Two-layer particle filter for multiple target detection and tracking

This paper deals with the detection and tracking of an unknown number of targets using a Bayesian hierarchical model with target labels. To approximate the posterior probability density function, we develop a two-layer particle filter. One deals with track initiation, and the other with track maintenance. In addition, the parallel partition method is proposed to sample the states of the surviving targets.

La geoquímica de los sedimentos de llanuras de inundación como guía para la valoración ambiental en España

En la presente tesis se estudian los contenidos geoquímicos de los sedimentos de llanura de inundación en diversas cuencas fluviales seleccionadas con el objetivo de contribuir a un mejor conocimiento de sus condiciones medioambientales. En cada cuenca, se ha muestreado un perfil vertical de llanura de inundación, dividiéndolo en tramos que generalmente corresponden a diferentes episodios de inundación. Estos sedimentos se depositan durante los episodios de crecidas una vez que la corriente sobrepasa los límites del canal. Se caracterizan normalmente por tener un tamaño de grano fino y una estructura en capas horizontales que corresponden a los sucesivos episodios de inundación. Muestran dos ventajas principales con respecto a otros medios de muestreo en geoquímica como suelos o sedimentos de corriente: • Pueden almacenar sedimento antiguo así como actual, con lo que se puede estudiar la historia geoquímica de una zona específica en un mismo punto de muestreo (perfil vertical). • Los sedimentos de llanuras de inundación son capaces de caracterizar grandes áreas de drenaje. El origen de los sedimentos es más diverso que en el sedimento de corriente, debido a las mayores áreas de donde proceden las aguas de las avenidas. Las cuencas han sido seleccionadas según las actividades antropogénicas que en ellas se llevan a cabo, en concreto, actividades urbanas e industriales, minería y agricultura. Así mismo, se han estudiado, como referencia, dos cuencas donde no se espera encontrar ningún tipo de actividad contaminante. Una vez hecha la selección, los sedimentos aluviales de las cuencas se han estudiado cuidadosamente para asegurar que no existen depósitos de acreción lateral en el punto seleccionado. Posteriormente se ha procedido al muestreo del perfil vertical. Las muestras han sido analizadas mediante ICP-MS (ataque total) e INAA para conocer los contenidos totales de los elementos traza y mayoritarios. Los análisis de la fracción extraíble se han llevado a cabo mediante ICP-MS (ataque con agua regia). Así mismo, algunas muestras seleccionadas han sido sometidas a una extracción secuencial para un estudio más detallado. La presencia de materia orgánica ha sido estimada mediante el análisis de Carbono Orgánico Total (TOC). Finalmente, se ha llevado a cabo un análisis de isótopos de Pb en muestras escogidas en los perfiles, con el objetivo de hacer una evaluación ambiental. Los contenidos metálicos aumentan hacia la superficie en algunos de los perfiles, mientras en otros muestran una distribución muy constante exceptuando algún nivel específico con un aumento de los contenidos de la mayoría de los metales. Ha sido posible determinar la influencia de las actividades antropogénicas en algunos de los perfiles. Aquellos que pertenecen a cuencas mineras, urbanas o industrializadas muestran generalmente altos contenidos en elementos metálicos. Es el caso de los perfiles muestreados en los ríos Odiel y Tinto, Besaya, Besós y Manzanares. Algunos de estos perfiles pueden incluso correlacionarse con periodos de tiempo en los que ha tenido lugar una actividad antropogénica más intensa. Los perfiles que mejor se correlacionan con la actividad antropogénica de la cuenca son el perfil de Rivas en el río Manzanares (Madrid), que refleja un crecimiento de la contaminación producida por las actividades urbana e industrial en las últimas décadas, y el río Tinto, que muestra un crecimiento llamativo de los contenidos en su mayoría metálicos que puede estar relacionado con el incremento de la actividad minera que tuvo lugar hace aproximadamente 125 años. El análisis de los isótopos de Pb ha resultado ser una herramienta útil en la evaluación ambiental de estos sedimentos. Con este estudio y mediante la comparación con fuentes naturales y antropogénicas, ha sido posible diferenciar las muestras afectadas por diferentes fuentes de plomo, así como detectar las más afectadas antropogénicamente. ABSTRACT The geochemical composition of overbank sediments of some selected river basins is studied in this thesis in order to contribute to a better knowledge of the environmental conditions surrounding them. In each basin a vertical overbank profile has been sampled, dividing it into stretches that usually correspond to different flood events. The overbank sediments are those deposited during a flood event once the flow spills over the channel banks. They are usually characterized by a very fine grain size and a structure of horizontal layers, which correspond to successive flood events. These sediments show two main advantages regarding other sampling media in geochemistry, like soils or stream sediments: • They can store sediment deposited in the past as well as in current times, so that the history of a specific location can be studied at the very same point (vertical profile). • The overbank sediments are able to characterize a large drainage area. The origin of the sediment is wider than in the stream sediments due to the larger areas where the flood water comes from. The basins have been selected depending on the anthropogenic activities developed in them, namely, urban and industrial activities, mining activities and agricultural activities. As well, two pristine basins have been studied as a reference. Afterwards, the alluvial sediments in the basins have been carefully studied in order to sample a vertical profile and make sure that lateral accretion materials are not present in the profile. The samples have been analysed by ICP-MS (total digestion) and INAA to know the total contents of trace and major elements. Analysis of the mobile fraction has been carried out by ICP-MS (aqua regia); as well some of the samples have been subjected to sequential extraction for a more detailed study. The presence of organic matter has been estimated by the analysis of the Total Organic Carbon (TOC). Finally, a lead isotope analysis of some of the samples in the profiles was carried out in order to make an environmental assessment. Metal contents grow towards the surface in some of the profiles, while others show a very steady distribution, except for some of them with a growth of most of the metals in a specific level. XI It has been possible to determine the influence of the anthropogenic activities in some of the profiles. The ones that belong to mining and urban or industrialized basins show generally high contents of metal elements. This is the case of the profiles sampled in the Odiel and Tinto Rivers, the Besaya River, the Besós River and the Manzanares River. Some of these profiles can even correlate with the periods of time when a more intense activity in their respective basins has taken place. The profiles which best correlate with the anthropogenic activity are the Rivas profile in the Manzanares River, which reflects a growth of the pollution produced by urban and industrial activities in the city of Madrid in the last decades and the Tinto profile, which shows a very dramatic growth of the elemental contents (mostly metals) which can be related to the increase of the mining activities that took place in the last 125 years. The analysis of lead isotopes has turned out to be a powerful tool in the environmental assessment in this type of sediments. With this study and through the comparison with natural and anthropogenic sources, it has been possible to determine samples affected by different sources of lead and to detect the most anthropogenicaly affected ones.

