Repositório RCAAP

O negócio das Companhias de seguros baseia-se essencialmente numa gestão de riscos, de forma a garantir que se mantêm sempre solventes. Face a fatores como os desenvolvimentos no mercado eu ropeu, o aumento da concorrência, a volatilidade dos mercados financeiros, entre outros, sentiu-se a necessidade de desenvolver novas exigências regulamentares que permitissem a estabilidade das organi zações e a proteção dos tomadores de seguros. Para tal, a União Europeia emitiu a diretiva de Solvência II, que entrou em vigor a 1 de janeiro de 2016 e trouxe, numa visão económica, uma abordagem sensível aos riscos, dependendo os Requisitos de Capital de Solvência dos riscos subscritos e da estratégia de investimento da Companhia. Estes montantes podem ser apurados atráves de uma fórmula padrão, um modelo interno desenvolvido à medida das características da Companhia ou um modelo parcial interno, que utiliza a fórmula padrão mas com parâmetros calibrados ao nível dos dados da mesma, os Parâmetros Específicos da Empresa. Estes parâmetros têm em conta a exposição da Companhia face aos riscos que se pretende quantificar e podem ser calculados através dos métodos previstos no Regulamento Delegado (UE) 2015/35 da Comissão de 10 de outubro de 2014, que completa a Diretiva 2009/138/CE do Parla mento Europeu e do Conselho relativa ao acesso à atividade de Seguros e Resseguros e ao seu exercício (Diretiva de Solvência II). O foco deste projeto passa pelo estudo e implementação dos métodos anteriormente mencionados para o cálculo dos Parâmetros Específicos da Empresa utilizados na componente de Prémios e Reservas de duas Companhias: uma do ramo Não Vida e a outra do ramo de Saúde. Pretende-se com isto efetuar a análise dos resultados e consequentes impactos no Requisito de Capital e rácio de Solvência das duas Companhias, face ao cenário base de uso dos parâmetros padrão. Tal análise permite concluir se é possível o uso dos Parâmetros Específicos da Empresa para as duas Companhias em estudo e de que forma se posicionam as mesmas em termos de volatilidade nos Prémios e Reservas face ao benchmark europeu.

Despite the high maturity levels of CTI (Cyber Threat Intelligence) tools, techniques, procedures and frameworks, there are still gaps that must be considered and addressed. More than 50% of the world’s population is now online and growing, as the COVID-19 pandemic is pushing the large-scale adoption of technology in the most diverse areas. This context, aligned to the emerging technologies (e.g.: Cloud-computing, IoT, 5G) is enabling, allowing, and amplifying more complex and faster cyber-attacks. “Security-by design” is not yet the main principle, as products need to be quickly deployed into the market, delivering vulnerable targets into the Internet ecosystem. It is estimated that cy bercrime inflict damages of 6 billion USD in 2021, growing 15% per year, positioning it as the world’ third-largest economy, reaching 10.5 billion USD in 2025 [1]. Cyberattacks on critical infrastructures was considered the fifth top risk in 2020, as structural industries and sectors are juicy targets. On the other hand, the likelihood of detection and prosecu tion is estimated to be 0.05% in the USA [2]. To fight this threat and reduce the risk, it is essential that CTI parties join forces to improve coordination and cooperation, to reduce the time between the generation of CTI and its dissemination and achieve the balance between CTI in-time-dissemination and high-quality CTI. The quality of CTI is a huge barrier: most of the platforms ingest data from paid feeds and OSINT sources, gathering, filtering, analyzing, and aggregating, usually with little or no data-quality assessment. This increases the pressure on cyber-security analysts, who deal with plenty of generated alerts. IoCs (Indicator of Compromise) must go through an assessment process and be scored, so CTI consumers can decide and suit the measures accordingly. According to ENISA 2020 CTI survey [3], only 4% of CTI users can implement processes to measure CTI efficiency. This dissertation presents an overview of the existing CTI methodologies and technologies, proposing one solution to be adopted and integrated in CTI tools to assess, qualify, score and advise cyber-security analysts.

Background selection (BGS) and associative overdominance (AOD) are processes describing the effect of removal of deleterious mutations on linked neutral variation. Mathematical modelling and sim ulations predict that, in comparison to neutrality, BGS leads to reduced genetic diversity while AOD leads to an increase, with greater effects in low recombining regions. However, little is known about the impact of migration and hemizygosity (X-chromosomes or haplodiploid sex-determining system). Moreover, studies about divergent selection with migration usually ignore deleterious mutations. The objective of this thesis was to use simulations to predict genomic patterns of diversity and differenti ation resulting from the interaction between deleterious mutations, divergent selection, migration and haplodiploidy. We used forward-in-time simulations under an isolation-with-migration model with two populations, investigating variation in dominance, migration rate, recombination rate, and selective co efficients. To assess the impact of interactions between deleterious mutations and divergent selection, we performed simulations according to the demographic history from two sister species of haplodiploid Neodiprion sawflies to simulate a real-life based model. We found that the pattern of variation of the ratio of within population nucleotide diversity over its neutral expectation (π/πneutral) as a function of deleterious selective coefficient is only slightly influenced by gene flow. However, the ratio of between population differentiation (FST /FSTneutral) becomes closer to 1 as migration rate increases, suggesting a migration threshold above which linked selection has no effect on FST . We found a lower migration threshold for genomic regions under BGS than AOD. For fully and partially recessive deleterious mutations we discovered a transition between AOD and BGS as selective coefficients increased, both with and without migration. Finally, we introduce predictions for the genomic patterns of haplodiploid populations experiencing divergent selection and asymmetric gene flow.

