Search

Auditory alarms are commonly badly designed, providing little to no information or guidance. In the healthcare context, the poor acoustics of alarms is one contributor for the noise problem. The goal of this thesis is to propose a human-centred methodology for the design of clinical auditory alarms, by making them less disruptive and more informative, thus improving the healthcare soundscape. It implements this methodology from concept to evaluation and validation, combining psychoacoustics with usability and user experience methods. Another aim of this research consisted in understanding the limitations and possibilities offered by online tools for scientific studies. Thus, different processes and methodologies were implemented, and corresponding results were discussed. To understand the acoustic healthcare environment, field visits, interviews, and surveys were performed with healthcare professionals. Additionally, sound pressure levels and frequency analysis of several surgeries in different hospitals provided specific sound design requirements, which were added to an existent body of knowledge on clinical alarm design. A second stage consisted in prototyping very simple sounds to comprehend which temporal and spectral parameters of sound could be manipulated to communicate clinical information. Parameters such as frequency, speed, onset, and rhythm were studied, and relations between subjective perception and physical parameters were established. In parallel, and heavily influenced by the new IEC 60601-1-8 - General requirements, tests and guidance for alarm systems in medical electrical equipment and medical electrical systems, a design strategy with auditory icons was created. This strategy intended to provide as much information as possible in an auditory alarm. To do so, it involved two main components: a priority pointer indicating the priority of the alarm; an auditory icon indicating the cause of the alarm. A third component indicating increasing or decreasing tendency of the vital sign was designed, but not validated with users. After online validation of the priority pointer and auditory icon for eight categories (cardiac, drug administration, ventilation, blood pressure, perfusion, oxygen, temperature, and power down), a new library of clinical auditory alarms is proposed. Os alarmes auditivos são habitualmente mal concebidos, dando poucas informações ou orientações perante a situação que despoletou o aviso. No contexto da saúde, a má acústica dos alarmes é um dos contribuidores para o problema do ruído. O objetivo desta tese é o de melhorar a paisagem sonora em ambientes clínicos, propondo uma metodologia centrada no Humano para o design de alarmes auditivos clínicos, tornando-os menos disruptivos e mais informativos. Essa metodologia é implementada desde o conceito até a avaliação e validação, combinando métodos da psicoacústica com métodos de usabilidade e experiência do utilizador. Outro objetivo desta investigação é o de compreender as limitações e possibilidades oferecidas pelas ferramentas online para estudos científicos. Assim, diversos processos e metodologias foram implementados, e os respetivos resultados são discutidos. Para compreender o ambiente acústico clínico, foram realizadas visitas de campo, entrevistas e inquéritos com profissionais de saúde. Além disso, avaliou-se o nível de pressão sonora e frequências de várias cirurgias em diferentes hospitais. Esta atividade forneceu requisitos específicos de design de som que foram adicionados a um corpo existente de conhecimento sobre design de alarmes clínicos. Uma segunda etapa consistiu na prototipagem de sons simples para compreender que parâmetros temporais e espectrais do som poderiam ser manipulados para comunicar informações clínicas. Parâmetros como frequência, velocidade, envelope e ritmo foram estudados, e as relações entre a perceção subjetiva e os parâmetros físicos foram estabelecidas. Paralelamente, e fortemente influenciado pela nova norma IEC 60601-1-8 - Requisitos gerais, testes e orientações para sistemas de alarme em equipamentos médicos elétricos e sistemas médicos elétricos, foi criada uma estratégia de design com ícones auditivos. Essa estratégia pretendia incorporar o máximo de informações num alarme auditivo. Para isso, envolveu dois componentes principais: um ponteiro de prioridade que indica a prioridade do alarme; e um ícone auditivo que indica a causa do alarme. Um terceiro componente de tendência (aumento ou diminuição do valor do sinal vital) foi criado, mas não validado com utilizadores. Após a validação do ponteiro de prioridade e ícone auditivo para oito categorias (cardíaco, administração de medicamentos, ventilação, pressão arterial, perfusão, oxigénio, temperatura e falha de equipamento), propõe-se uma nova biblioteca de alarmes auditivos clínicos.

Due to the increasing technological innovation over the last decades, the average life expectancy of a human being has been increasing exponentially. Although this is an excellent step forward for humanity, it has led older population to being more prone to illness, making them more vulnerable to accidents such as falls. In this article a study is made on the existing literature in non-intrusive remote health monitoring systems, towards the design and implementation of an IoT system capable of identifying fall situations and monitor cardiac data. A Systematic Literature Review (SLR) method was considered in this work, focused on reviewing the existing literature on remote health monitoring systems, having fall detection algorithms, based in IoT. The Design Science Research (DSR) methodology was used to seek to enhance technology and science knowledge about this paper's topic, through the creation of an innovative artifact.The system includes a smart watch (Lily-Go T-Watch-2020 V2), programmable in C under Arduino IDE to detect falls and a photoplethysmography monitoring unit (PPG) based on a Onyx 9560 Bluetooth oximeter, capable of measuring the user's blood oxygen percentage (SpO2) and heart rate, in real time. It also provides remote monitoring through a user-friendly website to visualize live data about the status of the user. The system was tested in volunteers to show the effectiveness of remote health monitoring systems for the elderly population. info:eu-repo/semantics/acceptedVersion

Submitted by Margarida Vilhena Carvalho (mvc@ucp.pt) on 2021-12-23T12:27:42Z No. of bitstreams: 1 202830772.pdf: 3512407 bytes, checksum: 7e2132fe5a1219a6998d6bc532647cf0 (MD5) Made available in DSpace on 2021-12-23T12:27:42Z (GMT). No. of bitstreams: 1 202830772.pdf: 3512407 bytes, checksum: 7e2132fe5a1219a6998d6bc532647cf0 (MD5) Previous issue date: 2021-11-03

Submitted by Ana Moreira (anadsmoreira@ufp.pt) on 2021-09-27T11:16:51Z No. of bitstreams: 1 DM_37718.pdf: 4298392 bytes, checksum: ad8d7e3d55ec8198a56484159c61c721 (MD5) Approved for entry into archive by azevedo@ufp.pt (azevedo@ufp.pt) on 2021-09-27T12:18:12Z (GMT) No. of bitstreams: 1 DM_37718.pdf: 4298392 bytes, checksum: ad8d7e3d55ec8198a56484159c61c721 (MD5) Made available in DSpace on 2021-09-27T12:18:13Z (GMT). No. of bitstreams: 1 DM_37718.pdf: 4298392 bytes, checksum: ad8d7e3d55ec8198a56484159c61c721 (MD5) Previous issue date: 2021-07-27

Submitted by Ana Moreira (anadsmoreira@ufp.pt) on 2021-09-08T15:07:06Z No. of bitstreams: 1 DM_37717.pdf: 1381917 bytes, checksum: 228c08028dd91b97e46216f64a9315ff (MD5) Approved for entry into archive by azevedo@ufp.pt (azevedo@ufp.pt) on 2021-09-13T09:18:21Z (GMT) No. of bitstreams: 1 DM_37717.pdf: 1381917 bytes, checksum: 228c08028dd91b97e46216f64a9315ff (MD5) Made available in DSpace on 2021-09-13T09:18:21Z (GMT). No. of bitstreams: 1 DM_37717.pdf: 1381917 bytes, checksum: 228c08028dd91b97e46216f64a9315ff (MD5) Previous issue date: 2021-07-07

