly(ethylene glycol) copolymers, PLL-g-PEG, of varying grafting ratios g from 2.2 up to 22.7. Stable surface assembly of poly(I:C) was achieved by incubation of polymer-coated microspheres with aqueous poly(I:C) solutions. Notably, recognition of both surface-assembled and free poly(I:C) by extracellular TLR3 on HFFs halted their phagocytic activity. Ligation of surface-assembled poly(I:C) with extracellular TLR3 on HFFs could be controlled by tuning the grafting ratio g and thus the chain density of the PEG corona. When assembled on PLL-5.7-PEG-coated microspheres, poly(I:C) was blocked from triggering class I MHC molecule expression on HFFs. Secretion of interleukin (IL)-6 by HFFs after exposure to surface-assembled poly(I:C) was distinctly lower as compared to free poly(I:C). Overall, surface assembly of poly(I:C) may have potential to contribute to the clinical safety of this vaccine adjuvant candidate.

Metastatic leishmaniasis. A chronic inflammatory response to Leishmania's viral endosymbiont.

Leishmania parasites have been plaguing humankind for centuries as a range of skin diseases named the cutaneous leishmaniases (CL). Carried in a hematophagous sand fly, Leishmania usually infests the skin surrounding the bite site, causing a destructive immune response that may persist for months or even years. The various symptomatic outcomes of CL range from a benevolent self- healing reddened bump to extensive open ulcerations, resistant to treatment and resulting in life- changing disfiguration. Many of these more aggressive outcomes are geographically isolated within the habitats of certain Neotropical Leishmania species; where about 15% of cases experience metastatic complications. However, despite this correlation, genetic analysis has revealed no major differences between species causing the various disease forms. We have recently identified a cytoplasmic dsRNA virus within metastatic L. guyanensis parasites that acts as a potent innate immunogen capable of worsening lesionai inflammation and prolonging parasite survival. The dsRNA genome of Leishmania RNA virus (LRV) binds and stimulates Toll-Like-Receptor-3 (TLR3), inducing this destructive inflammation, which we speculate as a factor contributing to the development of metastatic disease. This thesis establishes the first experimental model of LRV-mediated leishmanial metastasis and investigates the role of non-TLR3 viral recognition pathways in LRV-mediated pathology. Viral dsRNA can be detected by various non-TLR3 pattern recognition receptors (PRR); two such PRR groups are the RLRs (Retinoic acid-inducible gene 1 like receptors) and the NLRs (nucleotide- binding domain, leucine-rich repeat containing receptors). The RLRs are designed to detect viral dsRNA in the cytoplasm, while the NLRs react to molecular "danger" signals of cell damage, often oligomerizing into molecular scaffolds called "inflammasomes" that activate a potent inflammatory cascade. Interestingly, we found that neither RLR signalling nor the inflammasome pathway had an effect on LRV-mediated pathology. In contrast, we found a dramatic inflammasome independent effect for the NLR family member, NLRP10, where a knockout mouse model showed little evidence of disease. This phenotype was mimicked in an NLR knockout with which NLRP10 is known to interact: NLRC2. As this pathway induces the chronic inflammatory cell lineage TH17, we investigated the role of its key chronic inflammatory cytokine, IL-17A, in human patients infected by L. guyanensis. Indeed, patients infected with LRV+ parasites had a significantly increased level of IL-17A in lesionai biopsies. Interestingly, LRV presence was also associated with a significant decrease in the correlate of protection, IFN-y. This association was repeated in our murine model, where after we were able to establish the first experimental model of LRV-dependent leishmanial metastasis, which was mediated by IL-17A in the absence of IFN-y. Finally, we tested a new inhibitor of IL-17A secretion, SR1001, and reveal its potential as a Prophylactic