UiS Brage
Not a member yet
19747 research outputs found
Sort by
Noen gode HINTS om den svimle pasienten?
Full text linkBakgrunn:
Akutt svimmelhet er en vanlig problemstilling som ambulansepersonell møter. Selv om tilstanden oftest skyldes godartede tilstander, kan det imidlertid være et symptom på hjerneslag i bakre kretsløp. Vi opplever en dissonans mellom de nasjonale retningslinjene og fraværet av prehospitale prosedyrer for vurderingen av disse pasientene, og ønsker derfor et kompetanseløft på temaet i ambulansetjenesten.
Hensikt:
Denne bacheloroppgaven undersøker potensialet for bruk av HINTS-undersøkelsen (Head Impulse, Nystagmus og Test of Skew) på pasienter med akutt vestibulært syndrom (AVS) for å utelukke sentrale årsaker til svimmelhet prehospitalt, med fokus på hjerneslag i bakre kretsløp.
Metode:
Vi gjennomførte en litteraturstudie og søkte etter kvantitative studier med forskning gjennomført i akuttmottak grunnet manglende prehospital forskning på temaet. Vi gjennomførte søk i databasene Cochrane, Epistemonikos, BMJ, MEDLINE og CINAHL.
Resultat:
Resultatene fra de inkluderte studiene antyder at HINTS-undersøkelsen har høy sensitivitet og spesifisitet når den utføres av tilstrekkelig trent personell, spesielt nevrologer. Oppgaven fremhever imidlertid betydelig heterogenitet i studiedesign og pasientutvalg, og understreker et behov for standardisert opplæring. Det er begrenset forskning på bruken av HINTS-undersøkelsen utført av andre enn legespesialister.
Avslutning:
Selv om HINTS-undersøkelsen viser lovende resultater som et verktøy for å ekskludere sentrale årsaker til svimmelhet, er det behov for mer forskning spesielt rettet mot prehospitale tjenester. Oppgaven understreker behovet for standardisert opplæring og et kompetanseløft for prehospital vurdering av den akutt svimle pasienten.
Nøkkelord: HINTS, svimmelhet, prehospitalt, AVS, hjernesla
Effekten av sosiale medier på forbrukeratferd: Hvordan påvirker annonsering på sosiale medier kjøpsbeslutning/kjøpsintensjon blant unge voksne?
No full textMålet med denne oppgaven er å undersøke hvordan annonsering på sosiale medier påvirker forbrukeratferd blant unge voksne, med særlig fokus på kjøpsbeslutning og kjøpsintensjon. Studien tar for seg hvordan ulike former for annonsering, som personaliserte annonser, trendbasert innhold og påvirkning fra influensere, påvirker unge voksnes holdninger, intensjoner og faktiske kjøpsvalg.
Underveis ble det benyttet relevante teorier som blant annet inkluderte «sosial aksept», «den kognitive beslutningsmodellen» og «sosial identitesteori». Vi presenterte også tidligere forskning som andre har gjort på lignende tema. Det var med på å belyse vår problemstilling og gav en bedre forståelse av oppgaven. I løpet av denne oppgaven ble det presentert ulike figurer og tabeller som var basert på svarene vi fikk fra spørreundersøkelsen som vi i begynnelsen av dette prosjektet hadde delt via sosiale medier.
Disse verdifulle svarene vi fikk fra respondentene var et viktig grunnlag for analysen i bacheloroppgaven vår. I tillegg til fremvisninger av tabeller, ble de også gjort grundige analyser av svarene fra respondentene, slik at vi fikk et best mulig bilde av hva som påvirket deres kjøpsatferd. Analysen inkluderte faktor-analyse, regresjonsanalyse og krysstabeller. Vi testet også reliabilitet og validitet av oppgaven.