Ser saudável é muito mais do que não estar doente. É essencialmente viver o dia-a-dia sob um bem-estar físico e mental. Uma sociedade moderna, cada vez mais agitada e exigente, dificulta esse objectivo. Presentemente, grande parte da população sofre de depressão, ansiedade e/ou stress. Este problema de saúde pública, agravado recentemente pelo COVID-19, carece de uma solução urgente. O efeito benéfico da prática continuada de meditação Mindfulness poderá contribuir para uma melhor qualidade de vida do indivíduo, vivendo o presente em atenção plena. Recorrendo a uma tarefa cognitiva, a um desafio motor e a um estímulo visual, são avaliados 25 voluntários (média idades = 26 anos, DP = 7, 9 masculinos), num período de 4 meses, no qual, durante 8 semanas, frequentaram o curso Mindfulnes-Based Stress Reduction (MBSR). Nas 4 sessões programadas (Pré, Peri, Pós-MBSR e follow-up 2 meses depois) são recolhidos, simultaneamente, três sinais electrofisiológicos: electroencefalograma (EEG), electrocardiograma (ECG) e actividade electrodérmica (EDA). Paralelamente, são preenchidos três questionários: qualidade de vida (WHOQOL), perfil de estado de humor (POMS) e escala de ansiedade, depressão e stress (DASS). Na análise DASS ao total de participantes, o estado de ansiedade diminui -66.0% (p

Submitted by António Freitas (amsf@uma.pt) on 2021-10-14T13:43:47Z No. of bitstreams: 1 Master Thesis - Sistema Realidade Virtual em Oncologia Pediá.pdf: 36385756 bytes, checksum: 127f3b661010d4df147b5232650be77e (MD5) Made available in DSpace on 2021-10-14T13:43:47Z (GMT). No. of bitstreams: 1 Master Thesis - Sistema Realidade Virtual em Oncologia Pediá.pdf: 36385756 bytes, checksum: 127f3b661010d4df147b5232650be77e (MD5) Previous issue date: 2021-06-20

Loss of mobility is among the most impactful consequences of sustaining a spinal cord injury. Wheelchairs provide a considerable degree of mobility to their users, but are not without their drawbacks, most of which are caused by requiring prolonged periods of time in the sitting position. Recently, the field of lower limb exoskeletons has seen considerable developments, and the use of external power technologies has made it possible for users to walk for longer periods of time. However, current exoskeletons do not ensure balance during standing and walking conditions, which leaves their users vulnerable to situations of instability and falls. The goal of this thesis was to investigate safe fall strategies to reduce the severity of the impact in case of a loss of balance with a lower limb exoskeleton. The backwards fall scenario is examined, and a fall strategy is implemented in a simulation environment, using a combination of center of mass and hip joint angle reference signals. The results verified the model’s ability to execute the proposed strategy by following the reference signals, and the strategy was shown to result in safer falls. Further work should be conducted to test this strategy in real-life human-exoskeleton fall scenarios, and to develop strategies for other fall scenarios. Perda de mobilidade é das consequências com maior impacto na vida de quem sofre uma lesão da medula espinhal. Cadeiras de rodas providenciam um nível considerável de mobilidade aos seus utilizadores, mas trazem os seus próprios problemas, sendo que a maioria dos quais são resultantes da necessidade dos seus utilizadores permanecerem sentados durante longos períodos de tempo. Mais recentemente, a área dos exosqueletos para membros inferiores tem visto desenvolvimento notável, e o uso de fontes de energia externas tem permitido que os seus utilizadores se desloquem a pé durante intervalos de tempo maiores. Contudo, estas tecnologias ainda não garantem equilíbrio constante dos seus utilizadores durante a marcha ou em pé, o que faz com que estes estejam vulneráveis a situações de instabilidade e quedas. Esta tese teve como objetivo investigar estratégias de queda segura para reduzir a severidade do impacto em caso de uma perda de equilíbrio utilizando um exosqueleto para membros inferiores. É examinado o cenário de uma queda para trás, e é implementada uma estratégia de queda num ambiente de simulação, recorrendo a uma combinação de sinais de referência de centro de massa e ângulo da articulação da anca. Os resultados verificaram a capacidade do modelo utilizado executar a estratégia proposta ao seguir os sinais de referência, e a estratégia mostrou resultar em quedas mais seguras. No futuro, a estratégia deve ser testada em cenários reais de quedas de um exosqueleto e o seu utilizador, e devem ser desenvolvidas estratégias para os restantes cenários de queda.

Cardiovascular diseases are the leading global death cause. Their treatment and prevention rely on electrocardiogram interpretation, which is dependent on the physician’s variability. Subjectiveness is intrinsic to electrocardiogram interpretation and hence, prone to errors. To assist physicians in making precise and thoughtful decisions, artificial intelligence is being deployed to develop models that can interpret extent datasets and provide accurate decisions. However, the lack of interpretability of most machine learning models stands as one of the drawbacks of their deployment, particularly in the medical domain. Furthermore, most of the currently deployed explainable artificial intelligence methods assume independence between features, which means temporal independence when dealing with time series. The inherent characteristic of time series cannot be ignored as it carries importance for the human decision making process. This dissertation focuses on the explanation of heartbeat classification using several adaptations of state-of-the-art model-agnostic methods, to locally explain time series classification. To address the explanation of time series classifiers, a preliminary conceptual framework is proposed, and the use of the derivative is suggested as a complement to add temporal dependency between samples. The results were validated on an extent public dataset, through the 1-D Jaccard’s index, which consists of the comparison of the subsequences extracted from an interpretable model and the explanation methods used. Secondly, through the performance’s decrease, to evaluate whether the explanation fits the model’s behaviour. To assess models with distinct internal logic, the validation was conducted on a more transparent model and more opaque one in both binary and multiclass situation. The results show the promising use of including the signal’s derivative to introduce temporal dependency between samples in the explanations, for models with simpler internal logic. As doenças cardiovasculares são, a nível mundial, a principal causa de morte e o seu tratamento e prevenção baseiam-se na interpretação do electrocardiograma. A interpretação do electrocardiograma, feita por médicos, é intrinsecamente subjectiva e, portanto, sujeita a erros. De modo a apoiar a decisão dos médicos, a inteligência artificial está a ser usada para desenvolver modelos com a capacidade de interpretar extensos conjuntos de dados e fornecer decisões precisas. No entanto, a falta de interpretabilidade da maioria dos modelos de aprendizagem automática é uma das desvantagens do recurso à mesma, principalmente em contexto clínico. Adicionalmente, a maioria dos métodos inteligência artifical explicável assumem independência entre amostras, o que implica a assunção de independência temporal ao lidar com séries temporais. A característica inerente das séries temporais não pode ser ignorada, uma vez que apresenta importância para o processo de tomada de decisão humana. Esta dissertação baseia-se em inteligência artificial explicável para tornar inteligível a classificação de batimentos cardíacos, através da utilização de várias adaptações de métodos agnósticos do estado-da-arte. Para abordar a explicação dos classificadores de séries temporais, propõe-se uma taxonomia preliminar, e o uso da derivada como um complemento para adicionar dependência temporal entre as amostras. Os resultados foram validados para um conjunto extenso de dados públicos, por meio do índice de Jaccard em 1-D, com a comparação das subsequências extraídas de um modelo interpretável e os métodos inteligência artificial explicável utilizados, e a análise de qualidade, para avaliar se a explicação se adequa ao comportamento do modelo. De modo a avaliar modelos com lógicas internas distintas, a validação foi realizada usando, por um lado, um modelo mais transparente e, por outro, um mais opaco, tanto numa situação de classificação binária como numa situação de classificação multiclasse. Os resultados mostram o uso promissor da inclusão da derivada do sinal para introduzir dependência temporal entre as amostras nas explicações fornecidas, para modelos com lógica interna mais simples.

