14277 research outputs found
Sort by
Population trends in an Atlantic puffin (Fratercula arctica) colony determined from a ~500-year sediment core
No full textThe global population of Atlantic puffin (Fratercula arctica) is declining. This decline has been linked to numerous anthropogenic impacts. However, untangling these impacts from fluctuations due to colony dynamics is complicated by the lack of historical population data.
Paleoecological reconstructions based on chemical and biological signatures in sediment cores taken from seabird colonies have been successfully used to infer historical trends in nesting seabird numbers for several seabird species, but not for Atlantic puffin. This research uses diatom assemblages, stable isotopes, metal(loid)s, sterols and stanols, and chlorophyll a to reconstruct an Atlantic puffin colony located in the western North Atlantic near Fogo Island, Newfoundland. The analysis shows that numbers of Atlantic Puffins nesting at this colony remained relatively stable from ~1450 CE, and began increasing in ~1966 CE, with a sharp and rapid population increase suggested from the 1990’s to present. The five proxies used had good congruence with one another. This corresponds with other evidence that Atlantic puffin populations in Newfoundland are stable or increasing, in contrast to colonies elsewhere in their range. The apparent increase at this colony also coincided with several significant changes in conditions, including the collapse of the cod fishing industry in Newfoundland.Includes bibliographical references (pages 64-81
Nobody is lonely amongst the stars: sense of religious community in the Star Trek fandom
No full textReligion in modernity has evolved, and media franchises may now form the basis for experiencing religious belonging and community in the changing nature of secular Western society. This thesis seeks to explore how Star Trek fandom, as a “cultural religion,” presents a similar sense of community as traditional religions and the impact it has on the sense of loneliness and mental health of participants. The research consisted of eighty-six completed questionnaires, four unstructured online interviews, and qualitative participant observation at conventions and in online fan forums, while also referencing scholarly works. The resulting portrait of Star Trek fandom does resemble traditional religion: the fandom supplies a unique sense of belonging; it provides comfort through the narratives in the series, a mission, and parasocial relationships. Fans often look towards the series and characters for escapism, understanding/speaking on serious issues, and guidance. Members experience increased meaning, sense of identity, confidence, self-awareness, social circles, coping ability, and ability to form relationships outside fandom, which all have a positive effect on loneliness and the fan’s mental health. This research may benefit both the study of religion and its changing landscape and provide insight into the nature of how fandom involvement can serve as a community-seeking and mental health-bolstering endeavour.Includes bibliographical references (pages 113-121
Storytelling as a tool to foster resilience: an arts-based study with postsecondary international students in the Newfoundland context
No full textThis paper is a study of resilience regarding international students studying at Memorial University of Newfoundland, living in the St. John’s metro area. Using visual-based narrative inquiry, a form of arts-based research, participants were invited to share their resilience stories through discussions and book making in virtual group settings (Kim, 2019, p. 143). The benefits of arts-based research methods were explored: it equalized the power dynamic between the researcher and the researched, employed aesthetics to deepen the understanding of human experiences, and widened the readership and audience for research. Through this study, it is noted that resilience can be fostered and can vary based on the time, contexts, and challenges faced. The study concludes that arts-based practices can be used as a mental health tool to help foster resilience. Other areas as well as forms of support that Memorial University can employ and expand on to support the mental health and well-being of international students were also discussed, such as changing the university’s culture to that of learning for learning’s sake, involving all to become advocates through university-wide intercultural training and incorporating intercultural learning as well as mental health support within classroom settings, providing reflective spaces to further support staff and students. The author of this study intentionally explained her social contexts and subjective stance to increase transparency. In doing so, she hoped to provide a different perspective to this study of resilience: as a person of colour, a student support staff member at the University and a former newcomer.Includes bibliographical references (pages 159-182
Exploring multi-modal behavioural taxis in leach's storm-petrels (Hydrobates leucorhous): Interactive effects of odour and anthropogenic light
No full textWith the increase of urban development, wildlife are increasingly exposed to sensory pollutants. Marine bird populations are declining, partly due to sensory traps, which can distract them from natural behaviours, increasing subsequent risk. Here, I explore the influence of a potential multisensory trap on Leach’s Storm-Petrels (Hydrobates leucorhous), a seabird that exhibits positive phototaxis towards nocturnal light, often resulting in them becoming stranded. I investigated the stranding of storm-petrels at an illuminated and odorous seafood processing plant near their largest colony and performed a full factorial experiment at a major colony to determine how counts of birds are affected when presented with light and scent cues individually and combined. Crab processing odours did not influence strandings at the plant nor behaviour at the colony. Nightly strandings were highest at the plant when scent and light were presented together, but no significant differences in strandings were observed when processing odours were present versus not present, and lights were on. Interestingly, storm-petrel strandings increased when light was presented at the plant, but counts were significantly lower at the colony when light was present. These findings demonstrate the importance of incorporating multisensory information into experimental designs and how multisensory interactions may increase risk
Multi-class motor imagery EEG classification for brain-computer interfaces: a comparative study of traditional and deep learning models
No full textBrain–computer interfaces (BCIs) have emerged as promising assistive technologies, enabling users to control external devices through brain activity alone. Among the various paradigms in BCI research, motor imagery (MI) stands out due to its intuitive control mechanism and compatibility with non-invasive electroencephalography (EEG) systems. However, the practical utility of MI-BCIs remains constrained by their limited number of distinguishable control commands and their reliance on subject-specific calibration. To address these challenges, this thesis investigates whether deep learning algorithms—particularly convolutional and hybrid architectures—can improve classification accuracy in high-class-count MI-BCI systems and enhance generalizability across subjects.
