Estimation of signal subspace on hyperspectral data

Dimensionality reduction plays a crucial role in many hyperspectral data processing and analysis algorithms. This paper proposes a new mean squared error based approach to determine the signal subspace in hyperspectral imagery. The method first estimates the signal and noise correlations matrices, then it selects the subset of eigenvalues that best represents the signal subspace in the least square sense. The effectiveness of the proposed method is illustrated using simulated and real hyperspectral images.

Unsupervised hyperspectral signal subspace identification

Hyperspectral imaging sensors provide image data containing both spectral and spatial information from the Earth surface. The huge data volumes produced by these sensors put stringent requirements on communications, storage, and processing. This paper presents a method, termed hyperspectral signal subspace identification by minimum error (HySime), that infer the signal subspace and determines its dimensionality without any prior knowledge. The identification of this subspace enables a correct dimensionality reduction yielding gains in algorithm performance and complexity and in data storage. HySime method is unsupervised and fully-automatic, i.e., it does not depend on any tuning parameters. The effectiveness of the proposed method is illustrated using simulated data based on U.S.G.S. laboratory spectra and real hyperspectral data collected by the AVIRIS sensor over Cuprite, Nevada.

Hyperspectral signal subspace estimation

Given an hyperspectral image, the determination of the number of endmembers and the subspace where they live without any prior knowledge is crucial to the success of hyperspectral image analysis. This paper introduces a new minimum mean squared error based approach to infer the signal subspace in hyperspectral imagery. The method, termed hyperspectral signal identification by minimum error (HySime), is eigendecomposition based and it does not depend on any tuning parameters. It first estimates the signal and noise correlation matrices and then selects the subset of eigenvalues that best represents the signal subspace in the least squared error sense. The effectiveness of the proposed method is illustrated using simulated data based on U.S.G.S. laboratory spectra and real hyperspectral data collected by the AVIRIS sensor over Cuprite, Nevada.

Caracterização da função do monóxido de azoto e glutationo hepáticos na sensabilidade periférica à insulina