Medical Microwave Imaging (MMI) has been studied in the past years to develop techniques to detect breast cancer at the earliest stages of development. Particularly, ultra-wideband (UWB) micro-wave radar imaging systems can detect and classify tumors as benign or malignant since this technique yields information about the size and shape of tumors. In this study we used this technology to classify tumors. The primary goal of this dissertation is two-folded. First, producing breast tumor numerical mod-els and using them in 2D MMI simulations that recreate the conditions of a UWB microwave radar imaging system. The breast tumor numerical produced resemble real tumor morphologies since they are made from breast MRI exams segmentations. Second, the data of the backscattered UWB microwave signals produced by the MMI simulations was used to classify tumors according to their size and histol-ogy, which is relevant to assess potential of UWB microwave radar imaging systems as a reliable alter-native method for the classification of breast tumors in the field of Medical Microwave Imaging. The Classification Algorithms used in this work were Pseudo Linear Discriminant Analysis (Pseudo-LDA), Pseudo Quadratic Discriminant Analysis (pseudo-QDA), and k-Nearest Neighbors (KNN), alongside with a feature extraction algorithm – Principal Component Analysis (PCA). A Imagem Médica por Microondas (do inglês, MMI) tem sido estudada nos últimos anos de forma a desenvolver técnicas de deteção do cancro da mama nas primeiras fases de desenvolvimento. Em particular, os sistemas de imagem de radar por microondas em banda ultralarga (do inglês UWB) podem detetar e classificar os tumores como benignos ou malignos, uma vez que esta técnica produz informação sobre o tamanho e a forma dos tumores. Neste estudo, utilizámos esta tecnologia para classificar os tumores. A dissertação tem dois objetivos principais. Primeiro, produzir fantomas de tumores mamários e utilizá-los em simulações de MMI em 2D que recriam as condições de um sistema de imagem de radar por microondas UWB. Os fantomas numéricos de tumores mamários produzidos possuem morfologias semelhantes a tumores reais, uma vez que são feitos a partir de segmentações de exames de ressonância magnética da mama. Em segundo lugar, as reflexões dos sinais de microondas UWB produzidos pelas simulações de MMI foram utilizados para classificar tumores de acordo com o seu tamanho e histologia, o que é relevante para avaliar o potencial dos sistemas de imagem de radar por microondas UWB como um método alternativo e fiável para a classificação de tumores mamários no campo da MMI. Os Algo-ritmos de Classificação utilizados neste trabalho foram a Pseudo Linear Discriminant Analysis (Pseudo-LDA), Pseudo Quadratic Discriminant Analysis (pseudo-QDA), e a K-Nearest Neighbors (KNN), jun-tamente com um algoritmo de extração de features - Análise de Componentes Principais (do inglês PCA).

Submitted by azevedo@ufp.pt (azevedo@ufp.pt) on 2021-03-11T17:31:47Z No. of bitstreams: 1 DM_37707.pdf: 2046326 bytes, checksum: 30e390d32ab8a7876e42dd1b53d8f7a5 (MD5) Approved for entry into archive by azevedo@ufp.pt (azevedo@ufp.pt) on 2021-03-11T18:39:51Z (GMT) No. of bitstreams: 1 DM_37707.pdf: 2046326 bytes, checksum: 30e390d32ab8a7876e42dd1b53d8f7a5 (MD5) Made available in DSpace on 2021-03-11T18:39:51Z (GMT). No. of bitstreams: 1 DM_37707.pdf: 2046326 bytes, checksum: 30e390d32ab8a7876e42dd1b53d8f7a5 (MD5) Previous issue date: 2021-02-22

The healthcare system, which is ubiquitously recognized as one of the most influential system in society, is facing new challenges since the start of the decade.The myriad of physiological data generated by individuals, namely in the healthcare system, is generating a burden on physicians, losing effectiveness on the collection of patient data. Information systems and, in particular, novel deep learning (DL) algorithms have been prompting a way to take this problem. This thesis has the aim to have an impact in biosignal research and industry by presenting DL solutions that could empower this field. For this purpose an extensive study of how to incorporate and implement Convolutional Neural Networks (CNN), Recursive Neural Networks (RNN) and Fully Connected Networks in biosignal studies is discussed. Different architecture configurations were explored for signal processing and decision making and were implemented in three different scenarios: (1) Biosignal learning and synthesis; (2) Electrocardiogram (ECG) biometric systems, and; (3) Electrocardiogram (ECG) anomaly detection systems. In (1) a RNN-based architecture was able to replicate autonomously three types of biosignals with a high degree of confidence. As for (2) three CNN-based architectures, and a RNN-based architecture (same used in (1)) were used for both biometric identification, reaching values above 90% for electrode-base datasets (Fantasia, ECG-ID and MIT-BIH) and 75% for off-person dataset (CYBHi), and biometric authentication, achieving Equal Error Rates (EER) of near 0% for Fantasia and MIT-BIH and bellow 4% for CYBHi. As for (3) the abstraction of healthy clean the ECG signal and detection of its deviation was made and tested in two different scenarios: presence of noise using autoencoder and fully-connected network (reaching 99% accuracy for binary classification and 71% for multi-class), and; arrhythmia events by including a RNN to the previous architecture (57% accuracy and 61% sensitivity). In sum, these systems are shown to be capable of producing novel results. The incorporation of several AI systems into one could provide to be the next generation of preventive medicine, as the machines have access to different physiological and anatomical states, it could produce more informed solutions for the issues that one may face in the future increasing the performance of autonomous preventing systems that could be used in every-day life in remote places where the access to medicine is limited. These systems will also help the study of the signal behaviour and how they are made in real life context as explainable AI could trigger this perception and link the inner states of a network with the biological traits. O sistema de saúde, que é ubiquamente reconhecido como um dos sistemas mais influentes da sociedade, enfrenta novos desafios desde o ínicio da década. A miríade de dados fisiológicos gerados por indíviduos, nomeadamente no sistema de saúde, está a gerar um fardo para os médicos, perdendo a eficiência no conjunto dos dados do paciente. Os sistemas de informação e, mais espcificamente, da inovação de algoritmos de aprendizagem profunda (DL) têm sido usados na procura de uma solução para este problema. Esta tese tem o objetivo de ter um impacto na pesquisa e na indústria de biosinais, apresentando soluções de DL que poderiam melhorar esta área de investigação. Para esse fim, é discutido um extenso estudo de como incorporar e implementar redes neurais convolucionais (CNN), redes neurais recursivas (RNN) e redes totalmente conectadas para o estudo de biosinais. Diferentes arquiteturas foram exploradas para processamento e tomada de decisão de sinais e foram implementadas em três cenários diferentes: (1) Aprendizagem e síntese de biosinais; (2) sistemas biométricos com o uso de eletrocardiograma (ECG), e; (3) Sistema de detecção de anomalias no ECG. Em (1) uma arquitetura baseada na RNN foi capaz de replicar autonomamente três tipos de sinais biológicos com um alto grau de confiança. Quanto a (2) três arquiteturas baseadas em CNN e uma arquitetura baseada em RNN (a mesma usada em (1)) foram usadas para ambas as identificações, atingindo valores acima de 90 % para conjuntos de dados à base de eletrodos (Fantasia, ECG-ID e MIT -BIH) e 75 % para o conjunto de dados fora da pessoa (CYBHi) e autenticação, atingindo taxas de erro iguais (EER) de quase 0 % para Fantasia e MIT-BIH e abaixo de 4 % para CYBHi. Quanto a (3) a abstração de sinais limpos e assimptomáticos de ECG e a detecção do seu desvio foram feitas e testadas em dois cenários diferentes: na presença de ruído usando um autocodificador e uma rede totalmente conectada (atingindo 99 % de precisão na classificação binária e 71 % na multi-classe), e; eventos de arritmia incluindo um RNN na arquitetura anterior (57 % de precisão e 61 % de sensibilidade). Em suma, esses sistemas são mais uma vez demonstrados como capazes de produzir resultados inovadores. A incorporação de vários sistemas de inteligência artificial em um unico sistema pederá desencadear a próxima geração de medicina preventiva. Os algoritmos ao terem acesso a diferentes estados fisiológicos e anatómicos, podem produzir soluções mais informadas para os problemas que se possam enfrentar no futuro, aumentando o desempenho de sistemas autónomos de prevenção que poderiam ser usados na vida quotidiana, nomeadamente em locais remotos onde o acesso à medicinas é limitado. Estes sistemas também ajudarão o estudo do comportamento do sinal e como eles são feitos no contexto da vida real, pois a IA explicável pode desencadear essa percepção e vincular os estados internos de uma rede às características biológicas.

