Animal health surveillance is an important part of animal health care, particularly in countries dependent on livestock for food production and international trade. There are two major issues related to the provision of e®ective surveillance activities. Firstly, for good information to become available, the design and conduct of data collection activ- ities should be carried out following sound statistical principles. In reality, constraints such as imperfect tests and unavoidably-biased sampling strategies hinder straightfor- ward analysis and interpretation of survey results. Risk-based surveillance is used to target high-risk sub-populations to increase e±ciency of disease detection; however, biased datasets are generated. This thesis develops methodologies to design risk-based surveillance systems and al- low statistically valid analysis of the inherently biased data they generate. The ¯rst example describes the development of a method to analyse surveillance data gathered for bovine spongiform encephalopathy (BSE). The data are collected from four dif- ferent surveillance streams of animals tested for BSE, with each stream containing unavoidable biases and limitations. In the BSurvE model, these data are combined with demographic information for each birth cohort to estimate the proportion of each birth cohort infected with BSE. The prevalence of BSE in a national herd can then be estimated using the method of moments, whereby the observed number of infected animals is equated with the expected number. The upper 95% con¯dence limit for the prevalence is estimated both for infected countries and for those where no BSE has previously been detected. A similar approach to that used in BSurvE is then applied to surveillance data for trichinellosis, for which risk-based post-mortem testing is also performed. Negative results from multiple species using di®erent, imperfect tests are combined to give an estimate of the upper 95% con¯dence limit of the national prevalence of trichinellosis in a reference population. This method is used to provide support for freedom from trichinellosis in Great Britain. A di®erent approach to risk-based surveillance is explored as the surveillance strategy for detection of exotic causes of abortion in sheep and goats in New Zealand is examined. Using a geographic information system (GIS) maps of disease risk factors were overlain to produce a risk landscape for the lower North Island. This was used to demonstrate how areas of high- and low-risk of disease occurrence can be identi¯ed and used to guide the design of a risk-based surveillance programme. Secondly, within one surveillance objective there may be many ways in which the available funds or human resources could be distributed. This thesis develops a method to assess BSE surveillance programmes, and provides tools to facilitate BSE detection on the basis of infection risk and to increase the e±ciency of surveillance strategies. A novel approach to allocation of resources is developed, where portfolio theory con- cepts from ¯nance are applied to animal health surveillance. The example of surveil- lance for exotic causes of sheep and goat abortion is expanded upon. Risk of disease occurrence is assessed for a population over di®erent time periods and geographical areas within a country, and portfolio theory used to allocate the number of tests to be carried out within each of these boundaries. This method is shown to be more likely to detect disease in a population when compared to proportional allocation of the available resources. The studies presented here show new approaches that allow better utilisation of imperfect data and more e±cient use of available resources. They allow development of surveillance programmes containing an appropriate balance of scanning and targeted surveillance activities. Application of these methods will enhance the implementation and value of surveillance in animal health.

Animal health surveillance is an important part of animal health care, particularly in countries dependent on livestock for food production and international trade. There are two major issues related to the provision of e®ective surveillance activities. Firstly, for good information to become available, the design and conduct of data collection activ- ities should be carried out following sound statistical principles. In reality, constraints such as imperfect tests and unavoidably-biased sampling strategies hinder straightfor- ward analysis and interpretation of survey results. Risk-based surveillance is used to target high-risk sub-populations to increase e±ciency of disease detection; however, biased datasets are generated. This thesis develops methodologies to design risk-based surveillance systems and al- low statistically valid analysis of the inherently biased data they generate. The ¯rst example describes the development of a method to analyse surveillance data gathered for bovine spongiform encephalopathy (BSE). The data are collected from four dif- ferent surveillance streams of animals tested for BSE, with each stream containing unavoidable biases and limitations. In the BSurvE model, these data are combined with demographic information for each birth cohort to estimate the proportion of each birth cohort infected with BSE. The prevalence of BSE in a national herd can then be estimated using the method of moments, whereby the observed number of infected animals is equated with the expected number. The upper 95% con¯dence limit for the prevalence is estimated both for infected countries and for those where no BSE has previously been detected. A similar approach to that used in BSurvE is then applied to surveillance data for trichinellosis, for which risk-based post-mortem testing is also performed. Negative results from multiple species using di®erent, imperfect tests are combined to give an estimate of the upper 95% con¯dence limit of the national prevalence of trichinellosis in a reference population. This method is used to provide support for freedom from trichinellosis in Great Britain. A di®erent approach to risk-based surveillance is explored as the surveillance strategy for detection of exotic causes of abortion in sheep and goats in New Zealand is examined. Using a geographic information system (GIS) maps of disease risk factors were overlain to produce a risk landscape for the lower North Island. This was used to demonstrate how areas of high- and low-risk of disease occurrence can be identi¯ed and used to guide the design of a risk-based surveillance programme. Secondly, within one surveillance objective there may be many ways in which the available funds or human resources could be distributed. This thesis develops a method to assess BSE surveillance programmes, and provides tools to facilitate BSE detection on the basis of infection risk and to increase the e±ciency of surveillance strategies. A novel approach to allocation of resources is developed, where portfolio theory con- cepts from ¯nance are applied to animal health surveillance. The example of surveil- lance for exotic causes of sheep and goat abortion is expanded upon. Risk of disease occurrence is assessed for a population over di®erent time periods and geographical areas within a country, and portfolio theory used to allocate the number of tests to be carried out within each of these boundaries. This method is shown to be more likely to detect disease in a population when compared to proportional allocation of the available resources. The studies presented here show new approaches that allow better utilisation of imperfect data and more e±cient use of available resources. They allow development of surveillance programmes containing an appropriate balance of scanning and targeted surveillance activities. Application of these methods will enhance the implementation and value of surveillance in animal health.