Estudios de especiación de arsénico y acumulación de metales en muestras de interés medioambiental : Arsenic speciation and metal accumulation studies in environmental samples

Se ha estudiado la determinación de especies de arsénico y de contenidos totales de arsénico y metales pesados, específicamente cadmio, cromo, cobre, níquel, plomo y cinc, en muestras de interés medioambiental por su elevada capacidad acumuladora de metales, concretamente algas marinas comestibles y plantas terrestres procedentes de suelos contaminados por la actividad minera. La determinación de contenidos totales se ha llevado a cabo mediante espectrometría de emisión atómica con plasma de acoplamiento inductivo (ICP‐AES), así como por espectrometría de fluorescencia atómica con generación de hidruros (HG‐AFS), para bajos contenidos de arsénico. Las muestras fueron mineralizadas en medio ácido y calentamiento en horno de microondas. Los métodos fueron validados a través de su aplicación a materiales de referencia de matriz similar a la de las muestras, certificados en contenidos totales de los elementos seleccionados. Los resultados obtenidos mostraron su elevada capacidad de bioabsorción, especialmente en relación a los elevados contenidos de arsénico encontrados en algunas especies de algas pardas (Phaeophytas). En las plantas, se calcularon los factores de translocación, acumulación y biodisponibilidad de los elementos estudiados, permitiendo identificar a la especie Corrigiola telephiifolia como posible acumuladora de plomo e hiperacumuladora de arsénico. La determinación de especies de arsénico hidrosolubles en las muestras objeto de estudio, se llevó a cabo por cromatografía líquida de alta eficacia (HPLC) acoplado a ICP‐AES, HG‐ICP‐AES y HG‐AFS, incluyendo una etapa previa de foto‐oxidación. Los métodos desarrollados, mediante intercambio aniónico y catiónico, permitieron la diferenciación de hasta once especies de arsénico. Para el análisis de las muestras, fue necesaria la optimización de métodos de extracción, seleccionándose la extracción asistida por microondas (MAE) con agua desionizada. Asimismo, se realizaron estudios de estabilidad de arsénico total y de las especies hidrosolubles presentes en las algas, tanto sobre la muestra sólida como en sus extractos acuosos, evaluando las condiciones de almacenamiento adecuadas. En el caso de las plantas, la aplicación del diseño factorial de experimentos permitió optimizar el método de extracción y diferenciar entre las especies de arsénico presentes en forma de iones sencillos de mayor movilidad y el arsénico más fuertemente enlazado a componentes estructurales. Los resultados obtenidos permitieron identificar la presencia de arseniato (As(V)) y arsenito (As(III)) en las plantas, así como de ácido monometilarsónico (MMA) y óxido de trimetilarsina (TMAO) en algunas especies. En la mayoría de las algas se encontraron especies tóxicas, tanto mayoritarias (arseniato) como minoritarias (ácido dimetilarsínico (DMA)), así como hasta cuatro arsenoazúcares. Los resultados obtenidos y su estudio a través de la legislación vigente, mostraron la necesidad de desarrollar una reglamentación específica para el control de este tipo de alimentos. La determinación de especies de arsénico liposolubles en las muestras de algas se llevó a cabo mediante HPLC, en modo fase inversa, acoplado a espectrometría de masas con plasma de acoplamiento inductivo (ICP‐MS) y con ionización por electrospray (ESI‐MS), permitiendo la elucidación estructural de estos compuestos a través de la determinación de sus masas moleculares. Para ello, fue necesaria la puesta a punto de métodos extracción y purificación de los extractos. La metodología desarrollada permitió identificar hasta catorce especies de arsénico liposolubles en las algas, tres de ellas correspondientes a hidrocarburos que contienen arsénico, y once a arsenofosfolípidos, además de dos especies desconocidas. Las masas moleculares de las especies identificadas fueron confirmadas mediante cromatografía de gases acoplada a espectrometría de masas (GC‐MS) y espectrometría de masas de alta resolución (HR‐MS). ABSTRACT The determination of arsenic species and total arsenic and heavy metal contents (cadmium, chromium, cooper, nickel, lead and zinc) in environmental samples, with high metal accumulator capacity, has been studied. The samples studied were edible marine algae and terrestrial plants from soils polluted by mining activities. The determination of total element contents was performed by inductively coupled plasma atomic emission spectrometry (ICP‐AES), as well as by hydride generation atomic fluorescence spectrometry (HG‐AFS) for low arsenic contents. The samples studied were digested in an acidic medium by heating in a microwave oven. The digestion methods were validated against reference materials, with matrix similar to sample matrix and certified in total contents of the elements studied. The results showed the high biosorption capacity of the samples studied, especially regarding the high arsenic contents in some species of brown algae (Phaeophyta division). In terrestrial plants, the translocation, accumulation and bioavailability factors of the elements studied were calculated. Thus, the plant species Corrigiola telephiifolia was identified as possible lead accumulator and arsenic hyperaccumulator. The determination of water‐soluble arsenic species in the samples studied was carried out by high performance liquid chromatography (HPLC) coupled to ICP‐AES, HG‐ICP‐AES and HG‐AFS, including a prior photo‐oxidation step. The chromatographic methods developed, by anion and cation exchange, allowed us to differentiate up to eleven arsenic species. The sample analysis required the optimization of extraction methods, choosing the microwave assisted extraction (MAE) with deionized water. On the other hand, the stability of total arsenic and water‐soluble arsenic species in algae, both in the solid samples and in the water extracts, was studied, assessing the suitable storage conditions. In the case of plant samples, the application of a multivariate experimental design allowed us to optimize the extraction method and differentiate between the arsenic species present as simple ions of higher mobility and the arsenic more strongly bound to structural components. The presence of arsenite (As(III)) and arsenate (As(V)) was identified in plant samples, as well as monomethylarsonic acid (MMA) and trimethylarsine oxide (TMAO) in some cases. Regarding algae, toxic arsenic species were found in most of them, both As(V) and dimethylarsinic acid (DMA), as well as up to four arsenosugars. These results were discussed according to the current legislation, showing the need to develop specific regulations to control this kind of food products. The determination of lipid‐soluble arsenic species in alga samples was performed by reversed‐phase HPLC coupled to inductively coupled plasma and electrospray mass spectrometry (ICP‐MS and ESI‐MS), in order to establish the structure of these compounds by determining the corresponding molecular masses. For this purpose, it was necessary to develop an extraction method, as well as a clean‐up method of the extracts. The method developed permitted the identification of fourteen lipid‐soluble arsenic compounds in algae, corresponding to three arsenic‐hydrocarbons and eleven arsenosugarphospholipids, as well as two unknown compounds. Accurate mass measurements of the identified compounds were performed by gas chromatography coupled to mass spectrometry (GC‐MS) and high resolution mass spectrometry (HR‐MS).