Submitted by azevedo@ufp.pt (azevedo@ufp.pt) on 2021-03-11T16:00:42Z No. of bitstreams: 1 DM_37701.pdf: 2304714 bytes, checksum: 9c85e8d7a8b7414bd99f0a31d1506fe9 (MD5) Approved for entry into archive by azevedo@ufp.pt (azevedo@ufp.pt) on 2021-03-11T17:29:35Z (GMT) No. of bitstreams: 1 DM_37701.pdf: 2304714 bytes, checksum: 9c85e8d7a8b7414bd99f0a31d1506fe9 (MD5) Made available in DSpace on 2021-03-11T17:29:36Z (GMT). No. of bitstreams: 1 DM_37701.pdf: 2304714 bytes, checksum: 9c85e8d7a8b7414bd99f0a31d1506fe9 (MD5) Previous issue date: 2021-02-08

Retina image analysis is an important screening tool for early detection of multiple dis eases such as diabetic retinopathy which greatly impairs visual function. Image analy sis and pathology detection can be accomplished both by ophthalmologists and by the use of computer-aided diagnosis systems. Advancements in hardware technology led to more portable and less expensive imaging devices for medical image acquisition. This promotes large scale remote diagnosis by clinicians as well as the implementation of computer-aided diagnosis systems for local routine disease screening. However, lower cost equipment generally results in inferior quality images. This may jeopardize the reliability of the acquired images and thus hinder the overall performance of the diagnos tic tool. To solve this open challenge, we carried out an in-depth study on using different deep learning-based frameworks for improving retina image quality while maintaining the underlying morphological information for the diagnosis. Our results demonstrate that using a Cycle Generative Adversarial Network for unpaired image-to-image trans lation leads to successful transformations of retina images from a low- to a high-quality domain. The visual evidence of this improvement was quantitatively affirmed by the two proposed validation methods. The first used a retina image quality classifier to confirm a significant prediction label shift towards a quality enhance. On average, a 50% increase of images being classified as high-quality was verified. The second analysed the perfor mance modifications of a diabetic retinopathy detection algorithm upon being trained with the quality-improved images. The latter led to strong evidence that the proposed solution satisfies the requirement of maintaining the images’ original information for diagnosis, and that it assures a pathology-assessment more sensitive to the presence of pathological signs. These experimental results confirm the potential effectiveness of our solution in improving retina image quality for diagnosis. Along with the addressed con tributions, we analysed how the construction of the data sets representing the low-quality domain impacts the quality translation efficiency. Our findings suggest that by tackling the problem more selectively, that is, constructing data sets more homogeneous in terms of their image defects, we can obtain more accentuated quality transformations.

Surgical patients are at risk of experiencing clinical deterioration events, especially when transferred to general wards during the postoperative period of their hospital stay. Cur rently, such events are detected by combining Early Warning Scores (EWS) with manual and periodical vital signs measurements, performed by nurses every 4 to 6 hours. Hence, deterioration may remain unnoticed for hours, delaying patient treatment, which might lead to increased morbidity and mortality. Also, EWS are inadequate to predict events so physiologically complex. So that early warning of deterioration could be provided, it was investigated the potential of warning systems that combine machine learning-based prediction models with continuous vital signs monitoring, provided by wearable sensors. This dissertation presents the development of such a warning system, fully indepen dent of manual measurements and based on a logistic regression prediction model with 85% sensitivity, 79% precision and 98% specificity. Additionally, a new personalized ap proach to handle missing data periods in vital signs and a novel variation of a RR-interval preprocessing technique were developed. The results obtained revealed a relevant im provement in the detection of deterioration events and a significant reduction in false alarms, when comparing the warning system with a commonly employed EWS (42% sensitivity, 14% precision and 90% specificity). It was also found that the developed sys tem can assess patient’s condition much more frequently and with timely deterioration detection, without even requiring nurses to interrupt their workflow. These findings sup port the idea that these warning systems are reliable, more practical, more appropriate and produce smarter alarms than current methods, making early deterioration detection possible, thus contributing for better patients outcomes. Nonetheless, the performance achieved may yet reveal insufficient for application in real clinical contexts. Therefore, further work is necessary to improve prediction performance to a greater extent and to confirm these systems reliability.

Although nuclear medicine (NM) procedures are highly effective diagnostic tools, they have been contributing significantly, together with other medical diagnostic and therapeutic methodologies, to the increase in exposure to ionizing radiation in recent years. There is an urgent need to opti mize NM techniques, to maintain diagnostic quality at a minimum possible radiation absorbed dose. 68Ga-prostate-specific membrane antigen positron emission tomography/computed tomog raphy (68Ga-PSMA-11 PET/CT) imaging has rapidly gained notoriety in the NM field and, at the same time, personalized dosimetry studies using voxel-based methods have been performed. This study aimed to calculate the absorbed dose at the voxel level in the kidneys, liver, spleen, and red bone marrow, compare the results with other studies and draw conclusions regarding the safety of using 68Ga-PSMA-11 in NM clinics. Whole-body PET/CT images from six patients were acquired after a single 68Ga-PSMA 11 injection. After registration of the CT and PET images, the target organs were manually seg mented in the CT and resampled to the PET voxel size. Voxel S-values were computed for spe cific tissues using the Monte-Carlo N-Particle transport 6.1 code. The absorbed dose rates were obtained by convolution of the PET activity images with the specific S-values of each tissue. A time integral was then applied to each distribution to account for all 68Ga decay. Statistical dose values were computed and compared with the available literature. Considering all the patients included in this study, the kidneys received the highest radi ation, with a mean overall absorbed dose of 0.0561 mGy/MBq and a median overall absorbed dose of 0.0499 mGy/MBq. In contrast, the red bone marrow received the lowest absorbed dose values (mean dose: 0.0015 mGy/MBq, median dose: 0.0013 mGy/MBq). The present study showed lower dosimetry values than the literature, resulting in deviations ranging from -38.1% (in the liver) to -91.3% (in the red bone marrow). The present study employs a voxel-based approach, which considers a non-uniform bio distribution of the radiopharmaceutical in the organs and leads to dosimetry estimates closer to the real ones. The reasonable low absorbed doses in the four organs herein studied is an argument in favor of using 68Ga-PSMA-11 in NM clinics. In the Future Work chapter, a more specific dynamic NM imaging methodology, taking into consideration the radiopharmaceutical pharma cokinetics, is presented.

Motivation: Cognitive impairment is an important symptom of Parkinson’s Disease (PD), usually having a substantial negative impact on the quality of life of patients, families, and caregivers. Cognitive Training (CT) have been proven effective in halting the process of cognitive decline in PD. However, the efficacy of CT is unpredictable from subject to subject. Objective: Investigate the possibility of predicting the outcome of CT in PD patients with Mild Cognitive Impairment using structural and functional Magnetic Resonance Imaging (MRI) data. Methods: Before CT, a sample of 42 PD patients underwent structural and functional MRI. Graph measures were then extracted from their structural and functional con nectomes and used as features for random forest (RFo) and decision tree (DT) machine learning (ML) regression algorithms with and without prior latent component analysis (LCA). CT response was evaluated by assessing the outcomes of the Tower of London task pre- and post-treatment. Finally, the 4 ML models were used to predict CT response and their performances were assessed. Post hoc analyses were conducted to investigate whether these algorithms could predict age using connectomic measures on a sample of 80 PD patients. Results: The performances of the aforementioned algorithms did not differ signifi cantly from the baseline performance predicting the subject-specific CT outcome. The performance of the RFo without LCA differed significantly from the baseline performance in the age prediction task for the sample of 80 patients. Conclusion: Notwithstanding the lack of statistical significance in predicting our xicognitive outcomes, the relative success of the age prediction task points towards the potential of this approach. We hypothesise that bigger sample sizes are needed in order to predict the outcome of CT using ML.