The alignment of optical systems is a fundamental step during metrological tests to obtain meaningful and traceable results. Depending on the application, several approaches can be used to achieve the alignment of the components. However, for large and complex systems, the typical approach relies on the materialization of the optic axis using a gaussian-like light source. The work developed on this thesis is focused on this approach. Using a gaussian-like beam that serves as a reference source for axis materialization along with a bidimensional image sensor to monitor this light source, it is possible to align an optical system. Another requirement is an imaging processing technique. Depending on the application, several methods are commonly used to characterise and determine the center of a spot. The first step for this thesis was to make a brief study based on beam alignment techniques that are commonly used to have a better grasp on the state of the art associated with this process. The ideal beam processing method must be applicable to large images in real-time to allow the alignment to occur in an iterative way, while being as precise as possible with all the necessary information displayed in a computer screen to aid the operator in positioning optical components in the setups. To verify the potential of the proposed method, other commonly used methods were tested alongside it, using generated images with gaussian spots, benchmarking their performance and, consequently, obtaining positive results. Afterwards, a setup was constructed to test the proposed method to a wide variety of extreme conditions, comparing the results to another commonly used method for the same application. The automation of the process was achieved in LabVIEW 2017, a software capable of controlling the equipment and acquiring data automatically. The results obtained show that the method performed better than the other tested, and a further analysis can be done to fully characterise its performance. Some improvements can be made in the future, but nonetheless, the method is ready to be employed in a lab environment and achieve satisfactory results.

As comunidades escolares atuais, cada vez mais diversas do ponto de vista sociocultural, fenómeno associado aos movimentos migratórios globais, contêm uma complexidade multicultural intrínseca que cria contextos escolares muito desafiantes. Os docentes, primeiros interlocutores entre as culturas em diálogo, desenvolvem um conjunto de projetos tendo em vista criar respostas que ajudem a ultrapassar as dificuldades decorrentes de um diálogo escolar intercultural escasso e do predomínio dos currículos prescritivos monoculturais. O Projeto – ISI (Informar e Sensibilizar para a Intervenção) Contra a Violência de Género, Tolerância Zero à MGF/C (Mutilação Genital Feminina), realizado na Escola Secundária da Baixa da Banheira, é uma resposta escolar, construída em parceria com uma Organização Não-governamental (ONG), mandatada pela Secretaria de Estado da Igualdade, com o intuito de ser promotor da prevenção da MGF/C nos públicos escolares cuja origem sociocultural coloca as mulheres e meninas em risco iminente de serem sujeitas a esta prática milenar. Iniciando-se a reflexão em torno das questões da identidade e da diferença, passando pelos direitos humanos e a violência de género, procura-se articular teoricamente o potencial da educação e da escola na modelação de comportamentos, para que se caminhe em direção à hermenêutica diatópica enquanto processo coletivo, consciente, da nossa incompletude cultural mútua e promotor de uma educação em direitos humanos. No mundo cada vez mais interligado, a conceção multicultural dos direitos humanos (Santos, 1997, p. 15) tem cada vez mais sentido e levanta questões muito pertinentes em comunidades escolares, palco onde ocorre a ação do Projeto – ISI. O trabalho pretende identificar o potencial inovador do Projeto – ISI e o seu contributo para a prevenção da Mutilação Genital Feminina, a partir de dinâmicas educativas realizadas em ambiente escolar. A reflexão inicial pretende colocar em diálogo questões como a Identidade e a Diferença, os Direitos Humanos, o Relativismo Cultural e a Hermenêutica Diatópica e propõe-se a considerar a Educação Desenvolvente e a Educação em Direitos Humanos como uma ferramenta de ação preventiva da MGF/C junto de comunidades escolares onde a prática pode existir. Desenvolvido sob a forma de um Estudo de Caso, examina-se o caso da Escola Secundária da Baixa da Banheira e o Projeto que nela se desenvolveu, tendo em vista perscrutar o seu potencial inovador educativo, a partir de referentes propostos por diversos autores especialistas na matéria. A recolha de informação foi precedida de um desenho estratégico que pretende, dentro do possível, garantir a representação da diversidade real. As entrevistas semiestruturadas e a leitura das atas dos Conselhos de Diretores de Turma e dos Conselhos de Turma que participaram no projeto foram o manancial de informação a ser interpretada e a partir do qual se desenvolve o racional de análise do projeto. A descrição interpretativa que aqui se apresenta tem em vista evidenciar o potencial Inovador do projeto e o seu contributo para a prevenção da MGF/C, tendo sempre como pano de fundo que, enquanto prática, “viola um conjunto de direitos humanos fundamentais, normas e princípios de igualdade de género e não discriminação, bem como o direito inalienável à vida e o direito a ser livre de tortura ou tratamento cruel, desumano ou degradante” (Cerejo et al, 2015, p. 16).