Node sampling using random centrifugal walks

Sampling a network with a given probability distribution has been identified as a useful operation. In this paper we propose distributed algorithms for sampling networks, so that nodes are selected by a special node, called the source, with a given probability distribution. All these algorithms are based on a new class of random walks, that we call Random Centrifugal Walks (RCW). A RCW is a random walk that starts at the source and always moves away from it. Firstly, an algorithm to sample any connected network using RCW is proposed. The algorithm assumes that each node has a weight, so that the sampling process must select a node with a probability proportional to its weight. This algorithm requires a preprocessing phase before the sampling of nodes. In particular, a minimum diameter spanning tree (MDST) is created in the network, and then nodes weights are efficiently aggregated using the tree. The good news are that the preprocessing is done only once, regardless of the number of sources and the number of samples taken from the network. After that, every sample is done with a RCW whose length is bounded by the network diameter. Secondly, RCW algorithms that do not require preprocessing are proposed for grids and networks with regular concentric connectivity, for the case when the probability of selecting a node is a function of its distance to the source. The key features of the RCW algorithms (unlike previous Markovian approaches) are that (1) they do not need to warm-up (stabilize), (2) the sampling always finishes in a number of hops bounded by the network diameter, and (3) it selects a node with the exact probability distribution.

Caracterización e instrumentación de sensores de ultrasonidos

El propósito de este proyecto de �fin de carrera es la caracterización e instrumentación de un sensor de ultrasonidos modelado por el tutor de este proyecto: Don César Briso Rodrí��guez. Una vez realizado el modelado de dicho sensor, simulando tanto sus caracter�í�sticas f�í�sicas, como sus caracterí��sticas eléctricas, se procede a la intrumentación y uso del mismo. La parte de intrumentaci�ón incluye tanto la electrónica que ser��á necesaria para la excitación del piezoeléctrico, en el modo de emisi�ón, como para la recepción de los pulsos el�éctricos generados por el sensor, como respuesta a los ecos recibidos, y su adecuación a niveles de señal correctos para la adquisici�ón, en el modo de escucha. Tras la adecuaci�ón de las señales para la adquisici�ón, éstas ser�án digitalizadas, tratadas y representadas por pantalla en un PC, a trav�es de una tarjeta de adquisición de datos por puerto USB encargada del muestreo de las señales de respuesta ya tratadas y su posterior enví��o al software de control y representaci�ón desarrollado en este proyecto. El entorno de usuario, el software de control de la tarjeta de adquisición y el software de tratamiento y representaci�ón se ha desarrollado con Visual Basic 2008 y las utilidades gr�áfi�cas de las librer��ías OpenGL. ABSTRACT The purpose of this project is to limit the characterization and implementation of an ultrasonic sensor modeled by Mr. C�ésar Briso Rodr��íguez. Once the sensor modeling by simulating physical characteristics and electrical characteristics, we proceed to the instrumentation and use. This section includes electronic instrumentation that would be necessary for the piezoelectric excitation in the emission mode and for receiving electrical pulses generated by the sensor in response to the received echoes, and matching signal levels right to acquire, in the reception mode. After the adjustment of the signals for the acquisition, these signals will be digitalized, processed and represented on the screen on a PC through a data acquisition card by USB port. Acquisition card is able to sample the response signals and transmit the samples to representation and control software developed in this project. The user interface, the acquisition card control software and processing and representation software has been developed with Visual Basic 2008 and OpenGL graphical libraries.

Massively parallelizable list-mode reconstruction using a Monte Carlo-based elliptical Gaussian model

Purpose: A fully three-dimensional (3D) massively parallelizable list-mode ordered-subsets expectation-maximization (LM-OSEM) reconstruction algorithm has been developed for high-resolution PET cameras. System response probabilities are calculated online from a set of parameters derived from Monte Carlo simulations. The shape of a system response for a given line of response (LOR) has been shown to be asymmetrical around the LOR. This work has been focused on the development of efficient region-search techniques to sample the system response probabilities, which are suitable for asymmetric kernel models, including elliptical Gaussian models that allow for high accuracy and high parallelization efficiency. The novel region-search scheme using variable kernel models is applied in the proposed PET reconstruction algorithm. Methods: A novel region-search technique has been used to sample the probability density function in correspondence with a small dynamic subset of the field of view that constitutes the region of response (ROR). The ROR is identified around the LOR by searching for any voxel within a dynamically calculated contour. The contour condition is currently defined as a fixed threshold over the posterior probability, and arbitrary kernel models can be applied using a numerical approach. The processing of the LORs is distributed in batches among the available computing devices, then, individual LORs are processed within different processing units. In this way, both multicore and multiple many-core processing units can be efficiently exploited. Tests have been conducted with probability models that take into account the noncolinearity, positron range, and crystal penetration effects, that produced tubes of response with varying elliptical sections whose axes were a function of the crystal's thickness and angle of incidence of the given LOR. The algorithm treats the probability model as a 3D scalar field defined within a reference system aligned with the ideal LOR. Results: This new technique provides superior image quality in terms of signal-to-noise ratio as compared with the histogram-mode method based on precomputed system matrices available for a commercial small animal scanner. Reconstruction times can be kept low with the use of multicore, many-core architectures, including multiple graphic processing units. Conclusions: A highly parallelizable LM reconstruction method has been proposed based on Monte Carlo simulations and new parallelization techniques aimed at improving the reconstruction speed and the image signal-to-noise of a given OSEM algorithm. The method has been validated using simulated and real phantoms. A special advantage of the new method is the possibility of defining dynamically the cut-off threshold over the calculated probabilities thus allowing for a direct control on the trade-off between speed and quality during the reconstruction.

Hidrodinamica de planeadores submarinos y su impacto en la funcionalidad de sensores CTD