A doença de Parkinson é uma doença neurodegenerativa comum no mundo, em que um dos seus tratamentos pode passar pela cirurgia de estimulação cerebral profunda. Esta te rapêutica produz modificações na qualidade de vida dos doentes, com custos económicos associados. Ao contrário de outros países, em Portugal o estudo do impacto da cirurgia de Parkinson na qualidade de vida e carga sócio-económica ainda não foi publicado. Este tipo de estudo é importante para o Serviço Nacional de Saúde porque permite perceber o benefício clínico e económico da cirurgia. Surge, assim, a necessidade de abordar este tema neste país. Com o objetivo de avaliar a alteração da qualidade de vida dos doentes sujeitos à cirurgia e os custos relacionados com a doença de Parkinson, foram aplicados questionários de qualidade de vida antes e depois da cirurgia, nomeadamente UPDRS MDS, a doentes seguidos no Hospital de Santa Maria (CHULN). Os dados recolhidos mostraram que a situação da doença de Parkinson em Portugal é semelhante à de outros países referidos na literatura científica. O estudo estatístico permitiu concluir que há diferenças significativas na qualidade de vida dos doentes quando comparado o antes e o depois da intervenção cirúrgica. O estudo económico avaliou a custo-efetividade da inter venção cirúrgica numa perspetiva da sociedade que agrega as perspetivas do doente e do Serviço Nacional de Saúde. Pode-se concluir que a cirurgia melhorou significativamente a qualidade de vida, em termos clínicos, e que o investimento feito neste tratamento é recompensado a longo prazo.

O Sistema Nervoso Autónomo (SNA) controla a atividade inconsciente do organismo e regula-a de modo a garantir a homeostasia. Diversas doenças manifestam-se em consequência de um mau funcionamento do SNA, como a diabetes.Por isso, compreender o SNA é essencial para melhorar o diagnóstico e tratamento destas doenças.Umas das técnicas de medição do SNA é a pupilometria. Assim, desenvolveu-se um pupilómetro de vídeo de baixo custo que se pretende utilizar e melhorar nesta dissertação de modo aperceber se é possível observar-se alterações do SNA, em indivíduos diabéticos.Para tal,recolheram-se sinais de pupilometria, de eletrocardiografia e atividade eletrodérmica, que foram analisados por métodos lineares nos domínios dotempo e da frequência.Concluiu-se que é necessário realizar um estudo mais alargado e aprofundado para que se possa confirmar a capacidade do pupilómetropara estas aplicações, embora os resultados indiquem uma tendência para que tal seja possível. The Autonomic Nervous System (ANS) controls inconscient activity of the organism and regulates it to ensure homeostasis. Many illnesses manifest inconsequence of the malfunction of the ANS, such as diabetes. For that reason, understanding the ANS is essential to improve the diagnosis and treatment of these diseases. One technique to measure ANS is pupillometry. Thus, a low-cost video pupillometer was developed,and this dissertation intends to use and improve it, to understand if it can be used to observe ANS alterations, in diabetic individuals.For that, pupillometry, electrocardiographic and electrodermal activity signals were collected and analysed through linear methods in time and frequency domains.It was concluded that itis necessary to perform a broader and profound study to confirm the pupillometer’s capacity for these applications, although the results indicate that it is likely to be possible.

O estilo de vida actual é promotor de estados de depressão, ansiedade ou stress, cujo tratamento passa maioritariamente pela prescrição de fármacos psicoactivos. Este panorama tem preocupado a área da Medicina, que tem procurado respostas alternativas. Uma delas é a prática de mindfulness, cujos benefícios têm sido amplamente testemunhados pelos seus praticantes. No entanto, continua sem se conhecer exactamente a sua influência na dinâmica do organismo, nomeadamente na actividade cerebral, pelo que o estudo de eventuais alterações tem merecido crescente atenção da área da Neurociência. Com este trabalho pretende-se estudar os efeitos da prática de mindfulness na conectividade cerebral, através de uma abordagem objectiva e longitudinal. Para isso, foram utilizados os dados de Electroencefalografia (EEG) recolhidos em 13 indivíduos, durante um trabalho de investigação anterior. Os dados foram adquiridos em 4 sessões distintas: antes, durante, no fim e após a realização de um curso de mindfulness. Em cada sessões, os indivíduos realizaram a mesma tarefa, alternando entre o estado de repouso e de concentração. Os registos de EEG foram utilizados para analisar a evolução da conectividade cerebral entre as diferentes regiões do cérebro, ao longo das sessões. Nesta análise foram consideradas as banda de frequências Teta, Alfa e Beta. Os resultados obtidos para as sessões 1 e 2 são idênticos e evidenciam a latência dos efeitos da prática de mindfulness. Através dos resultados obtidos, conclui-se a existência de uma reorganização na dinâmica funcional do cérebro, sendo observado um padrão de conectividade mais focalizado ao longo do tempo, mais evidente a partir da 3ª sessão, permanecendo após o fim do curso, pois entre a sessão 3 e a 4 (separadas por 2 meses) os padrões de conectividade apresentam semelhanças. Tal observação sustenta a hipótese da promoção da neuroplasticidade através prática de mindfulness. Modern society’s lifestyle is a promoter of depression, anxiety and stress, which leads to the rise of psychoactive drugs consumption. This scenario constitutes a serious public health concern. To decrease this consumption, the field of medicine has been looking for alternative answers. Mindfulness practice is one of the alternative ways, its benefits have been widely known by its practitioners. However, its exact influence on the human organism remains unknown, namely about brain activity. This is the reason for the increasing attention from neuroscientists about mindfulness neurological effects. This work aims to study the effects of mindfulness practice on brain connectivity through an objective and longitudinal approach. For this study, Electroencephalography (EEG) records collected from 13 subjects during a previous investigation work were used. The data were acquired in 4 different sessions: before, during, at the end and two months after the conclusion of a mindfulness course. During each session, each subject performed a task alternating between the resting state and a cognitive process. EEG records were used to assess the brain connectivity evolution of different brain’s regions. This analysis was performed for the frequency bands Theta, Alpha and Beta. The results obtained between sessions 1 and 2 do not differ somuch, which proves a latency of the mindfulness effects. The results obtained indicate a reorganization in the brain’s functional dynamics, observing a connectivity pattern more localized over time. This is expressly evident from the 3rd session. The results show that after the end of the course those effects remain for some time, since the results for sessions 3 and 4 reveal similarities. This fact supports the hypothesis for mindfulness-induced neural plasticity.

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Radiações em Diagnóstico e Terapia), Universidade de Lisboa, Faculdade de Ciências, 2021 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-06-02T16:53:32Z No. of bitstreams: 1 ulfc126367_tm_Joana_Leitão.pdf: 34179793 bytes, checksum: 57820f89feedef3d66a051f9598997cd (MD5) Made available in DSpace on 2021-06-02T16:53:41Z (GMT). No. of bitstreams: 1 ulfc126367_tm_Joana_Leitão.pdf: 34179793 bytes, checksum: 57820f89feedef3d66a051f9598997cd (MD5) Previous issue date: 2021

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Engenharia Clínica e Instrumentação Médica), Universidade de Lisboa, Faculdade de Ciências, 2021 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-05-05T18:08:20Z No. of bitstreams: 1 ulfc126204_tm_João_Lopes.pdf: 3048580 bytes, checksum: 7e5065ecbb660b40f37abebd03c97b37 (MD5) Made available in DSpace on 2021-05-05T18:08:30Z (GMT). No. of bitstreams: 1 ulfc126204_tm_João_Lopes.pdf: 3048580 bytes, checksum: 7e5065ecbb660b40f37abebd03c97b37 (MD5) Previous issue date: 2021

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Sinais e Imagens Médicas), Universidade de Lisboa, Faculdade de Ciências, 2021 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-06-04T16:24:15Z No. of bitstreams: 1 ulfc126380_tm_João_Gabriel.pdf: 3175623 bytes, checksum: a2162b7f8dbb6c56af652af38fbaaa51 (MD5) Made available in DSpace on 2021-06-04T16:24:26Z (GMT). No. of bitstreams: 1 ulfc126380_tm_João_Gabriel.pdf: 3175623 bytes, checksum: a2162b7f8dbb6c56af652af38fbaaa51 (MD5) Previous issue date: 2021