immunomodulator and potent parasitotoxic drug. Taken together, these findings provide a basis for anti-IL-17A as a feasible therapeutic intervention to prevent and treat the metastatic complications of cutaneous leishmaniasis. -- Les parasites Leishmania infectent l'homme depuis des siècles causant des affections cutanées, appelées leishmanioses cutanées (LC). Le parasite est transmis par la mouche des sables et réside dans le derme à l'endroit de la piqûre. Au niveau de la peau, le parasite provoque une réponse immunitaire destructrice qui peut persister pendant des mois voire des années. Les symptômes de LC vont d'une simple enflure qui guérit spontanément jusqu' à de vastes ulcérations ouvertes, résistantes aux traitements. Des manifestations plus agressives sont déterminées par les habitats géographiques de certaines espèces de Leishmania. Dans ces cas, environ 15% des patients développent des lésions métastatiques. Aucun «facteur métastatique» n'a encore été trouvé à ce jour dans ces espèces. Récemment, nous avons pu identifier un virus résidant dans certains parasites métastatiques présents en Guyane française (appelé Leishmania-virus, ou LV) et qui confère un avantage de survie à son hôte parasitaire. Ce virus active fortement la réponse inflammatoire, aggravant l'inflammation et prolongeant l'infection parasitaire. Afin de diagnostiquer, prévenir et traiter ces lésions, nous nous sommes intéressés à identifier les composants de la voie de signalisation anti-virale, responsables de la persistance de cette inflammation. Cette étude décrit le premier modèle expérimental de métastases de la leishmaniose induites par LV, et identifie plusieurs composants de la voie inflammatoire anti-virale qui facilite la pathologie métastatique. Contrairement à l'homme, les souris de laboratoire infectées par des Leishmania métastatiques (contenant LV, LV+) ne développent pas de lésions métastatiques et guérissent après quelques semaines d'infection. Après avoir analysé un groupe de patients atteints de leishmaniose en Guyane française, nous avons constaté que les personnes infectées avec les parasites métastatiques LV+ avaient des niveaux significativement plus faibles d'un composant immunitaire protecteur important, appelé l'interféron (IFN)-y. En utilisant des souris génétiquement modifiées, incapables de produire de l'IFN-y, nous avons observé de telles métastases. Après inoculation dans le coussinet plantaire de souris IFN-y7" avec des parasites LV+ ou LV-, nous avons démontré que seules les souris infectées avec des leishmanies ayant LV développent de multiples lésions secondaires sur la queue. Comme nous l'avons observé chez l'homme, ces souris sécrètent une quantité significativement élevée d'un composant inflammatoire destructeur, l'interleukine (IL)-17. IL-17 a été incriminée pour son rôle dans de nombreuses maladies inflammatoires chroniques. On a ainsi trouvé un rôle destructif similaire pour l'IL-17 dans la leishmaniose métastatique. Nous avons confirmé ce rôle en abrogeant IL-17 dans des souris IFN-y7- ce qui ralentit l'apparition des métastases. Nous pouvons donc conclure que les métastases de la leishmaniose sont induites par l'IL-17 en absence d'IFN-v. En analysant plus en détails les voies de signalisation anti-virale induites par LV, nous avons pu exclure d'autres voies d'activation de la réponse inflammatoire. Nous avons ainsi démontré que la signalisation par LV est indépendante de la signalisation inflammatoire de type « inflammasome ». En revanche, nous avons pu y lier plusieurs autres molécules, telles que NLRP10 et NLRC2, connues pour leur synergie avec les réponses inflammatoires. Cette nouvelle voie pourrait être la cible pour des médicaments inhibant l'inflammation. En effet, un nouveau médicament qui bloque la production d'IL-17 chez la souris s'est montré prometteur dans notre modèle : il a réduit le gonflement des lésions ainsi que la charge parasitaire, indiquant que la voie anti-virale /inflammatoire est une approche thérapeutique possible pour prévenir et traiter cette infection négligée.