Resultatene viser at sosiale medier spiller en sentral rolle i unge voksnes forbrukeratferd, men at påvirkning varierer med type innhold. Funnene viser at påvirkning fra influensere har størst effekt, mens personaliserte annonser også har tydelig virkning når innholdet oppleves som relevant. Trendbasert innhold har svakere, men målbar effekt. Studien viser at tillit, relevans og sosial påvirkning er sentrale faktorer for effektiv annonsering i denne målgruppen
Road Tunnel Safety: Communicating Efforts in Emergencies and Tunnel Users’ Learning
Full text linkPhD thesis in Risk management and societal safetyRoad tunnel safety concerns the management of systems and activities to keep road tunnels in a safe state. A number of social and technological systems are involved in both the design and operation of tunnels. Therefore, effective means of communication are needed between system actors. To understand the effect of communicating efforts, it is essential to uncover the knowledge and theories behind objects that are designed to create stimuli for a behavioural response among system actors. Hence, the following question was outlined as the major issue in this thesis: How is knowledge developed and applied to create change, confirmation or comprehension of current practice?
The management policy for road tunnel safety is rooted in the Vision Zero philosophy (the ethical principle, scientific knowledge and shared responsibility) and the Tunnel Safety Regulations. Norway has more than 1200 tunnels with different designs and standards. There is an average of 27 tunnel fires per year in Norway. Although it is a rare event, experience from previous accidents has shown that a major tunnel fire represents a threat to the life and health of road users. For example, the fires in the Mont Blanc Tunnel (France/Italy 1999, 39 fatalities), the Gotthard Tunnel (Switzerland 2001, 11 fatalities), the Oslofjord tunnel (Norway 2011, 34 people trapped in smoke, some seriously injured due to inhalation of toxic gases), and the Gudvanga tunnel (Norway 2013, 67 people trapped in smoke and inhaled toxic gases, 23 seriously injured and five with acute life-threatening injuries). Current expectations to evacuation is based on the self-rescue principle. For many tunnel designs, there are large uncertainties related to evacuation in the event of a fire.
Four research questions were developed to explore the major issue. The research questions were associated with: (1) understanding current tunnel safety management practices in light of the Vision Zero philosophy, (2) elaborating social construction as a scientific perspective in search for a set a rules to conceptualise an approach to the concept of risk, (3) exploring and understanding the use of virtual reality technology in training and research, (4) understanding mechanisms that are vital to promote learning effects of structured risk communication. The research is based on a holistic perspective on learning and systems safety, and takes an interdisciplinary approach in search for a more comprehensive understanding of the major issue. The research is presented in seven papers, concerning:
1) A discussion on how tunnel safety designs are communicated by national agencies with governmental tasks from a systems theory perspective applied to Vision Zero principles.
2) A comparison of theoretical perspectives related to risk acceptance in light of the Vision Zero philosophy and road tunnel safety, to identify the dominant mindset in the Norwegian approach.
3) An elaboration of social construction as a scientific perspective, showing that expressions of risk can be understood as causal, constitutive and pragmatic social constructions. An example from tunnel safety research is used to discuss five distinct risk concepts.
4) A tool to assess learning is applied to peer-reviewed literature to understand how and what knowledge is developed by using virtual reality technology in fire evacuation training.
5) A scoping review from a design science perspective is used to identify peer-reviewed literature documenting learning effects of using virtual reality in training and educational activities.
6) A randomized controlled experiment with 80 participants, using a virtual reality driving simulator to test stimuli-response effects of structured risk communication. The national information campaign “Safe behaviour in case of road tunnel incidents” is designed as structured risk communication and applied in the experiment as stimuli in the test-group.
7) A longitudinal study based on telephone interviews with participants from the experiment. Two interviews (total 125) were conducted over a two year period. The transcribed interviews constitute the empirical data for the longitudinal study, and are used to identify long-term learning effects of structured risk communication aimed at safe behaviour in road tunnels.