This study discusses how students at risk of educational exclusion can restore their identity and improve learning and scientific literacy. This qualitative research adopts an interpretative orientation. Twenty-eight secondary school students, aged 16 to 20 participated. Data was collected of the interactions between students in lessons, focus group interview and written documents. Data analysis was inductive, consistent with a naturalistic research paradigm, and consisted of uncovering salient patterns, singularities, and themes associated with research aims. Several students, at risk of educational exclusion, value this kind of practice. By changing practice, the learning social context varied, as well as students’ identities and their relationships with knowledge, teachers and school.

O presente relatório de prática de ensino supervisionada é resultado do estágio curricular, em contexto de iniciação à prática profissional docente, decorrido entre 2019 e 2021, na Escola Secundária de Cacilhas-Tejo. Trata-se da descrição da nossa experiência enquanto alunos do Mestrado em Ensino de História no 3º Ciclo do Ensino Básico e Ensino Secundário e responsáveis, ao longo de um mês, pela docência de uma turma do 11º ano de escolaridade. O período histórico lecionado foi o Antigo Regime, e os conteúdos trabalhados foram a demografia dos séculos XVII e XVIII, a sociedade de ordens e o absolutismo régio de Luís XIV. Estes conteúdos foram lecionados com o apoio e orientação de teorias e metodologias pedagógicas de autores como David Ausubel, Jerome Bruner e Richard Arends, com vista a conjugar o momento de ensino-aprendizagem em sala de aula, com a capacidade intrínseca e a curiosidade/motivação dos alunos. Concluímos que, para um ensino completo, que responda tanto às competências a desenvolver pelos alunos como aos conteúdos que estes devem apreender, se torna necessária uma avaliação formativa constante. A importância da avaliação formativa reside, não só no conhecimento dos alunos pelo professor, mas também no devido feedback que os resultados formativos oferecem aos métodos de lecionação que o docente colocou em prática. O retorno crítico revela-se, uma necessidade fulcral do processo de ensino-aprendizagem, pelo que a avaliação formativa deve ser tanto diversificada como capaz de motivar os alunos.

The late-time accelerated expansion of the Universe has been a challenge for modern Cosmology since its detection. General Relativity explains this phenomenon via the introduction of the cosmological constant, Λ. This what is known as the standard cosmological model or ΛCDM. However, the cosmological constant solution brings a series of shortcomings that have guided cosmologists to explore and propose alternative models to describe the late-time cosmic acceler ation. These proposed alternatives have ranged from models of a dynamical dark fluid, known as Dark Energy, to long-scale modifications of the gravitational interaction, known as Modified Gravity models. Inside the Modified Gravity approach, one can construct theories of gravity built from the scalars associated with non-metricity (Q). The latter is the symmetric teleparallel description which has been shown to be equivalent to General Relativity. Nonetheless, its gener alisation with a general function of the non-metricity, f(Q), has generated symmetric teleparallel theories of gravity with exciting properties beyond the standard cosmological scenario. This dissertation aims to analyse the viability of f(Q) gravity theories as a challenging model for ΛCDM. We focus on the analysis of two forms of the f(Q) function: one which only changes the cosmological perturbations at large linear scale compared to ΛCDM and a second one which introduces a modification in both the background expansion history and in linear perturbations. We explore both the background phenomenology of the models and their impact on the cosmological observables. To this purpose, we adapt an Einstein-Boltzmann code, dubbed MGCAMB, by implementing these models. We then provide for the first time cosmological constraints by using Markov chain Monte Carlo methods and large sets of data spanning from measurements of the background expansion of the Universe to those of gravitational potentials, matter density and temperature fluctuations of the Cosmic Microwave Background radiation. We conclude our investigation with a model selection analysis using χ 2 statistics and deviance information criteria. They will inform us whether the f(Q) models analysed are supported by data over the ΛCDM scenario. In this analysis we identify specific features in the temperature-temperature power spectrum, matter power spectrum and lensing power spectrum which make the f(Q) model distinguishable from the ΛCDM scenario. Finally from the model selection analysis, we find that the f(Q) gravity models can show, in some cases, a better fit to current data than the standard model. The capability of suppressing large-scale temperature anisotropies is the main sources of a better agreement with data.