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Engenharia Clínica e Instrumentação Médica), Universidade de Lisboa, Faculdade de Ciências, 2021 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-06-11T15:02:42Z No. of bitstreams: 1 TM_Ana_Catarina_Lourenço.pdf: 8025894 bytes, checksum: fd34696d2b39248855f66a418d8c8cbb (MD5) Made available in DSpace on 2021-06-11T15:02:52Z (GMT). No. of bitstreams: 1 TM_Ana_Catarina_Lourenço.pdf: 8025894 bytes, checksum: fd34696d2b39248855f66a418d8c8cbb (MD5) Previous issue date: 2021

This brief study focuses on skin as sensory interface, and explores the latest discoveries in bioelectronic science. The work analyzes at what extent invisibility is possible by emulating nature, and if military applications can really benefit from technology that combines epidermal systems and virtual reality. The study shows that, to be effective, camouflage needs to match the environment — and to be disruptive. Adaptive and epidermal VR joint together can be a on the edge technology that might find effective military applications., Extended abstract published in the Book of Abstracts - Proceedings of the 2nd World Conference on Advanced Materials for Defense, edited by Raul Fangueiro — ISBN: 978-989-54808-4-5. Full paper will be published in Key Engineering Materials. This work received financial support by the European Social Fund (ESF) and by the Fundação para a Ciência e a Tecnologia (FCT), Portugal, under research grant SFRH/BD/136170/2018.

Breast cancer (BC) is the type of cancer that most greatly affects women globally hence its early detection is essential to guarantee an effective treatment. Although digital mammography (DM) is the main method of BC detection, it has low sensitivity with about 30% of positive cases undetected due to the superimposition of breast tissue when crossed by the X-ray beam. Digital breast tomosynthesis (DBT) does not share this limi tation, allowing the visualization of individual breast slices due to its image acquisition system. Consecutively, DBT was the object of this study as a means of determining one of the main risk factors for BC: breast density (BD). This thesis was aimed at developing an algorithm that, taking advantage of the 3D nature of DBT images, automatically clas sifies them in terms of BD. Thus, a quantitative, objective and reproducible classification was obtained, which will contribute to ascertain the risk of BC. The algorithm was developed in MATLAB and later transferred to a user interface that was compiled into an executable application. Using 350 images from the VICTRE database for the first classification phase – group 1 (ACR1+ACR2) versus group 2 (ACR3+ACR4), the highest AUC value of 0,9797 was obtained. In the classification within groups 1 and 2, the AUC obtained was 0,7461 and 0,6736, respectively. The algorithm attained an accuracy of 82% for these images. Sixteen exams provided by Hospital da Luz were also evaluated, with an overall accuracy of 62,5%. Therefore, a user-friendly and intuitive application was created that prioritizes the use of DBT as a diagnostic method and allows an objective classification of BD. This study is a first step towards preparing medical institutions for the compulsoriness of assessing BD, at a time when BC is still a very present pathology that shortens the lives of thousands of people.

A marcha hemiparética, causada pelo Acidente Vascular Cerebral (AVC), é assimétrica e tem uma velocidade reduzida, resultando numa marcha menos funcional. Os pacientes desejam andar com uma maior velocidade e melhorar o seu padrão de marcha. Assim, os médicos e terapeutas trabalham de modo a garantir que o indivíduo adota uma marcha o mais funcional possível. A análise da marcha realizada em laboratórios especializados como o Laboratório de Marcha do Centro de Medicina de Reabilitação de Alcoitão (CMRA), é muito relevante para a descrição e compreensão da marcha, permitindo fazer uma descrição espacial, temporal, cinemática e dinâmica da marcha, com o objetivo de desenvolver programas de intervenção direcionados para a problemática de cada paciente. No entanto, esta descrição é baseada na observação de gráficos, tornando o processo demorado e subjetivo, existindo uma lacuna na identificação dos parâmetros que tornam a marcha mais funcional. Desta forma, um dos objetivos deste projeto passou pelo desenvolvimento de uma pla taforma de análise biomecânica da marcha que permite tornar o processo mais objetivo e padronizado, calculando e extraindo 43 parâmetros biomecânicos que previamente eram analisados por observação visual dos gráficos. Recorrendo à plataforma desenvolvida, foram extraídos os parâmetros biomecânicos de 59 pacientes com hemiparesia, e foram calculadas as suas assimetrias. Com base na premissa de que a velocidade é um indicador de funcionalidade, foi realizado um teste de correlação entre os parâmetros biomecânicos extraídos e a velocidade, e entre as assimetrias e a velocidade, de modo a perceber que parâmetros estão mais relacionados com a velocidade e, como tal, com a funcionalidade. A plataforma desenvolvida facilita, de facto, o trabalho dos profissionais do Labora tório de Marcha do CMRA. Os resultados obtidos levam a crer que as assimetrias dos parâmetros biomecânicos não aparentam ser um bom indicador de funcionalidade, mas que a funcionalidade do membro não-parético parece estar relacionada com as articula ções proximais e a funcionalidade do membro parético com as articulações distais.

Submitted by Paula Guerreiro (passarinho@reitoria.ulisboa.pt) on 2021-11-10T12:41:21Z No. of bitstreams: 1 ulsd737048_td_Ana_Sá.pdf: 38904888 bytes, checksum: ef319d2367a2151bcfa59d96f0d53e65 (MD5) Made available in DSpace on 2021-12-13T15:42:14Z (GMT). No. of bitstreams: 1 ulsd737048_td_Ana_Sá.pdf: 38904888 bytes, checksum: ef319d2367a2151bcfa59d96f0d53e65 (MD5) Previous issue date: 2021-03 Bolsa de doutoramento Caixa Geral de Depósitos/IPL-2017

The ultrastructure of metaphase-I (MI) human immature oocytes is largely unreported. Which organelles are more frequent and their location are questions without a comprehensive answer. The aim of this dissertation is to describe the ultrastructure of MI human oocytes, qualitatively by describing their morphology, and quantitatively, using stereological analysis to analyse the relative volumes (Vv) of each organelle and their distribution. Moreover, the quantitative analysis of organelles in MI oocytes is compared to prophase-I oocytes, quantitatively analysed in a previous study. MI oocytes, retrieved after ovarian stimulation, were processed for transmission electron microscopy, serially sliced and photographed. Stereological analysis using a classical manual stereological technique based on point-counting was performed to compute relative volumes occupied by each organelle. Kruskal-Wallis test and Mann-Whitney U-test (with Bonferroni correction) were used to compare Vv occupied by organelles in the entire oocyte and three oocyte regions – cortex, subcortex and inner ooplasm. Mitochondria and smooth endoplasmic reticulum (SER) elements are the organelles with highest Vv in the entire oocyte and each oocyte region. No significant differences were detected in Vv of mitochondria and SER tubules, tubular aggregates, small vesicles and medium vesicles among oocyte regions. SER large vesicles show significant higher Vv (P = 0.011) in the inner ooplasm than in the subcortex (0.2% vs 0%). Conversely, cortical vesicles present significant higher Vv (P = 0.004) in the cortex than in the subcortex (0.96% vs 0.1%) and inner ooplasm (0.96% vs 0.1%); and vesicles with dense granular contents present significant higher Vv (P = 0.005) in the cortex than in the subcortex (0.1% vs 0%). No significant differences were observed in Vv of dictyosomes and lysosomes. These descriptive and quantitative data, combined with similar analysis of metaphase-II human oocytes and molecular approaches, could contribute to the improvement of stimulation protocols and in-vitro maturation methods.