Machine learning for geospatial data : algorithms, software tools and case studies

Résumé Cette thèse est consacrée à l'analyse, la modélisation et la visualisation de données environnementales à référence spatiale à l'aide d'algorithmes d'apprentissage automatique (Machine Learning). L'apprentissage automatique peut être considéré au sens large comme une sous-catégorie de l'intelligence artificielle qui concerne particulièrement le développement de techniques et d'algorithmes permettant à une machine d'apprendre à partir de données. Dans cette thèse, les algorithmes d'apprentissage automatique sont adaptés pour être appliqués à des données environnementales et à la prédiction spatiale. Pourquoi l'apprentissage automatique ? Parce que la majorité des algorithmes d'apprentissage automatiques sont universels, adaptatifs, non-linéaires, robustes et efficaces pour la modélisation. Ils peuvent résoudre des problèmes de classification, de régression et de modélisation de densité de probabilités dans des espaces à haute dimension, composés de variables informatives spatialisées (« géo-features ») en plus des coordonnées géographiques. De plus, ils sont idéaux pour être implémentés en tant qu'outils d'aide à la décision pour des questions environnementales allant de la reconnaissance de pattern à la modélisation et la prédiction en passant par la cartographie automatique. Leur efficacité est comparable au modèles géostatistiques dans l'espace des coordonnées géographiques, mais ils sont indispensables pour des données à hautes dimensions incluant des géo-features. Les algorithmes d'apprentissage automatique les plus importants et les plus populaires sont présentés théoriquement et implémentés sous forme de logiciels pour les sciences environnementales. Les principaux algorithmes décrits sont le Perceptron multicouches (MultiLayer Perceptron, MLP) - l'algorithme le plus connu dans l'intelligence artificielle, le réseau de neurones de régression généralisée (General Regression Neural Networks, GRNN), le réseau de neurones probabiliste (Probabilistic Neural Networks, PNN), les cartes auto-organisées (SelfOrganized Maps, SOM), les modèles à mixture Gaussiennes (Gaussian Mixture Models, GMM), les réseaux à fonctions de base radiales (Radial Basis Functions Networks, RBF) et les réseaux à mixture de densité (Mixture Density Networks, MDN). Cette gamme d'algorithmes permet de couvrir des tâches variées telle que la classification, la régression ou l'estimation de densité de probabilité. L'analyse exploratoire des données (Exploratory Data Analysis, EDA) est le premier pas de toute analyse de données. Dans cette thèse les concepts d'analyse exploratoire de données spatiales (Exploratory Spatial Data Analysis, ESDA) sont traités selon l'approche traditionnelle de la géostatistique avec la variographie expérimentale et selon les principes de l'apprentissage automatique. La variographie expérimentale, qui étudie les relations entre pairs de points, est un outil de base pour l'analyse géostatistique de corrélations spatiales anisotropiques qui permet de détecter la présence de patterns spatiaux descriptible par une statistique. L'approche de l'apprentissage automatique pour l'ESDA est présentée à travers l'application de la méthode des k plus proches voisins qui est très simple et possède d'excellentes qualités d'interprétation et de visualisation. Une part importante de la thèse traite de sujets d'actualité comme la cartographie automatique de données spatiales. Le réseau de neurones de régression généralisée est proposé pour résoudre cette tâche efficacement. Les performances du GRNN sont démontrées par des données de Comparaison d'Interpolation Spatiale (SIC) de 2004 pour lesquelles le GRNN bat significativement toutes les autres méthodes, particulièrement lors de situations d'urgence. La thèse est composée de quatre chapitres : théorie, applications, outils logiciels et des exemples guidés. Une partie importante du travail consiste en une collection de logiciels : Machine Learning Office. Cette collection de logiciels a été développée durant les 15 dernières années et a été utilisée pour l'enseignement de nombreux cours, dont des workshops internationaux en Chine, France, Italie, Irlande et Suisse ainsi que dans des projets de recherche fondamentaux et appliqués. Les cas d'études considérés couvrent un vaste spectre de problèmes géoenvironnementaux réels à basse et haute dimensionnalité, tels que la pollution de l'air, du sol et de l'eau par des produits radioactifs et des métaux lourds, la classification de types de sols et d'unités hydrogéologiques, la cartographie des incertitudes pour l'aide à la décision et l'estimation de risques naturels (glissements de terrain, avalanches). Des outils complémentaires pour l'analyse exploratoire des données et la visualisation ont également été développés en prenant soin de créer une interface conviviale et facile à l'utilisation. Machine Learning for geospatial data: algorithms, software tools and case studies Abstract The thesis is devoted to the analysis, modeling and visualisation of spatial environmental data using machine learning algorithms. In a broad sense machine learning can be considered as a subfield of artificial intelligence. It mainly concerns with the development of techniques and algorithms that allow computers to learn from data. In this thesis machine learning algorithms are adapted to learn from spatial environmental data and to make spatial predictions. Why machine learning? In few words most of machine learning algorithms are universal, adaptive, nonlinear, robust and efficient modeling tools. They can find solutions for the classification, regression, and probability density modeling problems in high-dimensional geo-feature spaces, composed of geographical space and additional relevant spatially referenced features. They are well-suited to be implemented as predictive engines in decision support systems, for the purposes of environmental data mining including pattern recognition, modeling and predictions as well as automatic data mapping. They have competitive efficiency to the geostatistical models in low dimensional geographical spaces but are indispensable in high-dimensional geo-feature spaces. The most important and popular machine learning algorithms and models interesting for geo- and environmental sciences are presented in details: from theoretical description of the concepts to the software implementation. The main algorithms and models considered are the following: multi-layer perceptron (a workhorse of machine learning), general regression neural networks, probabilistic neural networks, self-organising (Kohonen) maps, Gaussian mixture models, radial basis functions networks, mixture density networks. This set of models covers machine learning tasks such as classification, regression, and density estimation. Exploratory data analysis (EDA) is initial and very important part of data analysis. In this thesis the concepts of exploratory spatial data analysis (ESDA) is considered using both traditional geostatistical approach such as_experimental variography and machine learning. Experimental variography is a basic tool for geostatistical analysis of anisotropic spatial correlations which helps to understand the presence of spatial patterns, at least described by two-point statistics. A machine learning approach for ESDA is presented by applying the k-nearest neighbors (k-NN) method which is simple and has very good interpretation and visualization properties. Important part of the thesis deals with a hot topic of nowadays, namely, an automatic mapping of geospatial data. General regression neural networks (GRNN) is proposed as efficient model to solve this task. Performance of the GRNN model is demonstrated on Spatial Interpolation Comparison (SIC) 2004 data where GRNN model significantly outperformed all other approaches, especially in case of emergency conditions. The thesis consists of four chapters and has the following structure: theory, applications, software tools, and how-to-do-it examples. An important part of the work is a collection of software tools - Machine Learning Office. Machine Learning Office tools were developed during last 15 years and was used both for many teaching courses, including international workshops in China, France, Italy, Ireland, Switzerland and for realizing fundamental and applied research projects. Case studies considered cover wide spectrum of the real-life low and high-dimensional geo- and environmental problems, such as air, soil and water pollution by radionuclides and heavy metals, soil types and hydro-geological units classification, decision-oriented mapping with uncertainties, natural hazards (landslides, avalanches) assessments and susceptibility mapping. Complementary tools useful for the exploratory data analysis and visualisation were developed as well. The software is user friendly and easy to use.