The research in this thesis has identified several aspects that promote or inhibit communication and learning in road tunnel safety management. One aspect concerns the meaning of Vision Zero as a philosophy for road tunnel safety. From a systems perspective on road tunnel safety, the Vision Zero values are inadequately communicated by national agencies with governmental tasks. The lack of constrains means that that the responsibility for interpreting the meaning of the concept of Vision Zero is transferred to lower level actors. This implies that the Development and application of objects, are based on lower levels understanding of road tunnel safety. An example is related to the self-rescue principle. Until recently, the self-rescue principle has been inadequately communicated to road users. This means that contemporary road users lack emergency preparedness competence in the event of a road tunnel fire.
Another aspect is the concept of risk. Risk is not a universal concept that can be isolated from the context, but hold different ontological, epistemological and methodological presumptions. This means that risk is defined in very different ways and can be understood as causal, constitutive and pragmatic social constructions. The lack of constraints implies that the concept of risk hold formal and informal power structures. Therefore it is necessary to develop norms to ensure a knowledge-based methodological approach to risk analysis and management.
Virtual reality technologies are objects that are often designed and applied as tools in behavioural studies and emergency preparedness training across a wide range of disciplines. While ‘virtual reality’ can come close to ‘real reality’, they are socially constructed objects that are applied in knowledge development. This means that a holistic perspective on learning is crucial in the development and application of the technology to ensure quality in knowledge development. The research shows limited documented learning effects of applying virtual reality technology in educational activities, but also that this artefact can be designed and applied in different ways as a method for knowledge development.
In 2019, the Norwegian Public Roads Administration launched the national information campaign “Safe behaviour in case of road tunnel incidents” on social media. The campaign consisted of six videos developed to increase public knowledge about road tunnel safety. The campaign can be understood as structured risk communication, that is, an object designed to create stimuli for a desired behavioural response among relevant actors in the context of uncertainty. A randomized controlled trial, consisting of 80 participants, was designed to test the effect of structured risk communication. A virtual reality driving simulator was used as a method to collect quantitative data. Qualitative data was collected through three questionnaires, as well as observation of participants during the experiment. The statistical results showed strong positive effects of structured risk communication.
Safe behaviour in road tunnels implies a latent emergency preparedness competence among road users. This involves learning that persists over time. A longitudinal study was designed to identify features of long-term learning. Two telephone interviews (125 in total) with participants from the experiment were conducted over a two-year period. The interview guides were based on a holistic perspective on learning and created stimuli for the expression of competence. The analysis showed a change over time in the respondents’ expression of competence, from quite fragmented to a consistent level of high competence. Emotional commitment to the context and reflection on decision-making and action (response) were identified as essential features of long-term learning. The analysis also showed that the design of the interview guides were vital as structured risk communication, and that participation in the social context was essential to create stimuli for long-term learning.
The research in this thesis has documented that emotional commitment to the context and reflection are important features to create long-term learning effects of structured risk communication. The KATS project has been vital in order to explore the major issue. This provided a context for participation, cooperation, and learning. The mixed method approach enabled interdisciplinary research to develop a more comprehensive understanding of cause-effect relationships. Road tunnel safety cannot be meaningfully understood without a more comprehensive understanding of the effect of communicating efforts, and how concepts and artefacts are developed and applied to assess, evaluate and communicate knowledge. This includes the understanding that objects can be designed to communicate road tunnel safety and create stimuli for change, confirmation and/or comprehension of current practices.
The design of structured risk communication targeting safe behaviour in road tunnels needs further empirical testing. Road tunnel safety involves a socio-technical system of actors interacting at different hierarchical levels and who have a shared responsibility for safety. To control road tunnel safety, knowledge about the system is an important element among the actors. Competence requirements should therefore be emphasized as a further step in the management of road tunnel safety
Edu-Choreography – Choreographing-Researching-Teaching Performative Re-Turnings of Dance (Education) in Primary School
No full textThe thesis focuses on the choreographic-pedagogical insights of dance education in Norwegian primary school, through the collaboration between a choreographer-researcher-teacher (I), four primary school teachers, and their pupils in third grade. The research builds on fieldwork in an urban primary school in Norway, where I carried out creative dance workshops with the pupils.