As expressões faciais são uma forma de comunicação não-verbal muito importante para a transmissão de informações sociais entre seres-humanos e a sua descodificação é um processo eminentemente automático e espontâneo. Existem fatores que podem fazer variar a atenção visual dada a cada elemento da expressão facial, podendo esta atenção depender da emoção apresentada ou do tipo de observador. A monitorização da atenção visual durante este processo poderá ser feita através de um eye tracker. Porém, estes dispositivos são muito dispendiosos. Para estudar a dinâmica dos movimentos oculares durante a descodificação de expressões faciais com diferentes cargas emocionais, recorreu-se a um eye tracker de baixo custo, desenvolvido e otimizado em teses anteriores [1, 2]. Sendo este o primeiro estudo realizado com o referido dispositivo em contexto real, foi primeiramente necessário validá-lo. Para tal, seguiu-se a metodologia de um estudo realizado em 2011[3] cujo objetivo foi examinar a evolução da direção do olhar de indivíduos saudáveis enquanto estes examinam diferentes expressões faciais representando emoções básicas e investigar se existem padrões, na dinâmica de atenção visual, específicos a cada emoção apresentada. Depois de validado o dispositivo, foi realizado um estudo comparativo dos movimentos oculares, durante a descodificação de expressões faciais emocionais, entre indivíduos com formação musical e indivíduos sem este tipo de formação. De um modo geral observou-se um foco da atenção visual nos olhos, independentemente da emoção apresentada. A região da boca revelou ser um alvo mais frequente da atenção visual na emoção alegria em comparação com as outras emoções. Para o grupo de indivíduos com formação musical, verificou-se uma frequência e duração de fixações na região dos olhos superior aos valores apresentados para o grupo sem formação musical e foi também verificada uma dinâmica de olhar menos dispersa e menos errática para o grupo com formação musical. Facial expressions are a very important form of non-verbal communication for the transmission of social information between human beings and it’s decoding is an eminently automatic and spontaneous process. There are factors that can impact the visual attention given to each element of the facial expression, and this attention may depend on the emotion presented or the type of observer. Monitoring of visual attention during this process can be done using an eye tracker. However, these are very expensive. To study the dynamics of eye movements during the decoding of emotional expressions, a low-cost eye tracker was developed and optimized in previous theses [1, 2]. Since this is the first study carried out with the referred device in a real context, it was first necessary to validate it. To this end, we followed the methodology of a study carried out in 2011 [3] whose objective was to examine the scanpaths of healthy individuals while they observed different facial expressions representing basic emotions. After validating the device, a comparative study of eye movements was performed, during the decoding of emotional expressions, between individuals with musical training and individuals without this type of training. In general, there was a focus of visual attention in the eyes, regardless of the emotion presented. The mouth region proved to be a more frequent target of visual attention in the emotion of joy compared to other emotions. For the group of individuals withmusical training, there was a frequency and duration of fixations in the eye region higher than the values presented for the group without musical training and there was also a less dispersed and less erratic scanpath for the group with musical training.

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Engenharia Clínica e Instrumentação Médica), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-05-19T15:01:43Z No. of bitstreams: 1 ulfc126280_tm_Patrícia_Falcão.pdf: 653837 bytes, checksum: 3299ba958c6534a47012db24c4f86111 (MD5) Made available in DSpace on 2021-05-19T15:01:56Z (GMT). No. of bitstreams: 1 ulfc126280_tm_Patrícia_Falcão.pdf: 653837 bytes, checksum: 3299ba958c6534a47012db24c4f86111 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Sinais e Imagens Médicas), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-05-18T14:33:23Z No. of bitstreams: 1 ulfc126267_tm_Rita_Alves.pdf: 5255033 bytes, checksum: 47018e8ba196224e89f7a0fcc9539c59 (MD5) Made available in DSpace on 2021-05-18T14:33:33Z (GMT). No. of bitstreams: 1 ulfc126267_tm_Rita_Alves.pdf: 5255033 bytes, checksum: 47018e8ba196224e89f7a0fcc9539c59 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Sinais e Imagens Médicas), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-06-09T16:05:40Z No. of bitstreams: 1 ulfc126426_tm_Maria_Francisca_Canais.pdf: 547074 bytes, checksum: e61036e440a73cfed7dbd1a6acbb4c30 (MD5) Made available in DSpace on 2021-06-09T16:05:50Z (GMT). No. of bitstreams: 1 ulfc126426_tm_Maria_Francisca_Canais.pdf: 547074 bytes, checksum: e61036e440a73cfed7dbd1a6acbb4c30 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Engenharia Clínica e Instrumentação Médica), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2020-12-15T18:08:40Z No. of bitstreams: 1 ulfc125978_tm_Claudia_Espadinha.pdf: 5101899 bytes, checksum: cbd5699a5a8e5860919235a2f293dd01 (MD5) Made available in DSpace on 2020-12-15T18:08:59Z (GMT). No. of bitstreams: 1 ulfc125978_tm_Claudia_Espadinha.pdf: 5101899 bytes, checksum: cbd5699a5a8e5860919235a2f293dd01 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Sinais e Imagens Médicas), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2020-12-17T18:02:31Z No. of bitstreams: 1 ulfc125988_tm_Filipe_Galvão.pdf: 4403404 bytes, checksum: 65b0601e70430e28da460cd9dc29638b (MD5) Made available in DSpace on 2020-12-17T18:02:41Z (GMT). No. of bitstreams: 1 ulfc125988_tm_Filipe_Galvão.pdf: 4403404 bytes, checksum: 65b0601e70430e28da460cd9dc29638b (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Radiações em Diagnóstico e Terapia), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2020-12-23T16:51:10Z No. of bitstreams: 1 ulfc126047_tm_Joana_Antunes.pdf: 3394121 bytes, checksum: 685c379e8e8341b809fb85d3b282c475 (MD5) Made available in DSpace on 2020-12-23T16:51:20Z (GMT). No. of bitstreams: 1 ulfc126047_tm_Joana_Antunes.pdf: 3394121 bytes, checksum: 685c379e8e8341b809fb85d3b282c475 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Engenharia Clínica e Instrumentação Médica), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-06-08T17:43:47Z No. of bitstreams: 1 ulfc126416_tm_Marcia_Abreu.pdf: 2698834 bytes, checksum: 32ede00ef9af538f723b27038c7ae7f1 (MD5) Made available in DSpace on 2021-06-08T17:43:58Z (GMT). No. of bitstreams: 1 ulfc126416_tm_Marcia_Abreu.pdf: 2698834 bytes, checksum: 32ede00ef9af538f723b27038c7ae7f1 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Engenharia Clínica e Instrumentação Médica), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-05-06T17:52:25Z No. of bitstreams: 1 ulfc126205_tm_Joana_Chong.pdf: 2331813 bytes, checksum: 76a3ba9c56c2bd72053c25cdfa93b851 (MD5) Made available in DSpace on 2021-05-06T17:52:37Z (GMT). No. of bitstreams: 1 ulfc126205_tm_Joana_Chong.pdf: 2331813 bytes, checksum: 76a3ba9c56c2bd72053c25cdfa93b851 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Sinais e Imagens Médicas), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-05-20T17:16:15Z No. of bitstreams: 1 ulfc126288_tm_Natalia_Marques.pdf: 5312686 bytes, checksum: cda648e8708aa9bbdc03125a6ee62eac (MD5) Made available in DSpace on 2021-05-20T17:16:27Z (GMT). No. of bitstreams: 1 ulfc126288_tm_Natalia_Marques.pdf: 5312686 bytes, checksum: cda648e8708aa9bbdc03125a6ee62eac (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Biofísica Médica e Fisiologia de Sistemas), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-05-25T14:00:24Z No. of bitstreams: 1 ulfc126319_tm_Maria_João_Palma.pdf: 4257436 bytes, checksum: 9fd4ccef48b8c1a44118560a4485551c (MD5) Made available in DSpace on 2021-05-25T14:00:44Z (GMT). No. of bitstreams: 1 ulfc126319_tm_Maria_João_Palma.pdf: 4257436 bytes, checksum: 9fd4ccef48b8c1a44118560a4485551c (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Engenharia Clínica e Instrumentação Médica), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-06-07T17:35:37Z No. of bitstreams: 1 ulfc126406_tm_Luis_Venancio.pdf: 11010379 bytes, checksum: 57a80763041adf6c1f0efb7b0d6c7a2f (MD5) Made available in DSpace on 2021-06-07T17:35:48Z (GMT). No. of bitstreams: 1 ulfc126406_tm_Luis_Venancio.pdf: 11010379 bytes, checksum: 57a80763041adf6c1f0efb7b0d6c7a2f (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Sinais e Imagens Médicas), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-05-21T17:14:10Z No. of bitstreams: 1 ulfc126294_tm_Monica_Ferreira.pdf: 9562606 bytes, checksum: e872fcaba1188118b6be92f235653c76 (MD5) Made available in DSpace on 2021-05-21T17:14:20Z (GMT). No. of bitstreams: 1 ulfc126294_tm_Monica_Ferreira.pdf: 9562606 bytes, checksum: e872fcaba1188118b6be92f235653c76 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Engenharia Clínica e Instrumentação Médica), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-05-04T14:50:09Z No. of bitstreams: 1 ulfc126198_tm_Inês_Fragoeiro.pdf: 22227887 bytes, checksum: 09b932a1c795f7bf26919bbbaccae249 (MD5) Made available in DSpace on 2021-05-04T14:50:29Z (GMT). No. of bitstreams: 1 ulfc126198_tm_Inês_Fragoeiro.pdf: 22227887 bytes, checksum: 09b932a1c795f7bf26919bbbaccae249 (MD5) Previous issue date: 2020