This study compares the performance of traditional machine learning algorithms (e.g., logistic regression, LDA, random forest, GBDT) with modern deep learning models (EEGNet, TCNet, ATCNet) across subject-specific and subject-independent scenarios using three EEG datasets: the public BCI-IV-2a dataset and two original datasets incorporating novel MI tasks, including kinesthetic singing imagery (SI) and dual-task combinations. Classification problems ranged from two to eight classes, allowing for systematic evaluation of scalability and robustness.
The results demonstrate that traditional models, such as logistic regression and random forest, performed strongly in subject-specific, low-complexity tasks—for instance, logistic regression achieved 84% accuracy in 2-class classification. However, their effectiveness declined significantly in high-class-count and subject-independent settings, with accuracies dropping to around 25% in 8-class scenarios. In contrast, the hybrid ATCNet architecture consistently outperformed other models in more complex tasks, delivering robust results across datasets—for example, achieving 85% subject-independent accuracy on the BCI-IV-2a dataset and 61% subject-specific accuracy on the more challenging 6-class DI-data (dual imagery dataset). Notably, across all experiments, subject-specific classifiers yielded higher accuracy than their subject-independent counterparts, particularly as task complexity increased. These findings underscore the challenge of inter-subject variability and highlight the need for larger, more diverse datasets to improve the generalizability and scalability of MI-BCI systems for real-world applications.
This work affirms the promise of hybrid deep learning models in enhancing the scalability and user-independence of MI-BCI systems. Future research should focus on expanding training datasets and refining model architectures to achieve reliable, calibration-free BCI performance suitable for clinical and consumer applications
Modelling of ocean circulation in the Newfoundland Basin
No full textThe average temperature of the Earth's surface has increased significantly since preindustrial
levels and continues to rise. This global warming is caused by the increased
level of greenhouse gases in the atmosphere, particularly carbon dioxide (CO₂). To
reduce the rate of temperature increase, it is necessary to decrease the amount of
CO₂ in the atmosphere. One way to achieve this is by allowing oceans to absorb more
CO₂ through a process known as ocean alkalinity enhancement (OAE), which occurs
when alkaline particles are added to the ocean surface. The observational tracking
of particle trajectories through the ocean presents significant challenges; therefore,
ocean models are primary tools for predicting particle trajectories. This research
focuses on ocean modelling around the Newfoundland Basin. The Regional Ocean
Modelling System (ROMS) has been implemented for this region. ROMS is a hydrostatic,
free-surface ocean model which uses a terrain-following vertical coordinate.