Los recientes desarrollos tecnológicos permiten la transición de la oceanografía observacional desde un concepto basado en buques a uno basado en sistemas autónomos en red. Este último, propone que la forma más eficiente y efectiva de observar el océano es con una red de plataformas autónomas distribuidas espacialmente y complementadas con sistemas de medición remota. Debido a su maniobrabilidad y autonomía, los planeadores submarinos están jugando un papel relevante en este concepto de observaciones en red. Los planeadores submarinos fueron específicamente diseñados para muestrear vastas zonas del océano. Estos son robots con forma de torpedo que hacen uso de su forma hidrodinámica, alas y cambios de flotabilidad para generar movimientos horizontales y verticales en la columna de agua. Un sensor que mide conductividad, temperatura y profundidad (CTD) constituye un equipamiento estándar en la plataforma. Esto se debe a que ciertas variables dinámicas del Océano se pueden derivar de la temperatura, profundidad y salinidad. Esta última se puede estimar a partir de las medidas de temperatura y conductividad. La integración de sensores CTD en planeadores submarinos no esta exenta de desafíos. Uno de ellos está relacionado con la precisión de los valores de salinidad derivados de las muestras de temperatura y conductividad. Específicamente, las estimaciones de salinidad están significativamente degradadas por el retardo térmico existente, entre la temperatura medida y la temperatura real dentro de la celda de conductividad del sensor. Esta deficiencia depende de las particularidades del flujo de entrada al sensor, su geometría y, también se ha postulado, del calor acumulado en las capas de aislamiento externo del sensor. Los efectos del retardo térmico se suelen mitigar mediante el control del flujo de entrada al sensor. Esto se obtiene generalmente mediante el bombeo de agua a través del sensor o manteniendo constante y conocida su velocidad. Aunque recientemente se han incorporado sistemas de bombeo en los CTDs a bordo de los planeadores submarinos, todavía existen plataformas equipadas con CTDs sin dichos sistemas. En estos casos, la estimación de la salinidad supone condiciones de flujo de entrada al sensor, razonablemente controladas e imperturbadas. Esta Tesis investiga el impacto, si existe, que la hidrodinámica de los planeadores submarinos pudiera tener en la eficiencia de los sensores CTD. Específicamente, se investiga primero la localización del sensor CTD (externo al fuselaje) relativa a la capa límite desarrollada a lo largo del cuerpo del planeador. Esto se lleva a cabo mediante la utilización de un modelo acoplado de fluido no viscoso con un modelo de capa límite implementado por el autor, así como mediante un programa comercial de dinámica de fluidos computacional (CFD). Los resultados indican, en ambos casos, que el sensor CTD se encuentra fuera de la capa límite, siendo las condiciones del flujo de entrada las mismas que las del flujo sin perturbar. Todavía, la velocidad del flujo de entrada al sensor CTD es la velocidad de la plataforma, la cual depende de su hidrodinámica. Por tal motivo, la investigación se ha extendido para averiguar el efecto que la velocidad de la plataforma tiene en la eficiencia del sensor CTD. Con este propósito, se ha desarrollado un modelo en elementos finitos del comportamiento hidrodinámico y térmico del flujo dentro del CTD. Los resultados numéricos indican que el retardo térmico, atribuidos originalmente a la acumulación de calor en la estructura del sensor, se debe fundamentalmente a la interacción del flujo que atraviesa la celda de conductividad con la geometría interna de la misma. Esta interacción es distinta a distintas velocidades del planeador submarino. Específicamente, a velocidades bajas del planeador (0.2 m/s), la mezcla del flujo entrante con las masas de agua remanentes en el interior de la celda, se ralentiza debido a la generación de remolinos. Se obtienen entonces desviaciones significantes entre la salinidad real y aquella estimada. En cambio, a velocidades más altas del planeador (0.4 m/s) los procesos de mezcla se incrementan debido a la turbulencia e inestabilidades. En consecuencia, la respuesta del sensor CTD es mas rápida y las estimaciones de la salinidad mas precisas que en el caso anterior. Para completar el trabajo, los resultados numéricos se han validado con pruebas experimentales. Específicamente, se ha construido un modelo a escala del sensor CTD para obtener la confirmación experimental de los modelos numéricos. Haciendo uso del principio de similaridad de la dinámica que gobierna los fluidos incompresibles, los experimentos se han realizado con flujos de aire. Esto simplifica significativamente la puesta experimental y facilita su realización