Molecular characterization of the Carma1-related protein card9

ABSTRACT : Fungal infections have become a major source of diseases in immuncompromised patients, but are quite benign in healthy individuals. As fungi are eukaryotes, and share many biological processes with humans, many antifungal drugs can cause toxicity in the patients. Therefore, the characterization of signaling pathways specific to the anti-fungal immune response is relevant for the better understanding of the disease and the development of new therapeutic approaches. Dectin-1 is the major mammalian pattern recognition receptor for the fungal component zymosan. Dectin-1 is an innate non-Toll-like receptor containing immunoreceptor tyrosine-based activation motifs (ITAMs). Card9, Bc110 and Maltl are proteins that have been shown to play a key role in the Dectin-l-induced signaliñg pathway by controlling Dectin-l-mediated cell activation, cytokine production and innate anti-fungal immunity in mice. Here we investigate the role of the Card9-Bc110-Maltl complex in humans using the monocytic cell line THP-1. We show that Card9 interacts with Bc110 through a CARD-CARD interaction and that interaction of Card9 with Bc110 is required for NF-xB activation. We further demonstrate that Card9 is phosphorylated in its C-terminal part on serine residues. The phosphorylation status of Card9 can influence its ability to active NF-xB, since mutation of the phosphorylation sites increases its ability to activate NF-xB. We find that Card9 is expressed in myeloid derived cells, such as the human monocytic cell lines THP1 and U937, and in human monocyte-enriched PBLs and monocyte-derived DCs. Our findings demonstrate that Card9 is implicated in anti-fungal responses, since silencing of Card9 as well as of Bc110 and Maltl diminishes the capacity of THP1 cells to produce TNF-a in response to zymosan. Interestingly, activation of the NF-xB and MAPK pathway remained normal and levels of TNF-a mRNA produced were also not affected in THP 1 cells silenced for the expression of Card9, Bc110 or Malt1. Using a Malt1 inhibitor, we provide evidence that the proteolytic activity of Malt1 is needed for zymosan-induced TNF-a production in THP 1 cells and bone marrow-derived macrophages of mice, but further experiments are required to confirm these findings and identify the substrate(s) of Malt1. In conclusion, our results reveal an important role for Card9 in the innate immune response of human macrophages to fungi. RÉSUMÉ : Les infections fongiques sont une source majeure de maladie chez les patients immunodéprimés, alors qu'elles sont plutôt bénignes chez les individus sains. Comme les champignons sont des eucaryotes et partagent beaucoup de processus biologiques avec les humains, les médicaments antifongiques peuvent être source de toxicité chez les patients. Il est donc important de mieux caractériser les voies de signalisation intracellulaire des réponses anti-fongiques pour pouvoir développer de nouvelles approches thérapeutiques. La protéine Dectin-1 est le récepteur principal du composé fongique zymosan. Les protéines Card9, Bc110 et Maltl ont été décrites comme jouant un rôle primordial dans les signaux d'activation induits par Dectin-l, en contrôlant l'activité cellulaire, la production de cytokines et la défense anti-fongique dans les souris. Dans cette étude, nous investiguons le rôle du complexe Card9-Bc110-Maltl dans la lignée monocytaire humaine THP1. Nous montrons que Card9 interagit avec Bc110 par une interaction CARD-CARD et que cette interaction est requise pour activer le facteur de transcription NF-xB. Nous observons que Card9 est phosphorylé dans sa partie C-terminale sur des résidus serine et que l'état de phosphorylation de Card9 influence sa capacité à activer NF-xB. En effet, sa capacité à activer NF-xB est augmentée, après mutation des sites de phosphorylation. La génération d'un anticorps spécifique dirigé contre Card9 nous a permis de démontrer que Card9 est exprimé dans des cellules myéloïdes comme les lignées cellulaires monocytiques THP-1 et U-937, ainsi que dans les cellules dendritiques humaines. Nos résultats démontrent que Card9 est impliqué dans la réponse immunitaire antifongique puisque la réduction de l'expression de Card9 ainsi que de Bc110 et de Malt1 diminue la capacité des THP-1 à produire du TNF-a en réponse au zymosan. Par contre, les voies de signalisation NF-xB et MAPK ainsi que les niveaux de mRNA de TNF-a produits en réponse au zymosan ne sont pas affectés dans ces cellules. En utilisant un inhibiteur de Malt1, nous montrons que l'activité protéolytique de Malt1 est nécessaire pour la production de TNF-a induite par le zymosan dans les cellules THP-1 ainsi que dans les macrophages de souris, mais d'autres expériences seront nécessaires pour confirmer cette observation et identifier le(s) substrat(s) de Malt1 responsables de cet effet. En conclusion, nos résultats révèlent un rôle important de la protéine Card9 dans la réponse immunitaire innée antifongique dans les macrophages humains.