A crescente pressão sobre a infraestrutura das cidades, como resultado do movimento demográfico junto dos grandes centros, tem provocado o agravamento de vários problemas ambientais. Todavia, a promoção de soluções de mobilidade centradas nos interesses e dificuldades do utilizador surge como uma área de avanço tecnológico importante para promover o uso de transportes públicos e consequentemente reduzir o impacto ambiental e logístico das tendências mundiais de híper-urbanização e globalização. Em resposta a esta situação que surge o termo Mobility-as-a-Service (MaaS). Neste contexto, a empresa Card4B Systems S.A., que opera na vanguarda do setor em Portugal, lançou o projeto i4MaaS que procura explorar a interseção entre a tecnologia Blockchain e o conceito de MaaS. Com este cenário em mente, o projeto visa o desenvolvimento de um sistema de bilhética sustentável com o foco na privacidade do utilizador e na colaboração entre as empresas do setor. O trabalho apresentado consiste na continuação do projeto i4MaaS, nomeadamente na implementação e integração de uma solução com recurso à abordagem Self-Sovereign Identity no sistema de bilhética previamente idealizado, para a gestão das identidades digitais dos passageiros, transversal aos elementos do ecossistema de MaaS. Por fim, foi desenvolvido um modelo de deployment distribuído baseado em computação em nuvem com o objetivo de avaliar a aplicabilidade da solução num contexto nacional e internacional. Esta avaliação, revela o valor adicionado pela solução nos campos da privacidade, segurança e cooperação entre entidades face aos sistemas tradicionais.

The own-race bias is a robust effect where participants show better memory for faces of their race. Hills and Lewis (2011) found that shifting attention to distinctive black face features reduces the bias for white participants. Nevertheless, social categories such as race are graded and their members can vary on their representativeness of the category (E. Rosch & Mervis, 1975). Tanaka and Corneille (2007) showed atypical faces are more easily recognised than typical faces and Kleider-Offutt et al. (2017) demonstrated that prototypicality can affect the activation of category association. Our hypothesis states that the more typical a face is, the less efficient the attentional shift is in reducing the own-race bias. Our results replicated the own-race bias (ORB), but we could not replicate Hills and Lewis’s (2011) results. We also did not confirm our hypothesis, since the only effect on typicality was a higher recogni tion for atypical faces of both racial groups.

O presente texto pretende fazer um relato da Prática de Ensino Supervisionada, no âmbito do mestrado em Ensino de Artes Visuais da Universidade de Lisboa, mais concretamente o estágio realizado na Escola Secundária de Santa Maria, em Sintra, no ano letivo 2020/2021, na disciplina de Design Comunicação e Audiovisuais da turma de 1º ano do Curso Profissional Técnico de Multimédia. A disciplina na qual se implementou a Unidade Didática tem como principais objetivos desenvolver nos alunos competências a nível da comunicação visual e multimédia, preparando-os para a sua prática profissional futura ao concluírem o curso. A implementação incidiu no módulo de Edição Vetorial da disciplina e consistiu na criação de um cartaz embebido nas temáticas relativas à educação para a cidadania, com o intuito de, através da mudança social, direcionar o design para um público diferente e, igualmente, visando explorar as dimensões estética, técnica, criativa e comunicacional. O que deu origem e orientou este trabalho foi a vontade de motivar os alunos através de um exercício que lhes fosse aplicável num contexto real, que tivesse um propósito para além do produto e que, desta forma, aliasse os conteúdos da disciplina com a educação para a cidadania, promovendo assim aprendizagens significativas. A proposta pedagógica assentou na metodologia projetual de Munari, através da qual se efetuaram as seguintes atividades: pesquisa de temáticas relacionadas com ativismo; pesquisa de referências ligadas à cultura visual; exploração do processo criativo, através do desenho analógico; a edição vetorial, a concretização do cartaz; apresentação de trabalhos; e criação de uma galeria virtual. Verificou-se, através da análise, que a Unidade Didática permitiu, aos alunos, apropriarem-se dos conhecimentos e refletirem a partir da investigação e dos debates em aula e ampliarem o seu repertório com obras de referência no mundo da cultura visual, o que contribuiu para que analisassem, lessem e criassem imagens com mais maturidade. A maioria dos objetivos de aprendizagem foram cumpridos e os alunos compreenderam, com a experiência deste projeto, que o design pode ser uma ferramenta de mudança social.