Animal health surveillance is an important part of animal health care, particularly in countries dependent on livestock for food production and international trade. There are two major issues related to the provision of e®ective surveillance activities. Firstly, for good information to become available, the design and conduct of data collection activ- ities should be carried out following sound statistical principles. In reality, constraints such as imperfect tests and unavoidably-biased sampling strategies hinder straightfor- ward analysis and interpretation of survey results. Risk-based surveillance is used to target high-risk sub-populations to increase e±ciency of disease detection; however, biased datasets are generated. This thesis develops methodologies to design risk-based surveillance systems and al- low statistically valid analysis of the inherently biased data they generate. The ¯rst example describes the development of a method to analyse surveillance data gathered for bovine spongiform encephalopathy (BSE). The data are collected from four dif- ferent surveillance streams of animals tested for BSE, with each stream containing unavoidable biases and limitations. In the BSurvE model, these data are combined with demographic information for each birth cohort to estimate the proportion of each birth cohort infected with BSE. The prevalence of BSE in a national herd can then be estimated using the method of moments, whereby the observed number of infected animals is equated with the expected number. The upper 95% con¯dence limit for the prevalence is estimated both for infected countries and for those where no BSE has previously been detected. A similar approach to that used in BSurvE is then applied to surveillance data for trichinellosis, for which risk-based post-mortem testing is also performed. Negative results from multiple species using di®erent, imperfect tests are combined to give an estimate of the upper 95% con¯dence limit of the national prevalence of trichinellosis in a reference population. This method is used to provide support for freedom from trichinellosis in Great Britain. A di®erent approach to risk-based surveillance is explored as the surveillance strategy for detection of exotic causes of abortion in sheep and goats in New Zealand is examined. Using a geographic information system (GIS) maps of disease risk factors were overlain to produce a risk landscape for the lower North Island. This was used to demonstrate how areas of high- and low-risk of disease occurrence can be identi¯ed and used to guide the design of a risk-based surveillance programme. Secondly, within one surveillance objective there may be many ways in which the available funds or human resources could be distributed. This thesis develops a method to assess BSE surveillance programmes, and provides tools to facilitate BSE detection on the basis of infection risk and to increase the e±ciency of surveillance strategies. A novel approach to allocation of resources is developed, where portfolio theory con- cepts from ¯nance are applied to animal health surveillance. The example of surveil- lance for exotic causes of sheep and goat abortion is expanded upon. Risk of disease occurrence is assessed for a population over di®erent time periods and geographical areas within a country, and portfolio theory used to allocate the number of tests to be carried out within each of these boundaries. This method is shown to be more likely to detect disease in a population when compared to proportional allocation of the available resources. The studies presented here show new approaches that allow better utilisation of imperfect data and more e±cient use of available resources. They allow development of surveillance programmes containing an appropriate balance of scanning and targeted surveillance activities. Application of these methods will enhance the implementation and value of surveillance in animal health.