xi RESUMO A acção da insulina no músculo esquelético depende de um reflexo parassimpático hepático que conduz à libertação de uma substância hepática sensibilizadora da insulina, designada por HISS, responsável por cerca de 55% do efeito hipoglicemiante da insulina. A acção da HISS é finamente regulada pelo monóxido de azoto (NO) hepático e pelo estado prandial, aumentando no período pós-prandial imediato e diminuindo progressivamente com as horas de jejum. A secreção da HISS pode ser inibida cirúrgica ou farmacologicamente, quer por desnervação selectiva do plexo anterior hepático, quer por administração de atropina, quer por inibição do sintase do NO (NOS) hepático. O objectivo geral do trabalho apresentado nesta dissertação foi a caracterização da via de transdução de sinal que conduz à libertação da HISS. O modelo utilizado neste estudo foi o rato Wistar. A sensibilidade à insulina foi avaliada através do teste rápido de sensibilidade à insulina (RIST). A primeira hipótese de trabalho testada foi que a sequência de eventos que conduzem à secreção da HISS inicia-se com a activação do sistema parassimpático hepático seguida de activação do NOS hepático com subsequente produção de NO e activação do guanilato ciclase (GC). Observou-se que a administração de um dador de NO reverteu a resistência à insulina induzida, quer por inibição do NOS hepático, quer por antagonismo dos receptores muscarínicos com atropina. Em contraste, a resistência à insulina produzida por inibição do NOS hepático não foi revertida por administração intraportal de acetilcolina (ACh). Constatou-se que a inibição do GC hepático diminuiu a sensibilidade à insulina. Estes resultados sugerem que: a ACh libertada no fígado induz a síntese de NO hepático que conduz à libertação da HISS, que por sua vez é modulada pelo GC hepático. A libertação da HISS em resposta à insulina é regulada pelo estado prandial. Uma vez que os níveis hepáticos de glutationo (GSH) se encontram, tal como a HISS, diminuídos no estado de jejum e aumentados após a ingestão de uma refeição, testou-se a hipótese de que o GSH hepático está envolvido na secreção da HISS. Observou-se que a depleção do GSH hepático induziu resistência à insulina, comparável à obtida após inibição do NOS hepático. Estes resultados suportam a hipótese de que o GSH hepático desempenha um papel crítico na acção periférica da insulina. Considerando que, no estado de jejum, tanto os níveis de GSH hepático como os níveis de NO hepático são baixos, testou-se a hipótese de que a co-administração intraportal de um dador de GSH e de um dador de NO promove um aumento da sensibilidade à insulina no estado de jejum, devido ao restabelecimento do mecanismo da HISS. Observou-se que a administração sequencial de dadores de GSH e de NO no fígado provocou um aumento na sensibilidade à insulina, dependente da dose de dador de GSH administrada. Concluiu-se portanto que ambos, GSH e NO, são essenciais para que o mecanismo da HISS esteja completamente funcional. O GSH e o NO reagem para formar um S-nitrosotiol, o S-nitrosoglutationo (GSNO). Os resultados supra-mencionados conduziram à formulação da hipótese de que a secreção/acção da HISS depende da formação de GSNO. Observou-se que a administração intravenosa de S-nitrosotióis (RSNOs) aumentou a sensibilidade à insulina, em animais submetidos a um período de jejum, ao contrário da administração intraportal destes fármacos, o que RSNOs têm uma acção periférica, mas não hepática, na sensibilidade à insulina. Os resultados obtidos conduziram à reformulação da hipótese da HISS, sugerindo que a ingestão de uma refeição activa os nervos parassimpáticos hepáticos levando à libertação de ACh no fígado que, por sua vez activa o NOS. Simultaneamente, ocorre um aumento dos níveis de GSH hepático que reage com o NO hepático para formar um composto nitrosado, o GSNO. Este composto mimetiza a acção hipoglicemiante da HISS no músculo esquelético. SUMMARY Insulin action at the skeletal muscle depends on a hepatic parasympathetic reflex that promotes the release of a hepatic insulin sensitizing substance (HISS) from the liver, which contributes 55% to total insulin action. HISS action is modulated by hepatic nitric oxide (NO) and also by the prandial status so as to, in the immediate ostprandial state, HISS action is maximal, decreasing with the duration of fasting. HISS secretion may be inhibited by interruption of the hepatic parasympathetic reflex, achieved either by surgical denervation of the liver or by cholinergic blockade with atropine, or by prevention of hepatic NO release, using NO synthase (NOS) antagonists. The main objective of this work was to characterize the signal transduction pathways that lead to HISS secretion by the liver. Wistar rats were used and insulin sensitivity was evaluated using the rapid insulin sensitivity test (RIST). The first hypothesis tested was that the sequence of events that lead to HISS secretion starts with an increase in the hepatic parasympathetic tone, followed by the activation of hepatic NOS and subsequent triggering of guanylate cyclase (GC). We observed that insulin resistance produced either by muscarinic receptor antagonism with atropine or by hepatic NOS inhibition was reversed by the intraportal administration of an NO donor. In contrast, intraportal acetylcholine (ACh) did not restore insulin sensitivity after NOS inhibition. We also observed that GC inhibition lead to a decrease in insulin sensitivity.These results suggest that the release of ACh in the liver activates hepatic NO synthesis in order to allow HISS secretion, through a signaling pathway modulated by GC. HISS release in response to insulin is controlled by the prandial status. The second hypothesis tested was that glutathione (GSH) is involved in HISS secretion since the hepatic levels of GSH are, like HISS action, decreased in the fasted state and increased after ingestion of a meal. We observed that hepatic GSH depletion led to insulin resistance of the same magnitude of that observed after inhibition of hepatic NOS. These results support the hypothesis that hepatic GSH is crucial in peripheral insulin action. Since, in the fasted state, both hepatic GSH and NO levels are low, we tested the hypothesis that intraportal o-administration of a GSH donor and an NO donor enhances insulin sensitivity in fasted Wistar rats, by restoring HISS secretion. We observed that GSH and NO increased insulin sensitivity in a GSH dose-dependent manner. We concluded that HISS secretion requires elevated levels of both GSH and NO in the liver. GSH and NO react to form a S-nitrosothiol, S-nitrosoglutathione (GSNO). The last hypothesis tested in this work was that HISS secretion/ action depends on the formation of GSNO. We observed that intravenous administration of -nitrosothiols (RSNOs) increased insulin sensitivity in animals fasted for 24 h, in contrast with the intraportal administration of the drug. This result suggests that RSNOs enhanced insulin sensitivity through a peripheral, and not hepatic, mechanism. The results obtained led to a restructuring of the HISS hypothesis, suggesting that the ingestion of a meal triggers the hepatic parasympathetic nerves, leading to the release of Ach in the liver, which in turn activates NOS. Simultaneously, hepatic GSH levels increase and react with NO to form a nitrosated compound, GSNO. S-nitrosoglutathione mimics HISS hypoglycaemic action at the skeletal muscle.

Parametric analog signal amplification applied to nanoscale cmos wireless digital transceivers

Thesis presented in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the subject of Electrical and Computer Engineering by the Universidade Nova de Lisboa,Faculdade de Ciências e Tecnologia

S100A6 Amyloid Fibril Formation Is Calcium-modulated and Enhances Superoxide Dismutase-1 (SOD1) Aggregation

S100A6 is a small EF-hand calcium- and zinc-binding protein involved in the regulation of cell proliferation and cytoskeletal dynamics. It is overexpressed in neurodegenerative disorders and a proposed marker for Amyotrophic Lateral Sclerosis (ALS). Following recent reports of amyloid formation by S100 proteins, we investigated the aggregation properties of S100A6. Computational analysis using aggregation predictors Waltz and Zyggregator revealed increased propensity within S100A6 helices HI and HIV. Subsequent analysis of Thioflavin-T binding kinetics under acidic conditions elicited a very fast process with no lag phase and extensive formation of aggregates and stacked fibrils as observed by electron microscopy. Ca2+ exerted an inhibitory effect on the aggregation kinetics, which could be reverted upon chelation. An FT-IR investigation of the early conformational changes occurring under these conditions showed that Ca2+ promotes anti-parallel β-sheet conformations that repress fibrillation. At pH 7, Ca2+ rendered the fibril formation kinetics slower: time-resolved imaging showed that fibril formation is highly suppressed, with aggregates forming instead. In the absence of metals an extensive network of fibrils is formed. S100A6 oligomers, but not fibrils, were found to be cytotoxic, decreasing cell viability by up to 40%. This effect was not observed when the aggregates were formed in the presence of Ca2+. Interestingly, native S1006 seeds SOD1 aggregation, shortening its nucleation process. This suggests a cross-talk between these two proteins involved in ALS. Overall, these results put forward novel roles for S100 proteins, whose metal-modulated aggregation propensity may be a key aspect in their physiology and function.