COVID-19 is a contagious respiratory disease caused by the SARS-CoV-2 coronavirus, which has sparked a pandemic that will last into 2020 and 2021. Identifying decisive risk factors is critical to minimizing the risk of adverse events, especially death. This study aimed to perform a survival analysis of Portuguese COVID-19-positive individuals regis tered in Sistema Nacional de Vigilancia Epidemiol ˆ ogica (SINAVE) using a competing risk model, with ´ the aim of analyzing the factors associated with the occurrence of death. The dependent variable is the time until one of the possible outcomes occurs.The events of interest considered were Death from COVID-19, Recovery and Death from another cause. The independent variables were gender, age group and Regional Health Service. For statistical analysis, several statistical inference methods were used, namely estimation of cumulative correlation function, Gray’s test, and Fine and Gray’s model. During the period covered by the study, from March 2 to December 31 in 2020, 360 914 people were registered, of whom 5 197 had COVID-19 as the underlying cause of death, 411 died from other causes, and 313 377 recovered. The median time from infection to death from COVID-19 was 11,2 days, death from other causes was 13,5 days, and recovery was 16 days. Recovery was the most likely event for all groups defined by sex, age group, and Regional Health Service. Men are estimated to have a higher risk of dying from COVID-19 after contracting the virus than from any other cause. Based on the identification of the characteristics of patients at higher risk of death, specific care strategies should be developed to prevent the evolution of the disease leading to death.

This project aimed to develop policy sales models that would allow to characterize possible life cycles, differences between products and predict future sales. The policyholders, insured people and policies of a previously marketed set of products (α range) were comprehensively analyzed, so that there was a greater understanding of the business and population studied. For the elaboration of the models, the database was reduced to the variables Sales and Active Policies (monthly observations) for each of the products. In order to simplify the procedure, groups were constituted empirically by comparing graphs and per centages of the active policies of the products and the computation of clusters. Subsequently, the multiple regression models were developed, illustrated graphically and analyzed, with variables established by the backward selection method . With access to the in force products database (β range) at the delivery date of this project, they were allocated to the stipulated groups and a significant growth between the two ranges was verified, aligned with what was observed in the country, regarding the acquisition of health insurances, which incited the elaboration of two proposals for the improvement of predictive models. The predictions calculated were exposed on the observed sales charts and the existing similarities were evaluated. Finally, the percentage differences between the observed sales and those predicted by each model were studied (periodically) and the most precise model was found for each of the groups. For the first proposed improvement model, the observed sales position was analyzed in the obtained graphs in relation to the extremes of the calculated prediction intervals, their amplitude and the expected sales. Ultimately, the developed project exceeded the expectations and returned predictions close to reality, in such a manner that it will continue to evolve in the future, with special application in sales projection and budgeting.

No contexto dos objetivos de desenvolvimento sustentável, a sustentabilidade é cada vez mais importante em todos os sectores económicos, em particular, na indústria do vinho de forma a diminuir as suas emissões de gases com efeito de estufa e o consumo de água. A análise do ciclo de vida (LCA – Life Cycle Assessement) do vinho (produto) é uma metodologia essencial para identificar os processos que necessitam de otimização, desde a viticultura e vinificação ao embalamento, distribuição, consumo e tratamento de resíduos e fim de vida da garrafa. Se a LCA for realizada de acordo com as mesmas fronteiras (cultivo da uva, ou cultivo + adega, ou cultivo + adega + transporte, ou cultivo + adega + transporte + embalamento, ou cultivo + adega + embalamento + distribuição + uso + fim-de-vida, etc) , unidade funcional ( tipicamente 1 L de vinho ou uma garrafa de vinho de 0,75L), e as mesmas bases de dados de fatores de emissão (gCO2eq/unidade de input), poder-se-á comparar o desempenho ambiental de várias produtoras e eventualmente construir um benchmarking para guiar as produtoras no tema de sustentabilidade. Nesta dissertação foram analisados 39 estudos referentes à pegada carbónica e outros impactes ambientais e 19 estudos referentes à pegada hídrica. Nestes estudos foi possível identificar qual a categoria de impacte mais estudada, reunir os focos que mais contribuem para as emissões com efeito de estufa, de acordo com as diferentes fronteiras, e avaliar a dispersão dos valores para auxílio a estabelecimento de um benchmaking. Na viticultura, a aplicação de fertilizantes e a queima de combustível nos tratores (diesel) e na fase de produção de materiais, a produção de vidro (embalamento), são tipicamente os processos com maior impacte. Foi realizada a comparação entre os diferentes estudos, discutindo aqueles que apresentavam resultados muito acima ou abaixo dos restantes (outliers). O mesmo método foi realizado para a pegada hídrica, sendo que a pegada verde, a maior contribuidora, consiste na água da chuva, algo que depende da localização das diferentes vinhas e não é possível alterar. A divisão por fases do consumo de água não foi possível, devido à falta de detalhe dos estudos reunidos. No entanto, podemos concluir que para a pegada azul o maior consumo é proveniente do embalamento, na pegada verde na área de cultivo da uva e por fim, para a pegada cinzenta novamente no embalamento. Os resultados obtidos de acordo com as diferentes fronteiras nem sempre se encontravam próximos uns dos outros destacando esses estudos fora dos limites habituais. Assim, esta dissertação vai facilitar a criação de um benchmarking e consequentemente medidas de diminuição das emissões e uso de água de forma a tornar a produção do vinho mais sustentável. O valor mais baixo encontrado de kg de CO2eq/garrafa foi de 0,06 e o mais elevado de 3,22, definindo um valor de benchmarking de 1 kg de CO2eq/garrafa para a fronteira mais comum, que se inicia na plantação da uva e termina no embalamento. Para o consumo de água azul os seus valores, mais alto e baixo são de 0,15 e 184 L/garrafa, respetivamente e para a água cinzenta 0,11 e 733 L/garrafa. O valor de benchmarking foi de 135 L/garrafa excluindo a pegada verde porque é dependente do clima, localização. A fase do ciclo de vida que mais contribui para os valores altos são: o uso no caso da pegada carbónica e o cultivo no caso da pegada hídrica.