en condiciones con medios limitados. Las pruebas experimentales han confirmado cualitativamente los resultados numéricos. Más aun, se sugiere en esta Tesis que la respuesta del sensor CTD mejoraría significativamente añadiendo un generador de turbulencia en localizaciones adecuadas al interno de la celda de conductividad. ABSTRACT Recent technological developments allow the transition of observational oceanography from a ship-based to a networking concept. The latter suggests that the most efficient and effective way to observe the Ocean is through a fleet of spatially distributed autonomous platforms complemented by remote sensing. Due to their maneuverability, autonomy and endurance at sea, underwater gliders are already playing a significant role in this networking observational approach. Underwater gliders were specifically designed to sample vast areas of the Ocean. These are robots with a torpedo shape that make use of their hydrodynamic shape, wings and buoyancy changes to induce horizontal and vertical motions through the water column. A sensor to measure the conductivity, temperature and depth (CTD) is a standard payload of this platform. This is because certain ocean dynamic variables can be derived from temperature, depth and salinity. The latter can be inferred from measurements of temperature and conductivity. Integrating CTD sensors in glider platforms is not exempted of challenges. One of them, concerns to the accuracy of the salinity values derived from the sampled conductivity and temperature. Specifically, salinity estimates are significantly degraded by the thermal lag response existing between the measured temperature and the real temperature inside the conductivity cell of the sensor. This deficiency depends on the particularities of the inflow to the sensor, its geometry and, it has also been hypothesized, on the heat accumulated by the sensor coating layers. The effects of thermal lag are usually mitigated by controlling the inflow conditions through the sensor. Controlling inflow conditions is usually achieved by pumping the water through the sensor or by keeping constant and known its diving speed. Although pumping systems have been recently implemented in CTD sensors on board gliders, there are still platforms with unpumped CTDs. In the latter case, salinity estimates rely on assuming reasonable controlled and unperturbed flow conditions at the CTD sensor. This Thesis investigates the impact, if any, that glider hydrodynamics may have on the performance of onboard CTDs. Specifically, the location of the CTD sensor (external to the hull) relative to the boundary layer developed along the glider fuselage, is first investigated. This is done, initially, by applying a coupled inviscid-boundary layer model developed by the author, and later by using a commercial software for computational fluid dynamics (CFD). Results indicate, in both cases, that the CTD sensor is out of the boundary layer, being its inflow conditions those of the free stream. Still, the inflow speed to the CTD sensor is the speed of the platform, which largely depends on its hydrodynamic setup. For this reason, the research has been further extended to investigate the effect of the platform speed on the performance of the CTD sensor. A finite element model of the hydrodynamic and thermal behavior of the flow inside the CTD sensor, is developed for this purpose. Numerical results suggest that the thermal lag effect is mostly due to the interaction of the flow through the conductivity cell and its geometry. This interaction is different at different speeds of the glider. Specifically, at low glider speeds (0.2 m/s), the mixing of recent and old waters inside the conductivity cell is slowed down by the generation of coherent eddy structures. Significant departures between real and estimated values of the salinity are found. Instead, mixing is enhanced by turbulence and instabilities for high glider speeds (0.4 m/s). As a result, the thermal response of the CTD sensor is faster and the salinity estimates more accurate than for the low speed case. For completeness, numerical results have been validated against model tests. Specifically, a scaled model of the CTD sensor was built to obtain experimental confirmation of the numerical results. Making use of the similarity principle of the dynamics governing incompressible fluids, experiments are carried out with air flows. This significantly simplifies the experimental setup and facilitates its realization in a limited resource condition. Model tests qualitatively confirm the numerical findings. Moreover, it is suggested in this Thesis that the response of the CTD sensor would be significantly improved by adding small turbulators at adequate locations inside the conductivity cell.