The research is a contribution to the post-qualitative and performative research field in dance education. The project was carried out as practice-led and a/r/tographic research, through bodily, relational, and affective approaches to research such as performative inquiry and diffractive analysis. The thesis has new-materialist and agential realist theoretical underpinnings, and builds on a transformative and performative approach to dance education, and an expanded view of choreography.
The thesis is article-based, with four material-discursive practices published as three research articles, a visual essay, and the mantle.
The contributions from the thesis reach from dance education to educational research practice. I propose a performative edu-choreographic approach where dance and dancing is a performative, relational, choreographic, and bodily dance education practice. Further I suggest dancing as a living, aesthetic, choreographic, and performative research practice part of dance education research. The entangled practice of choreographing-researching-teaching that this doctoral project promotes challenges the active separations between arts, research, and teaching.Forskningsprosjektet tar for seg de koreografi-didaktiske innsiktene som ble skapt gjennom et danseprosjekt i en norsk grunnskole. Her var jeg tilstede som koreograf-forsker-lærer og gjennomførte danseverksteder i kreativ dans i samarbeid med fire grunnskolelærere og deres elever i tredje trinn.
Avhandlingen er et bidrag til et post-kvalitativt og performativt forskningsfelt innenfor danseutdanning. Med kroppslige, relasjonelle og affektive forskningsmetoder, som performative inquiry og diffraktiv analyse, ble prosjektet skapt og gjennomført som praksisledet og a/r/tografisk forskning. I avhandlingen sammenfiltres nymaterialisme og agentisk realisme, en transformativ og performativ tilnærming til danseutdanning, og et utvidet syn på koreografi.
Avhandlingen er artikkelbasert, hvorav fire material-diskursive praksiser er publisert som tre forskningsartikler og et visuelt essay. Praksisene er videre syntetisert i denne kappen.
Forskningsbidragene i avhandlingen strekker seg fra danseutdanning til metodologisk praksis i danseutdanningsforskning. Jeg foreslår en performativ edu-koreografisk tilnærming til danseutdanning, med utgangspunkt i dans som kunstform og det å danse som en performativ, relasjonell, koreografisk og kroppslig praksis. Med utgangspunkt i denne dansepraksisen foreslår jeg dans som en levende, estetisk, koreografisk og performativ forskningspraksis i konteksten av danseutdanningsforskning. Med sammenfiltringen av kunst-forskningutdanning, og min rolle som koreograf-forsker-lærer, utfordrer dette doktorgradsprosjektet aktive separasjoner mellom dansekunst, forskning og undervisning
Mitigating oil spill risks from shipping on the Norwegian Continental Shelf: The impact of onboard response capabilities
Full text linkAn effective response to oil spills on the Norwegian Continental Shelf is significantly enhanced by the presence of efficient onboard response capabilities. Availability of such competencies is important as it can substantially affect the environmental impact and operational outcomes following an incident. Our focus area is the Norwegian Continental Shelf, a shallow seabed segment located off the coast of Norway, which represents the underwater extension of the land into the deep seas (Government.no, 2025a). Our focus area is oil spill incidents caused by shipping vessel accidents, which pose significant environmental and operational risks. Although the Norwegian Continental Shelf has a multi-actor oil spill preparedness system in place, many vessels operating in the region still lack sufficient onboard response capabilities. This master’s thesis analyses the current onboard capabilities for initial oil spill response, focusing on their effectiveness and identifying existing limitations that may increase risk. Conducting a literature-based analysis of past incidents, environmental conditions, risk factors, and regulatory frameworks, the findings highlight critical weaknesses in the current approach, including seasonal limitations, inadequate coordination, and limited access to remote spill locations. This master’s thesis emphasizes the need to enhance oil spill response capabilities, which is important for minimizing spill-related risks on the Norwegian Continental Shelf. However, these capabilities alone are insufficient. After conducting an analysis, the current onboard capabilities and their limitations, it is clear that their effectiveness depends on integration into a broader preparedness framework. This framework includes improved crew training, risk-oriented planning, robust infrastructure, and better coordination across the shipping industry. These insights aim to inform future decision-making actors within the shipping industry and public governance, particularly in adapting to growing environmental and operational challenges on the Norwegian Continental Shelf. This approach may resolve existing issues related to the shipping vessels and any potential oil loss during their operations (ITOPF, 1998).An effective response