Bacterial inter-species communication mediated by the autoinducer-2 signal

Dissertation presented to obtain the Ph.D degree in Biology by Universidade Nova de Lisboa, Instituto de Tecnologia Química e Biológica, Instituto Gulbenkian de Ciência.

Estudo dosimétrico para implementação da técnica radioterapêutica volumetric modulated arc therapy (VMAT)

Dissertação para obtenção do Grau de Mestre em Engenharia Biomédica

Myoelectric Signal Monitoring System

The Electromyography (EMG) is an important tool for gait analyzes and disorders diagnoses. Traditional methods involve equipment that can disturb the analyses, being gradually substituted by different approaches, like wearable and wireless systems. The cable replacement for autonomous systems demands for technologies capable of meeting the power constraints. This work presents the development of an EMG and kinematic data capture wireless module, designed taking into account power consumption issues. This module captures and converts the analog myoeletric signal to digital, synchronously with the capture of kinetic information. Both data are time multiplexed and sent to a PC via Bluetooth link. The work carried out comprised the development of the hardware, the firmware and a graphical interface running in an external PC. The hardware was developed using the PIC18F14K22, a low power family of microcontrollers. The link was established via Bluetooth, a protocol designed for low power communication. An application was also developed to recover and trace the signal to a Graphic User Interface (GUI), coordinating the message exchange with the firmware. Results were obtained which allowed validating the conceived system in static and with the subject performing short movements. Although it was not possible to perform the tests within more dynamic movements, it is shown that it is possible to capture, transmit and display the captured data as expected. Some suggestions to improve the system performance also were made.

Candida glabrata susceptibility to antifungals and phagocytosis is modulated by acetate

Candida glabrata is considered a major opportunistic fungal pathogen of humans. The capacity of this yeast species to cause infections is dependent on the ability to grow within the human host environment and to assimilate the carbon sources available. Previous studies have suggested that C. albicans can encounter glucose-poor microenvironments during infection and that the ability to use alternative non-fermentable carbon sources, such as carboxylic acids, contributes to the virulence of this fungus. Transcriptional studies on C. glabrata cells identified a similar response, upon nutrient deprivation. In this work, we aimed at analyzing biofilm formation, antifungal drug resistance, and phagocytosis of C. glabrata cells grown in the presence of acetic acid as an alternative carbon source. C. glabrata planktonic cells grown in media containing acetic acid were more susceptible to fluconazole and were better phagocytosed and killed by macrophages than when compared to media lacking acetic acid. Growth in acetic acid also affected the ability of C. glabrata to form biofilms. The genes ADY2a, ADY2b, FPS1, FPS2, and ATO3, encoding putative carboxylate transporters, were upregulated in C. glabrata planktonic and biofilm cells in the presence of acetic acid. Phagocytosis assays with fps1 and ady2a mutant strains suggested a potential role of FPS1 and ADY2a in the phagocytosis process. These results highlight how acidic pH niches, associated with the presence of acetic acid, can impact in the treatment of C. glabrata infections, in particular in vaginal candidiasis.

Analysis of Natural and Artificial Phenomena Using Signal Processing and Fractional Calculus

In this paper we study several natural and man-made complex phenomena in the perspective of dynamical systems. For each class of phenomena, the system outputs are time-series records obtained in identical conditions. The time-series are viewed as manifestations of the system behavior and are processed for analyzing the system dynamics. First, we use the Fourier transform to process the data and we approximate the amplitude spectra by means of power law functions. We interpret the power law parameters as a phenomenological signature of the system dynamics. Second, we adopt the techniques of non-hierarchical clustering and multidimensional scaling to visualize hidden relationships between the complex phenomena. Third, we propose a vector field based analogy to interpret the patterns unveiled by the PL parameters.

Parallel programming in biomedical signal processing

Dissertação para obtenção do Grau de Mestre em Engenharia Biomédica

Algorithms for information extraction and signal annotation on long-term biosignals using clustering techniques

Dissertação para obtenção do Grau de Mestre em Engenharia Biomédica

The role of Rac1-modulated gene transcription in tumorigenesis

Dissertation presented to obtain the Ph.D degree in Biology

Signal transduction pathways involving the hypertension-related WNK1 and WNK4 protein kinases

Dissertação apresentada para obtenção do Grau de Doutor em Biologia, na especialidade de Genética Molecular, pela Universidade Nova de Lisboa, Faculdade de Ciências e Tecnologia