Classifying minimally disabled multiple sclerosis patients from resting state functional connectivity.

Multiple sclerosis (MS), a variable and diffuse disease affecting white and gray matter, is known to cause functional connectivity anomalies in patients. However, related studies published to-date are post hoc; our hypothesis was that such alterations could discriminate between patients and healthy controls in a predictive setting, laying the groundwork for imaging-based prognosis. Using functional magnetic resonance imaging resting state data of 22 minimally disabled MS patients and 14 controls, we developed a predictive model of connectivity alterations in MS: a whole-brain connectivity matrix was built for each subject from the slow oscillations (<0.11Hz) of region-averaged time series, and a pattern recognition technique was used to learn a discriminant function indicating which particular functional connections are most affected by disease. Classification performance using strict cross-validation yielded a sensitivity of 82% (above chance at p<0.005) and specificity of 86% (p<0.01) to distinguish between MS patients and controls. The most discriminative connectivity changes were found in subcortical and temporal regions, and contralateral connections were more discriminative than ipsilateral connections. The pattern of decreased discriminative connections can be summarized post hoc in an index that correlates positively (ρ=0.61) with white matter lesion load, possibly indicating functional reorganisation to cope with increasing lesion load. These results are consistent with a subtle but widespread impact of lesions in white matter and in gray matter structures serving as high-level integrative hubs. These findings suggest that predictive models of resting state fMRI can reveal specific anomalies due to MS with high sensitivity and specificity, potentially leading to new non-invasive markers.

Model-based objects recognition in industrial environments for autonomous vehicles control

Behavior-based navigation of autonomous vehicles requires the recognition of the navigable areas and the potential obstacles. In this paper we describe a model-based objects recognition system which is part of an image interpretation system intended to assist the navigation of autonomous vehicles that operate in industrial environments. The recognition system integrates color, shape and texture information together with the location of the vanishing point. The recognition process starts from some prior scene knowledge, that is, a generic model of the expected scene and the potential objects. The recognition system constitutes an approach where different low-level vision techniques extract a multitude of image descriptors which are then analyzed using a rule-based reasoning system to interpret the image content. This system has been implemented using a rule-based cooperative expert system

Model-based objects recognition in man-made environments

We describe a model-based objects recognition system which is part of an image interpretation system intended to assist autonomous vehicles navigation. The system is intended to operate in man-made environments. Behavior-based navigation of autonomous vehicles involves the recognition of navigable areas and the potential obstacles. The recognition system integrates color, shape and texture information together with the location of the vanishing point. The recognition process starts from some prior scene knowledge, that is, a generic model of the expected scene and the potential objects. The recognition system constitutes an approach where different low-level vision techniques extract a multitude of image descriptors which are then analyzed using a rule-based reasoning system to interpret the image content. This system has been implemented using CEES, the C++ embedded expert system shell developed in the Systems Engineering and Automatic Control Laboratory (University of Girona) as a specific rule-based problem solving tool. It has been especially conceived for supporting cooperative expert systems, and uses the object oriented programming paradigm