to oil spills on the Norwegian Continental Shelf is significantly enhanced by the presence of efficient onboard response capabilities. Availability of such competencies is important as it can substantially affect the environmental impact and operational outcomes following an incident. Our focus area is the Norwegian Continental Shelf, a shallow seabed segment located off the coast of Norway, which represents the underwater extension of the land into the deep seas (Government.no, 2025a). Our focus area is oil spill incidents caused by shipping vessel accidents, which pose significant environmental and operational risks. Although the Norwegian Continental Shelf has a multi-actor oil spill preparedness system in place, many vessels operating in the region still lack sufficient onboard response capabilities. This master’s thesis analyses the current onboard capabilities for initial oil spill response, focusing on their effectiveness and identifying existing limitations that may increase risk. Conducting a literature-based analysis of past incidents, environmental conditions, risk factors, and regulatory frameworks, the findings highlight critical weaknesses in the current approach, including seasonal limitations, inadequate coordination, and limited access to remote spill locations. This master’s thesis emphasizes the need to enhance oil spill response capabilities, which is important for minimizing spill-related risks on the Norwegian Continental Shelf. However, these capabilities alone are insufficient. After conducting an analysis, the current onboard capabilities and their limitations, it is clear that their effectiveness depends on integration into a broader preparedness framework. This framework includes improved crew training, risk-oriented planning, robust infrastructure, and better coordination across the shipping industry. These insights aim to inform future decision-making actors within the shipping industry and public governance, particularly in adapting to growing environmental and operational challenges on the Norwegian Continental Shelf. This approach may resolve existing issues related to the shipping vessels and any potential oil loss during their operations (ITOPF, 1998)
Generating Synthetic Electroencephalography Data Using Diffusion Models
No full textDementia is a condition with many causes that affects the cognitive domain, dis-
rupting the ability to function in daily living. Alzheimer’s disease (AD) is the most
common cause of dementia and its preceding state, mild cognitive impairment
(MCI). In both cases, the brain activity is affected and differences between healthy
adults and those afflicted with MCI or dementia can be detected using electroen-
cephalography (EEG). However, analyzing the resulting electroencephalogram
(EEG) requires expert knowledge and experience. For this reason, automatic
analysis has been applied to EEG data to lessen the burden on trained profession-
als. Driven by recent releases of publicly available EEG datasets, deep learning
(DL) methods have become more popular and there has been success in differ-
ent fields, including classification. However, the available datasets are usually of
limited size or suffer from class imbalance.
In this thesis we explore the use of generative diffusion models and utilize
synthetic data as a form of data augmentation (DA) to examine the effect of both
expanding and balancing the classes in a public dementia EEG dataset. We test
different diffusion models including well established models such as denoising
diffusion probabilistic models (DDPM) and latent diffusion models (LDM). We
also design and perform experiments to give us an insight in the realism of gen-
erated samples and how suited such data is for augmenting EEG data in a classi-
fication task.
The results show that generated samples exhibit features characteristic of real
data. However, this was only seen for certain models with the majority generating
low quality samples. Furthermore, the classification experiments yield conflict-
ing results making the impact of generative DA uncertain. The finding from this
thesis show that it is possible to generate realistic EEG data, but that this is less
useful as DA compared to other works, despite sample quality being similar.Dementia is a condition with many causes that affects the cognitive domain, dis-
rupting the ability to function in daily living. Alzheimer’s disease (AD) is the most
common cause of dementia and its preceding state, mild cognitive impairment
(MCI). In both cases, the brain activity is affected and differences between healthy
adults and those afflicted with MCI or dementia can be detected using electroen-
cephalography (EEG). However, analyzing the resulting electroencephalogram
(EEG) requires expert knowledge and experience. For this reason, automatic
analysis has been applied to EEG data to lessen the burden on trained profession-
als. Driven by recent releases of publicly available EEG datasets, deep learning
(DL) methods have become more popular and there has been success in differ-
ent fields, including classification. However, the available datasets are usually of
limited size or suffer from class imbalance.