Modulation of cognition-specific cortical activity by gonadal steroids: A positron-emission tomography study in women

There is considerable evidence from animal studies that gonadal steroid hormones modulate neuronal activity and affect behavior. To study this in humans directly, we used H215O positron-emission tomography to measure regional cerebral blood flow (rCBF) in young women during three pharmacologically controlled hormonal conditions spanning 4–5 months: ovarian suppression induced by the gonadotropin-releasing hormone agonist leuprolide acetate (Lupron), Lupron plus estradiol replacement, and Lupron plus progesterone replacement. Estradiol and progesterone were administered in a double-blind cross-over design. On each occasion positron-emission tomography scans were performed during (i) the Wisconsin Card Sorting Test, a neuropsychological test that physiologically activates prefrontal cortex (PFC) and an associated cortical network including inferior parietal lobule and posterior inferolateral temporal gyrus, and (ii) a no-delay matching-to-sample sensorimotor control task. During treatment with Lupron alone (i.e., with virtual absence of gonadal steroid hormones), there was marked attenuation of the typical Wisconsin Card Sorting Test activation pattern even though task performance did not change. Most strikingly, there was no rCBF increase in PFC. When either progesterone or estrogen was added to the Lupron regimen, there was normalization of the rCBF activation pattern with augmentation of the parietal and temporal foci and return of the dorsolateral PFC activation. These data directly demonstrate that the hormonal milieu modulates cognition-related neural activity in humans.

Dynamic weighting in Monte Carlo and optimization

Dynamic importance weighting is proposed as a Monte Carlo method that has the capability to sample relevant parts of the configuration space even in the presence of many steep energy minima. The method relies on an additional dynamic variable (the importance weight) to help the system overcome steep barriers. A non-Metropolis theory is developed for the construction of such weighted samplers. Algorithms based on this method are designed for simulation and global optimization tasks arising from multimodal sampling, neural network training, and the traveling salesman problem. Numerical tests on these problems confirm the effectiveness of the method.

CD1c molecules broadly survey the endocytic system

The ability of antigen-presenting cells to sample distinct intracellular compartments is crucial for microbe detection. Major histocompatibility complex class I and class II molecules sample the cytosol or the late endocytic compartment, allowing detection of microbial peptide antigens that arise in distinct intracellular compartments. In contrast, CD1a and CD1b molecules mediate the presentation of lipid and glycolipid antigens and differentially sample early recycling endosomes or late endocytic compartments, respectively, that contain distinct sets of lipid antigens. Here, we show that, unlike the other CD1 isoforms or major histocompatibility complex molecules that each sample restricted only intracellular compartments, CD1c is remarkable in that it distributes broadly throughout the endocytic system and is expressed in both recycling endosomes and late endocytic compartments. Further, in contrast to CD1b, which requires an acidic environment to function, antigen presentation by CD1c was able to overcome dependence on vesicular acidification. Because CD1c is expressed on essential antigen-presenting cells, such as epidermal Langerhans cells (in the absence of CD1b), or on B cells (without CD1a or -b), we suggest that CD1c molecules allow a comprehensive survey for lipid antigens throughout the endocytic system even in the absence of other CD1 isoforms.