Alzheimer’s sjukdom är en neurodegenerativ sjukdom som leder till demens. Den kan börja tyst i de tidiga stadierna och fortsätta under åren till en allvarlig och obotlig fas. Språkstörningar uppstår ofta som ett av de tidiga symptomen och kan till slut leda till fullständig mutism i de avancerade stadierna av sjukdomen. Därför är tal- och språkbaserad analys en lovande och icke-invasiv metod för att upptäcka Alzheimer’s sjukdom i dess tidiga stadier. Vårt mål är att använda maskininlärning för att jämföra informationmängden hos språkliga representationer i stora språkmodeller och förtränade akustiska representationer. Såvitt vi vet är detta första gången som GPT-3 och wav2vec2.0 har använts tillsammans för klassificering av Alzheimer’s sjukdom. Dessutom utnyttjade vi för första gången en kombination av två stora språkmodeller, GPT-3 och BERT, för denna specifika uppgift. Genom att utvärdera vår metod på två datamängder på engelska och svenska kan vi också belysa språkskillnaderna mellan dessa två språk. / Alzheimer’s disease is a neurodegenerative disease that leads to dementia. It can begin silently in the early stages and progresses over the years to a severe and incurable stage. Language impairment often emerges as one of the early symptoms and can eventually progress to complete mutism in advanced stages of the disease. As a result, speech processing is a promising and non-invasive approach for detecting Alzheimer’s disease in its early stages. Our objective is to compare the informativeness levels of linguistic embedding derived from large language models and pre-trained acoustic embedding extracted using wav2vec2.0, in a machine learning-based approach. To the best of our knowledge, this is the first time that fusing GPT-3 text embedding and wav2vec2.0 acoustic embedding has been explored for Alzheimer’s disease classification. In addition, we utilized a combination of two large language models, GPT-3 and BERT, for the first time on this specific task. By evaluating our method on two datasets in English and Swedish, we can also highlight the language differences between these two languages.

In dieser Doktorarbeit wird die Frage untersucht, wie erfolgreich deep learning bei der Diagnostik von neurodegenerativen Erkrankungen unterstützen kann. In 5 experimentellen Studien wird die Anwendung von Convolutional Neural Networks (CNNs) auf Daten der Magnetresonanztomographie (MRT) untersucht. Ein Schwerpunkt wird dabei auf die Erklärbarkeit der eigentlich intransparenten Modelle gelegt. Mit Hilfe von Methoden der erklärbaren künstlichen Intelligenz (KI) werden Heatmaps erstellt, die die Relevanz einzelner Bildbereiche für das Modell darstellen. Die 5 Studien dieser Dissertation zeigen das Potenzial von CNNs zur Krankheitserkennung auf neurologischen MRT, insbesondere bei der Kombination mit Methoden der erklärbaren KI. Mehrere Herausforderungen wurden in den Studien aufgezeigt und Lösungsansätze in den Experimenten evaluiert. Über alle Studien hinweg haben CNNs gute Klassifikationsgenauigkeiten erzielt und konnten durch den Vergleich von Heatmaps zur klinischen Literatur validiert werden. Weiterhin wurde eine neue CNN Architektur entwickelt, spezialisiert auf die räumlichen Eigenschaften von Gehirn MRT Bildern. / Deep learning and especially convolutional neural networks (CNNs) have a high potential of being implemented into clinical decision support software for tasks such as diagnosis and prediction of disease courses. This thesis has studied the application of CNNs on structural MRI data for diagnosing neurological diseases. Specifically, multiple sclerosis and Alzheimer’s disease were used as classification targets due to their high prevalence, data availability and apparent biomarkers in structural MRI data. The classification task is challenging since pathology can be highly individual and difficult for human experts to detect and due to small sample sizes, which are caused by the high acquisition cost and sensitivity of medical imaging data. A roadblock in adopting CNNs to clinical practice is their lack of interpretability. Therefore, after optimizing the machine learning models for predictive performance (e.g. balanced accuracy), we have employed explainability methods to study the reliability and validity of the trained models. The deep learning models achieved good predictive performance of over 87% balanced accuracy on all tasks and the explainability heatmaps showed coherence with known clinical biomarkers for both disorders. Explainability methods were compared quantitatively using brain atlases and shortcomings regarding their robustness were revealed. Further investigations showed clear benefits of transfer-learning and image registration on the model performance. Lastly, a new CNN layer type was introduced, which incorporates a prior on the spatial homogeneity of neuro-MRI data. CNNs excel when used on natural images which possess spatial heterogeneity, and even though MRI data and natural images share computational similarities, the composition and orientation of neuro-MRI is very distinct. The introduced patch-individual filter (PIF) layer breaks the assumption of spatial invariance of CNNs and reduces convergence time on different data sets without reducing predictive performance. The presented work highlights many challenges that CNNs for disease diagnosis face on MRI data and defines as well as tests strategies to overcome those.