Age-related Macular Degeneration (AMD), a progressive neuroretinal degenerative disease, is the leading cause of irreversible vision loss in individuals over 65 in the western world. Although the mechanisms and progression of AMD are not fully understood, drusen and hyperreflective foci are among the key earliest signs of the disease and monitoring them using spectral domain optical coherence tomography (SD-OCT) imaging can help clinicians improve patient stratification and define the risk of disease’s progression. Observing patients over time with multiple assessments, as opposed to a single observation, increases the success in identifying the risk of progression. Also, the subjective task of assessment of huge amounts of imaging data in a clinical setting by hand is a problem that only recently the advances in computer-based image analysis have begun to solve. Moreover, computerized methods have opened the horizon of the search for under-recognized image biomarkers of AMD progression. We created tools and incorporated them in a unified graphical user interface (GUI) to automatically quantify and analyse drusen evolution over time in SD-OCT images of dry AMD patients using classic approaches and the App Designer functionality of MATLAB. The algorithm uses an improved segmentation of Bruch's membrane provided by the SPECTRALIS software to align the SD OCT image volumes relative to this layer. Then, within a user-defined central square around the fovea, it identifies and segments cuticular and soft drusen with a combined approach of two methods: a topographic analysis on slices near the maximum height of drusen and segmenting the RPE inner boundary (iRPE) in four different planes of the aligned SD-OCT volume. The watershed method is used to segregate drusen. Moreover, the algorithm automatically computes the drusen features relative to the overall region of interest (number of drusen, the percentage of drusen area in space within the user defined central region of interest, and the percentage of drusen volume within the same space, relative to the volume from Bruch’s membrane to the outer nuclear layer) and also per drusen (area of base, volume, median homogeneity inside and percentage of the presence of hyperreflective foci in all drusen identified). Finally, the algorithm can track the evolution of all of these features as the SPECTRALIS software performs intra-patient image alignment, although, in the future, it would be good to add a verification step to confirm they are properly aligned. Also, to increase usability, the GUI lets the user edit the segmentation of BM employed to align the images and to semi-manually revise the identified and segmented drusen. These algorithms were validated by comparing their performance with identifications performed manually using the GUI developed on a small sample of 10 patients and, using the algorithm in its entirety, achieved median sensitivity and precision of 0.88 and 0.92 respectively. The algorithm developed to automatically segment drusen also achieved medians accuracy, precision, sensitivity, and specificity of 0.92, 0.8, 0.9, and 0.92 respectively when validated on the same patients by evaluating its effectiveness in assessing the percentage of the total area occupied by drusen. Additionally, we computed the characteristics of the drusen of images of the 10 individual acquisitions used in the validation and also of images of 6 patients with 2 acquisitions spaced in time between 12 and 19 months. It proved the usability and good performance of all tools of the GUI, including the automatic longitudinal analysis tool. Nevertheless, both algorithms should be compared with identifications and segmentations performed by experts and tested in a higher number of patients. Also, the algorithms developed to perform volumetric segmentation and to obtain the drusen-specific features still need validation from specialists. Our results demonstrate the potential of the developed tools and GUI in a clinical setting to track the progression of dry AMD in patients and discover new biomarkers. It does not require a powerful processor while having a low processing time, making it easier to implement in the current clinical settings. At the same time, in the future, more tools can be incorporated such as a disease prediction algorithm regarding the established features and an interactive dashboard displaying the tracking of the features.