In this thesis we explore the use of generative diffusion models and utilize
synthetic data as a form of data augmentation (DA) to examine the effect of both
expanding and balancing the classes in a public dementia EEG dataset. We test
different diffusion models including well established models such as denoising
diffusion probabilistic models (DDPM) and latent diffusion models (LDM). We
also design and perform experiments to give us an insight in the realism of gen-
erated samples and how suited such data is for augmenting EEG data in a classi-
fication task.
The results show that generated samples exhibit features characteristic of real
data. However, this was only seen for certain models with the majority generating
low quality samples. Furthermore, the classification experiments yield conflict-
ing results making the impact of generative DA uncertain. The finding from this
thesis show that it is possible to generate realistic EEG data, but that this is less
useful as DA compared to other works, despite sample quality being similar
Resource Orchestration in 5G-MEC Systems
Full text linkPhD thesis in Information technologyThe advancement of Fifth Generation (5G) networks enables high-speed, massive connectivity and low-latency communication, driving innovation in applications such as autonomous driving, smart cities, and more. However, several challenges must be addressed, including the limitation of latency of the distance server, capacity distribution, and network congestion. Multi-access Edge Computing (MEC) mitigates these issues by bringing computation closer to the user, enhancing response times and overall network efficiency.
Resource orchestration in 5G MEC is a critical challenge due to the heterogeneity of resources, diverse service requirements, and dynamic network conditions. This thesis addresses these challenges by formulating both offline and online resource orchestration problems, developing the related solutions, and evaluating their performance. The research is guided by three main questions: (1) Which are the state of the art and broader impacts of resource orchestration in 5G-MEC systems? (2) How can relevant and unaddressed offline problems in 5G-MEC resource orchestration be formulated and solved? (3) How can relevant and unaddressed online problems in 5G-MEC resource orchestration be formulated and solved?
To answer the first question, the thesis provides a comprehensive survey of the state of the art in 5G-MEC resource orchestration, identifying key challenges. It also explores the broader impacts of resource efficient 5GMEC systems in smart city applications, highlighting their potential to drive economic growth, enhance quality of life, and promote sustainability. This foundational work sets the stage for addressing specific offline and online challenges. To address the second question about offline problems, the thesis focuses on well-defined but complex challenges, such as network slice allocation and multi-objective traffic path computation. This thesis proposes a mathematical optimization framework to minimize resource costs while meeting service requirements, demonstrating cost reductions of up to 20% compared to traditional methods. Additionally, this thesis develops an algorithm based on Reinforcement Learning (RL) to address multi-objective optimization, which is useful for 5G services that have diverse service requirements, showing the potential of computing traffic paths by using RL in 5G-MEC systems. These contributions provide scalable and efficient solutions to offline resource orchestration problems.
To answer the third question about online problems, the thesis addresses dynamic challenges such as service migration. A Deep Q-Network (DQN)-based algorithm is proposed to maximize service availability in the presence of failures of the MEC hosts, maintaining availability levels of 94%–96% with fewer migrations. For service relocation, this work develops a heuristic algorithm to balance service performance, availability, and user location privacy, demonstrating robust performance across diverse scenarios. These solutions address the need for adaptive and real-time decision-making in online resource orchestration.
By systematically addressing these research questions, this thesis advances the state of the art in 5G-MEC resource orchestration, providing both theoretical frameworks and practical algorithms to improve Resource efficiency and service performance in next-generation networks