Tese de mestrado integrado em Engenharia Biomédica e Biofísica (Biofísica Médica e Fisiologia de Sistemas), Universidade de Lisboa, Faculdade de Ciências, 2020 Submitted by Cristina Manessiez (camanessiez@fc.ul.pt) on 2021-06-04T17:54:38Z No. of bitstreams: 1 ulfc126382_tm_João_Campagnolo.pdf: 8597523 bytes, checksum: 13fb5f1b41afbd5c006a7a4428d548d5 (MD5) Made available in DSpace on 2021-06-04T17:54:50Z (GMT). No. of bitstreams: 1 ulfc126382_tm_João_Campagnolo.pdf: 8597523 bytes, checksum: 13fb5f1b41afbd5c006a7a4428d548d5 (MD5) Previous issue date: 2020

Diabetes mellitus is a chronic disease whose numbers of affected individuals are substantially increasing. Diabetes-associated complications include Diabetic Retinopathy (DR), the main cause of visual impairment and blindness. DR is a progressive pathology affecting the retina, caused by chronic hyperglycaemia. From the several associated risk factors, hypertension has been considered a key player in DR. Renin-Angiotensin System (RAS) is the hormonal cascade responsible for the control of blood pressure. This system can be divided in two main axes: a deleterious one that promotes vasoconstriction, inflammation, angiogenesis and increased oxidative stress and a parallel axis known to counterbalance these effects. Both axes’ components have been identified in the eye and we have shown that the prejudicial axis is modulated by glucose. In this work we aimed to characterize the expression of the RAS protective axis in retinal pigment epithelium (RPE) cells and evaluate the effects of its modulation by glucose and direct renin inhibition. The expression of Angiotensin-converting enzyme 2, Angiotensin (1-7) and Mas1 were detected in RPE cells. A decrease in the expression of these components was observed in high glucose conditions, showing a RAS imbalance. These results are supported by the results obtained in the retina of a type 1 diabetic animal, the Ins2Akita mouse. RAS dysregulation triggers DR hallmarks such as oxidative stress and angiogenesis. Thus, RAS blockade can affect DR development and progression. When simulating RAS overactivation, expression of the protective components increased over time, pointing to an initial protective role that needs further investigation. When treating cells with a renin-blocker, aliskiren, decreased levels of Angiotensin-converting enzyme 2 and Mas1 were observed, probably caused by the non-formation of angiotensin peptides. Overall our results are indicative of an early effect of the protective arm of RAS in the glucose-induced RAS dysregulation.

The emergence of counterfeit and substandard medicines is a worldwide problem that has been growing over the years. The use of these medicines puts people’s health at risk, which can even result in death. It is estimated that 5% of the antibiotics on the market are counterfeit and that most of these are -lactams. Despite being a worldwide problem, its incidence is higher in less developed countries due to limited access to quality medical products at an affordable price and the lack of sophisticated means for their detection. Most of the techniques used to validate the purity of an antibiotic require specialized technicians, adequate infrastructures and expensive equipment, which makes it impossible to use in resource-poor countries. The main goal of this dissertation is to develop an inexpensive and user-friendly device, able to detect counterfeit -lactam antibiotics. This detection is based on checking for the pres-ence of the active pharmaceutical ingredient (API) in the antibiotic and verifying if it is within the correct dosage. The device consists of a paper-based platform, obtained by Lab-on-Paper technology, with an enzyme-free detection, based on the use of gold nanoparticles. In the presence of the API, the HAuCl4 is reduced and gold nanoparticles (GNPs) are formed, allowing a colori-metric detection. This reaction was tested with solutions of different volumes and concentrations of antibiotic and calibration curves were obtained, from digital analysis, to obtain the dosage of that antibiotic. Solutions of common drug replacements were used to verify if there is formation of GNPs without the API.

A produção de matrizes fibrosas para a aplicação de estimulação eléctrica a células neuronais pode contribuir para a sua regeneração. A incorporação de poli (3,4 – etilenodioxitiofeno) (PE-DOT), na matriz poderá conferir-lhe a necessária condutividade. A dificuldade em dissolver este polímero condutor requer que a matriz seja produzida a partir de dispersões de partículas finas. Neste trabalho, o PEDOT foi obtido por polimerização química dos monómeros (EDOT), usan-do cloreto de ferro como agente oxidante. Foram seguidas duas vias: numa o EDOT foi disper-so em ácido acético e o oxidante adicionado dissolvido em etanol, na outra utilizou-se um lí-quido iónico, bis(trifluorometilsulfonil)imida de 1-butil-3-metilimidazólio[BMIM][T N], para dissolver EDOT e o oxidante. Para eliminar potenciais resíduos citotóxicos, os pós foram sujei-tos a um processo de lavagens controlado por espectroscopia de UV-vis e testes colorimétri-cos. Para a sua análise recorreu-se a medidas de condutividade, espectroscopias de UV-vis e de infravermelho por transformada de Fourier, dispersão dinâmica de luz e testes in vitro de citotoxicidade. Os pós com dimensões na ordem dos 300 nm, revelaram possuir estados con-dutores (estados de polarão) e ausência de citotoxicidade. Cada um dos pós obtidos foi disperso numa solução de gelatina de peixe, numa razão PE-DOT/gelatina de 20%, que foi eletrofiada para produzir matrizes fibrosas. Também foi produzi-da uma matriz de gelatina. As matrizes foram reticuladas e caracterizadas por FTIR, microsco-pia eletrónica de varrimento e sujeitas a um teste in vitro. As células epiteliais da linha Vero semeadas sobre as matrizes aderiram e proliferaram como indicaram os resultados dos testes colorimétricos e imagens de fluorescência dos núcleos.

The blood pressure is an important factor in the diagnosis and evaluation of several diseases, such as acute myocardial infarction and stroke. This way, continuous monitorization of this parameter is crucial to a correct health evaluation. The current methods, like the oscillometric method, have some major drawbacks, that can influence the output values or even make the measurements impossible. One example is the high frequency evaluation of the blood pressure, in the standard used methods the process of measuring can take up to 3 minutes, and a waiting time is necessary between consecutive measurements. This dissertation presents two different cuffless solution to solve those problems. One based on physical models of the human body, and the other using machine learning techniques. In the first solution seven models that correlate pulse transit time and blood pressure, deducted by different authors, were tested to evaluate which one performed better. The testes were performed in a custom dataset acquired at Fraunhofer AICOS and in clinical environment, with two different devices (low cost device and medical grade device). The results indicate that pulse transit time can be used to track blood pressure, the developed device/method was evaluated as grade A based in the Standard IEEE 1708-2014. The second solution it’s a proof of concept using a public database and three different machine learning methods (Random Forest, Neural Network and AdaBoost). Two sets of features are calculated from the ECG and PPG signals, one using TSFEL (spectral, frequency and time domain features) and a total of 15 custom features. The proposed method outperforms the methods presented in bibliography with mean absolute error of 3.6 mmHg and 2.0 mmHg to systolic and diastolic blood pressure respectively.