More than 50 million people worldwide suffer from epilepsy with a third of those being diagnosed with drug-resistant epilepsy where the seizures cannot be treated through pharmacotherapy. In these cases, surgical removal of the epileptic brain tissue in patients is presented as an effective solution for treatment. However, for surgery success, it is vital that the accurate location of epileptic regions in the brain are known. Neuroimaging, specifically magnetic resonance imaging (MRI) and positron emission tomography (PET), commonly are the doctor’s allies in identifying these lesions’ locations responsible for the seizures. Focal cortical dysplasias (FCDs) are the most common type of cortical lesions respon sible for drug-resistant epilepsy in children. These lesions have highly heterogeneous masses, occur in different brain regions and result in different levels of visibility, corresponding to the second most in tractable type of lesion in adults with epilepsy. Moreover, among drug-resistant epilepsy cases, a third of these lesions cannot be correctly identified by neuroimaging experts, resulting in unsuccessful surgical planning and consequently ineffective treatment for patients. Recently, Generative Adversarial Networks (GANs) have demonstrated their value in neuroimaging anomaly detection. Therefore, this work pro poses the application of two different GAN methods – WGAN and CycleGAN - for anomaly detection of FCDs, in PET-MRI data of epileptic patients. A 3D patch-basis anomaly detection approach was therefore developed, inspired by previous works, to detect FCDs location by deconfounding acquisition noise and normal cortical variabilities in PET-MR brain scans of epilepsy patients. Therefore, the GAN models applied two different approaches for lesion detection: detection through reconstruction (WGAN) and detection through translation (CycleGAN). Moreover, the combination of PET and MR modalities was studied and compared to training the networks with individual imaging modalities instead. Through the results, it was possible to understand and correct some issues GAN models have when training with multimodal 3D data. However, both methods for anomaly detection were able to detect diseased brain areas in patients with very visible FCDs, although failing to identify them in patients with very subtle lesions. Recent studies will be briefly discussed in the conclusion, which propose new approaches and architectures for multimodality training, with great potential to improve the performance of the networks for anomaly detection in future works.

In Portugal like in many other countries we can find statistics in the mathematics curriculum, and statistics is taught in mathematics classes by mathematics teachers until university level. During the compulsory levels most Portuguese students learn statistics in a traditional way. In this paper, we analyse dialogues of 7th grade students during collaborative work. The main objective was to understand some benefits of this kind of work on students' statistical reasoning. Our results suggest that collaborative work improve students' statistical reasoning as a result of the discussions and clarifications of their ideas and resolutions but also shows that this kind of work plays an important role in supporting students' development of positive orientations towards statistics.

Os dinoflagelados são um grupo de microalgas, que inclui algumas espécies marinhas produtoras de toxinas lipofílicas que causam intoxicação diarreica por bivalves (DSP). Entre essas espécies, o dinoflagelado Prorocentrum lima (Ehrenberg) F. Stein 1878, raramente foi associado a eventos de DSP. Os bivalves são organismos filtradores e acumulam toxinas lipofílicas ao alimentar-se. O consumo de bivalves contaminados causa DSP em humanos. Neste trabalho, realizaram-se, em laboratório, experiências para investigar a possível acumulação de toxinas DSP, com origem em P. lima, em bivalves da costa Portuguesa. Para isso, utilizou-se uma cultura de P. lima (estirpe IO66-01, Algoteca da Faculdade de Ciências da Universidade de Lisboa), isolado da costa Portuguesa, como fonte de toxinas. O berbigão, Cerastoderma edule (Linnaeus, 1758), foi utilizado como modelo no ensaio laboratorial de exposição ao P. lima. A estirpe de P. lima estudada, apresentou uma taxa de crescimento de 0.09d -1 e a concentração média de toxinas por célula atingiu 22pg cél -1 de ácido ocadáico (AO) e 13pg cél -1 de dinofisistoxina 1 (DTX1). A biomassa liofilizada de células de P. lima foi adicionada aos tanques contendo o berbigão, correspondendo a uma concentração final de 154,5x103 cél L-1 . A análise das toxinas no berbigão, ao longo do tempo, permitiu detetar a presença de AO, de DTX1 e dos seus derivados esterificados. A concentração total de toxinas DSP no berbigão ultrapassou o limite máximo regulamentado de 160µg eq OA kg-1 . Os resultados obtidos sugerem que o P. lima poderá também estar na origem da toxificação de bivalves por toxinas DSP na costa Portuguesa. A ocorrência de proliferações desta microalga pode, por isso, constituir um risco para a saúde pública e ter impacto nas atividades económicas relacionadas com a produção, apanha e comércio de bivalves em Portugal.