Ensemble codes involving hippocampal neurons are at risk during delayed performance tests

Multielectrode recording techniques were used to record ensemble activity from 10 to 16 simultaneously active CA1 and CA3 neurons in the rat hippocampus during performance of a spatial delayed-nonmatch-to-sample task. Extracted sources of variance were used to assess the nature of two different types of errors that accounted for 30% of total trials. The two types of errors included ensemble “miscodes” of sample phase information and errors associated with delay-dependent corruption or disappearance of sample information at the time of the nonmatch response. Statistical assessment of trial sequences and associated “strength” of hippocampal ensemble codes revealed that miscoded error trials always followed delay-dependent error trials in which encoding was “weak,” indicating that the two types of errors were “linked.” It was determined that the occurrence of weakly encoded, delay-dependent error trials initiated an ensemble encoding “strategy” that increased the chances of being correct on the next trial and avoided the occurrence of further delay-dependent errors. Unexpectedly, the strategy involved “strongly” encoding response position information from the prior (delay-dependent) error trial and carrying it forward to the sample phase of the next trial. This produced a miscode type error on trials in which the “carried over” information obliterated encoding of the sample phase response on the next trial. Application of this strategy, irrespective of outcome, was sufficient to reorient the animal to the proper between trial sequence of response contingencies (nonmatch-to-sample) and boost performance to 73% correct on subsequent trials. The capacity for ensemble analyses of strength of information encoding combined with statistical assessment of trial sequences therefore provided unique insight into the “dynamic” nature of the role hippocampus plays in delay type memory tasks.

Mnemonic neuronal activity in somatosensory cortex.

Single-unit activity was recorded from the hand areas of the somatosensory cortex of monkeys trained to perform a haptic delayed matching to sample task with objects of identical dimensions but different surface features. During the memory retention period of the task (delay), many units showed sustained firing frequency change, either excitation or inhibition. In some cases, firing during that period was significantly higher after one sample object than after another. These observations indicate the participation of somatosensory neurons not only in the perception but in the short-term memory of tactile stimuli. Neurons most directly implicated in tactile memory are (i) those with object-selective delay activity, (ii) those with nondifferential delay activity but without activity related to preparation for movement, and (iii) those with delay activity in the haptic-haptic delayed matching task but no such activity in a control visuo-haptic delayed matching task. The results indicate that cells in early stages of cortical somatosensory processing participate in haptic short-term memory.

Aprendizagem de substantivos e adjetivos por meio de leitura compartilhada para crianças com desenvolvimento típico e com Síndrome de Down

A leitura compartilhada de livros para crianças é uma atividade que tem sido estudada como forma de ensino incidental de vocabulário, que envolve, dentre outros processos, o responder por exclusão. O objetivo do presente trabalho foi investigar a ocorrência de aprendizagem de relações entre estímulos visuais (figuras) com seus respectivos estímulos auditivos (palavras) a partir de diferentes condições de leitura compartilhada de livros para crianças com Síndrome de Down (SD) e com desenvolvimento típico (DT). Para a pesquisa foram desenvolvidos dois estudos. No Estudo 1, participaram seis crianças com SD com seis a sete anos, e seis crianças com DT com três a quatro anos (amostras pareadas em função do nível de vocabulário). Foi utilizado um livro de história produzido pela pesquisadora, no qual havia dois substantivos e dois adjetivos desconhecidos (estímulos visuais S1, S2, A1, A2), apresentados uma única vez na história. Esse livro foi lido para cada criança duas vezes em sequência por sessão e em cada sessão foi realizada uma condição de leitura diferente. Foram apresentadas três condições de leitura e cada criança passou por todas, mas em diferentes ordens (contrabalanceamento). Na Condição 1, o livro foi lido para a criança sem intervenções. Na Condição 2, o livro foi lido para a criança e ela tinha que repetir o nome dos estímulos desconhecidos. Na Condição 3, o livro foi lido e foram realizadas perguntas relacionadas aos estímulos-alvo. Ao final de cada sessão foram realizadas sondas de aprendizagem (sondas de emparelhamento ao modelo e nomeação), e após uma semana da última sessão foi aplicada uma sonda de manutenção e uma de generalização. As crianças com DT apresentaram maior número de acertos que as com SD, e os acertos foram mais relacionados ao estímulo S1. As crianças não aprenderam a relação nome-cor. A análise dos resultados sugeriu que o número de estímulos-alvo era excessivo e com apresentações insuficientes no livro. No Estudo 2 participaram seis crianças com DT de 3 a 4 anos e seis crianças com SD, de 5 a 8 anos. O procedimento utilizado no Estudo 2 foi semelhante ao primeiro com as seguintes alterações no livro: utilização de apenas duas relações-alvo (um substantivo-alvo e um adjetivo-alvo - S2 e A3), cada uma sendo apresentada três vezes ao longo da história, em figuras que possibilitavam o responder por exclusão. Também foi acrescentada uma tentativa de exclusão nas sondas de aprendizagem. Nesse estudo, todas as crianças com DT conseguiram selecionar e nomear estímulo S2 e duas mostraram indícios de aprendizagem do estímulo A3. As crianças com SD apresentaram um menor número de acertos nas sondas de emparelhamento, mas apresentaram algumas nomeações corretas, o que não foi observado no Estudo 1. Os dados sugerem que as mudanças realizadas no livro melhoram o desempenho das crianças com DT, mas não o das crianças com SD. Não foram encontradas diferenças entre as condições de leituras nos dois estudos. No entanto, são necessários estudos adicionais para avaliar essas diferentes condições e as variáveis envolvidas na aprendizagem de palavras a partir da leitura compartilhada de livro.