Resting-state functional magnetic resonance imaging (rs-fMRI) has the potential to assist as a diagnostic or prognostic tool for a diverse set of neurological and neuropsychiatric disorders, which are often difficult to differentiate. fMRI focuses on the study of the brain functional Connectome, which is characterized by the functional connections and neuronal activity among different brain regions, also interpreted as communications between pairs of regions. This Functional Connectivity (FC) is quantified through the statistical dependences between brain regions’ blood-oxygen-level-dependent (BOLD) signals time-series, being traditionally evaluated by correlation coefficient metrics and represented as FC matrices. However, several studies underlined limitations regarding the use of correlation metrics to fully capture information from these signals, leading investigators towards different statistical metrics that would fill those shortcomings. Recently, investigators have turned their attention to Deep Learning (DL) models, outperforming traditional Machine Learning (ML) techniques due to their ability to automatically extract relevant information from high-dimensional data, like FC data, using these models with rs-fMRI data to improve diagnostic predictions, as well as to understand pathological patterns in functional Connectome, that can lead to the discovery of new biomarkers. In spite of very encouraging performances, the black-box nature of DL algorithms makes difficult to know which input information led the model to a certain prediction, restricting its use in clinical settings. The objective of this dissertation is to exploit the power of DL models, understanding how FC matrices created from different statistical metrics can provide information about the brain FC, beyond the conventionally used correlation family. Two publicly available datasets where studied, the ABIDE I dataset, composed by healthy and autism spectrum disease (ASD) individuals, and the ADHD-200 dataset, with typically developed controls and individuals with attention-deficit/hyperactive disorder (ADHD). The computation of the FC matrices of both datasets, using different statistical metrics, was performed in MATLAB using MULAN’s toolbox functions, encompassing the correlation coefficient, non-linear correlation coefficient, mutual information, coherence and transfer entropy. The classification of FC data was performed using two DL models, the improved ConnectomeCNN model and the innovative ConnectomeCNN-Autoencoder model. Moreover, another goal is to study the effect of a multi-metric approach in classification performances, combining multiple FC matrices computed from the different statistical metrics used, as well as to study the use of Explainable Artificial Intelligence (XAI) techniques, namely Layer-wise Relevance Propagation method (LRP), to surpass the black-box problem of DL models used, in order to reveal the most important brain regions in ADHD. The results show that the use of other statistical metrics to compute FC matrices can be a useful complement to the traditional correlation metric methods for the classification between healthy subjects and subjects diagnosed with ADHD and ASD. Namely, non-linear metrics like h2 and mutual information, achieved similar and, in some cases, even slightly better performances than correlation methods. The use of FC multi-metric, despite not showing improvements in classification performance compared to the best individual method, presented promising results, namely the ability of this approach to select the best features from all the FC matrices combined, achieving a similar performance in relation to the best individual metric in each of the evaluation measures of the model, leading to a more complete classification. The LRP analysis applied to ADHD-200 dataset proved to be promising, identifying brain regions related to the pathophysiology of ADHD, which are in broad accordance with FC and structural study’s findings.

Centrosome amplification (CA) is a hallmark of human cancers and is therefore a potential feature to explore for prognosis and therapy. However, the poor understanding of its origin and impact has limited its use in the clinic. Here, we used Barrett’s esophagus (BE) tumorigenesis as a human cancer model to test if CA has a contributory role in tumorigenesis and progression. Barrett’s esophagus is a premalignant condition and its neoplastic progression is a process with well-characterized and defined steps that go from metaplasia (premalignant condition - BE) to dysplasia (low- or high-grade intraepithelial neoplasia), adenocarcinoma (invasive neoplasia) and metastasis. Given that CA arises in BE metaplasia and that its incidence significantly increases from metaplasia to dysplasia, we posed the hypothesis that CA may play an important role in the acquisition of malignant properties. If this hypothesis is correct, then an increase of CA in metaplasia and/or a decrease of CA in dysplasia would be sufficient to respectively promote or reduce malignant properties such as invasiveness potential. To test these hypotheses, we first tested two different methods to reduce the CA levels in dysplasia cells: depletion of important molecules in the centrosome duplication cycle - PLK4, SAS6 and STIL - by siRNA, and inhibition of PLK4 activity using the specific chemical inhibitor centrinone-B. We then assessed changes in migration and invasive potential using 2D migration and invasion assays and 3D cell cultures. Both strategies were effective in reducing CA in dysplasia cells but as the treatment with the inhibitor acts directly on the protein's activity level, its effects were more quickly detected. Importantly, by providing the opportunity to reduce CA levels by affecting different molecules, the siRNA approach allowed us to confirm if the effects in the invasiveness properties of dysplasia cells is in fact a consequence of the reduction in the number of centrioles, and not just due to the dysfunction of specific molecules. Indeed, while results obtained with 2D migration and invasion assay were mostly inconclusive and further optimization is needed, we found that reduction of CA levels, with both approaches, reduced the invasiveness capacity of dysplasia cells in 3D cultures. Our findings therefore support a role for CA in promoting the invasiveness capacity in dysplasia, and provided important clues into their mechanisms that will be explored in the future. We then tested if an increase in CA is sufficient to induce invasion in metaplasia cells. To do this we first depleted p53 in metaplasia cells by siRNA, already known to increase CA in these cells, and then assessed their behavior using 3D cell cultures. Interestingly, our preliminary tests did not reveal any phenotypic differences in metaplasia cells with increased CA when compared to the controls. These findings indicate that increasing CA levels in metaplasia may not be sufficient, by itself, to induce an invasive capacity in metaplasia cells. In the future, testing different approaches to induce CA in these cells, as well as assessing their impact in cell migration and invasion using other assays, will be important to confirm this. By showing that CA contributes to the invasiveness potential in dysplasia, our results reveal the importance of CA in the acquisition of malignant properties in BE progression. These findings may contribute to new clinical tools, namely using CA as a biomarker of progression. Given widespread occurrence of CA in human tumors, our results may be extended to other cancers where CA is prevalent.