Otimização dos valores de referência da 17OH-progesterona neonatal de acordo com o peso ao nascimento no programa de triagem neonatal da APAE DE SÃO PAULO

Introdução: A Hiperplasia Adrenal Congênita por deficiência da 21-hidroxilase (HAC) é uma doença com mortalidade neonatal elevada sendo elegível para programas públicos de Triagem Neonatal (TN). A HAC é causada por mutações no gene CYP21A2, as quais acarretam diferentes comprometimentos da atividade enzimática e resultam em espectro amplo de manifestações clínicas. Apesar da eficiência da TN para diagnosticar os casos graves, a taxa elevada de resultados falso-positivos (RFP), principalmente relacionados à prematuridade, é um dos maiores problemas. Porém, resultados falso-negativos também podem ocorrer em coletas antes de 24 horas de vida. No Brasil, a coleta da amostra neonatal difere entre os municípios, podendo ser no terceiro dia de vida como após. Objetivo: Avaliar se os valores da 17OH-progesterona neonatal (N17OHP) das coletas no terceiro dia de vida diferem significativamente das coletas a partir do quarto dia. Determinar qual percentil (99,5 ou 99,8) pode ser utilizado como valor de corte para a N17OHP, de acordo com o peso ao nascimento e tempo de vida na coleta, a fim de que proporcione taxa menor de RFP. Métodos: Foi avaliada, retrospectivamente, a N17OHP de 271.810 recém-nascidos (Rns) de acordo com o tempo de vida na coleta (G1: 48 - = 72h) e peso ao nascimento (P1: <= 1.500g, P2: 1.501-2.000g, P3: 2.001-2.500g e P4: >= 2.500g), pelo método imunofluorimétrico. Testes com resultados alterados foram confirmados no soro por Espectrometria de Massas em Tandem - LC-MS/MS. Rns afetados e/ou assintomáticos e com valores persistentemente elevados de 17OHP sérica foram submetidos ao estudo molécular, sequenciamento do gene CYP21A2. Resultados: os valores da N17OHP no grupo G1 foram significativamente menores do que em G2 em todos os grupos de peso (p < 0.001). A taxa de RFP em G1 e G2 foi de 0,2% para o percentil 99,8 e de 0,5% para o percentil 99,5 em ambos os grupos. O percentil 99,8 da N17OHP foi o melhor valor de corte para distinguir os Rns não afetados dos afetados, cujos valores são: G1 (P1: 120; P2: 71; P3: 39 e P4: 20 ng /mL) e em G2 (P1: 173; P2: 90; P3: 66 e P4: 25 ng/mL). Vinte e seis Rns do grupo G1 apresentaram a forma perdedora de sal (PS) (13H e 13M), nestes a N17OHP variou de 31 a 524 ng/mL e vinte Rns no grupo G2 (8H e 12M), nestes a N17OHP variou de 53 a 736 ng/mL. Para ambos os grupos foram encontrados três Rns com a forma virilizante simples (1H e 2M) e os valores da N17OHP variaram de 36 a 51 ng/mL. Resultados falso-negativos não foram relatados. O valor preditivo positivo (VPP) no teste do papel filtro foi de 5,6% e 14,1% nos grupos G1 e G2, respectivamente, ao se utilizar o percentil 99,8, e de 2,3% e 7% nos grupos G1 e G2 ao se utilizar o percentil 99,5. Dentre os casos com TN alterada (RFP), 29 deles também apresentaram 17OHP sérica elevada quando dosada por LC-MS/MS. Os casos assintomáticos foram acompanhados até normalização da 17OHP sérica e/ou submetidos ao estudo molecular, que identificou dois Rns com genótipo que prediz a forma não clássica. Conclusão: a melhor estratégia para otimização do diagnóstico da HAC na triagem neonatal é se padronizar valores de corte da N17OHP em dois grupos de acordo com o tempo de vida na coleta (antes e depois de 72 horas), subdivididos em quatro grupos de peso. A utilização dos valores de corte do percentil 99,8 se mantém eficaz no diagnóstico da HAC-21OH na triagem neonatal, reduzindo de forma significativa a taxa de RFP, sem perda do diagnóstico da forma PS

Sampling scarab beetles in tropical forests: the effect of light source and night sampling periods

Light traps have been used widely to sample insect abundance and diversity, but their performance for sampling scarab beetles in tropical forests based on light source type and sampling hours throughout the night has not been evaluated. The efficiency of mercury-vapour lamps, cool white light and ultraviolet light sources in attracting Dynastinae, Melolonthinae and Rutelinae scarab beetles, and the most adequate period of the night to carry out the sampling was tested in different forest areas of Costa Rica. Our results showed that light source wavelengths and hours of sampling influenced scarab beetle catches. No significant differences were observed in trap performance between the ultraviolet light and mercury-vapour traps, whereas these two methods caught significantly more species richness and abundance than cool white light traps. Species composition also varied between methods. Large differences appear between catches in the sampling period, with the first five hours of the night being more effective than the last five hours. Because of their high efficiency and logistic advantages, we recommend ultraviolet light traps deployed during the first hours of the night as the best sampling method for biodiversity studies of those scarab beetles in tropical forests.