The continuous model equations of ROMS, along with their numerical implementations,
are described. The model initialization, as well as the definition of surface
forcing and boundary conditions, are presented. The method of particle tracking in
ROMS is described. Preliminary results from model simulations of ocean characteristics,
including ocean temperature, salinity, and sea surface height, for July 2020
are outlined. Particle trajectories after a month of circulation within the Newfoundland
Basin are shown. Future work, including extended model duration and different
particle distribution methods, is then discussed in the context of applications to OAE.Includes bibliographical references (pages 32-33
Unraveling the crucial role of huntingtin in synaptic plasticity and neuronal health in the adult brain
No full textHuntington’s disease (HD) is a monogenic neurodegenerative disease caused by a mutation in the huntingtin (HTT) gene. Many promising therapeutics have entered clinical trials that treat HD by targeting its root cause: mutant HTT (mHTT). These drugs deplete mHTT at the ribonucleic acid (RNA) or protein level; however, many are non-selective and reduce both mHTT and wild-type HTT (wtHTT). Non-pathogenic wtHTT is essential for nervous system development and regulates myriad crucial cellular functions, including axonal transport, transcription and autophagy. As most HD research has focused on the gain-of-function effects of mHTT, the consequences of wtHTT-lowering are not fully understood. Elucidating the consequences of wtHTT loss in the adult brain is essential as HD patients entering clinical trials have 50% wtHTT protein levels that are further reduced by non-selective therapeutics. The aim of this dissertation is to investigate the consequences of wtHTT-lowering in the adult brain. The present thesis begins by reviewing the role of wtHTT in synaptic function, as HD is considered, like many other neurodegenerative diseases, to be a synaptopathy, and synaptic dysfunction both precedes and predicts the onset of HD. This thesis highlights a plethora of studies that suggest a role for wtHTT as a major regulator of synaptic function in the adult brain. From here, the consequences of wtHTT-lowering in vitro and in vivo are examined, mainly focusing on wtHTT depletion in hippocampal neurons, as HD models display severe hippocampal synaptic dysfunction. Results from this dissertation show that wtHTT-lowered primary neurons have altered nuclear morphology and a loss of transcriptional repression. In vivo, 1-2 month conditional deletion of wtHTT results in widespread changes in morphology, extensive neuroinflammation, synaptic plasticity failures in the adult mouse hippocampus, and spatial
learning and memory deficits 6-8 months post-deletion. When compared to the HD-vulnerable striatum, 1-2 month wtHTT knockout in the mouse striatum was not found to impact morphology but did lead to neuroinflammation and reduced intrinsic excitability in a similar manner as what was seen in the hippocampus. This thesis emphasizes the importance of maintaining sufficient neuronal wtHTT levels in the adult brain, an essential consideration for the development of HTT-lowering therapeutics
From critical minerals to critical reclamation: Implementing an anticolonial ethics of reclamation for the Faro Mine, in Tsē Zūl, Dena Kēyeh (unceded Kaska Lands, Yukon, Canada)
No full textAs the Canadian government renews promotions for so-called ‘critical minerals’ extraction
across Northern Canada, local communities are grappling with both the legacies of abandoned
sites and new articulations (or co-optations) of sustainability in the context of green energy. With
this drive for increased mining, the need for nuanced discussions about reclamation are critical.
The Faro Mine, a lead-zinc mine that operated from 1969-1999, is one of the largest reclamation
sites in Canada. Located in Tsē Zūl, on unceded Tū Łídlīni (Ross River) Kaska Dena Land, the
mine inflicted countless harms on the community and Dena Kēyeh (Dena Land). For decades Tū
Łídlīni Dena have been demanding that both mining and reclamation be done differently, in a
way that respects Kaska stewardship and governance. Through participatory action and placebased
research, in partnership with Ross River Dena Council (RRDC), this PhD aims to re-think
reclamation and unearth possibilities for ethical, community-driven approaches to repairing
contaminated landscapes.
Grounded in the direction and priorities of Tū Łídlīni Dena Elders, this research focuses first on
analyzing the ‘infrastructures of theft’ at Faro, including the mineral permitting, welfare state
policies, and water licensing that facilitated the theft of unceded Kaska Land and Water. I then
trace how these historic mechanisms of theft have morphed into the contemporary impact
assessment and regulatory processes guiding reclamation work at Faro. A second key priority for
this community-based research was to build reclamation alternatives based in Kaska knowledge,
drawing on the long history of Tū Łídlīni Dena resistance to the Faro Mine. Therefore, part of
my work included supporting and documenting the implementation of a community-based
revegetation program, centered in healing. While pointing to the root causes of violence and
contamination at Faro, this research simultaneously celebrates all the relationships that have
persisted, that are hard fought for in the face of pervasive racism, colonialism, and extractivism.
This resistance is exemplified in the stories, experiences, community planning, and alternatives
that have been articulated by Tū Łídlīni Dena for decades. These alternatives are what form the
very basis of anti-colonial reclamation, and the imagining of future human-environment
relationships based in Indigenous lands, community, and governance.Includes bibliographical references (pages 416-464
Local government administrative efficiency and change in education outcome in Uganda
No full textThis study set out to investigate local government (LG) administrative efficiency and change in education outcomes in Uganda. The unit of study was the local governments. The study used achievement data from the local government performance assessment for 2022 undertaken by the Office of the Prime Minister. Change in Education outcomes was measured by change in the Primary Leaving Examinations (PLE) pass rate by primary schools in local government. The following variables were used as predictors of education outcomes; human resources practices, financial/budgetary practices, oversight, support supervision, and reporting. Data were analysed using descriptive statistics and binary logistic regressions between the predictors and outcomes variable. Human Resource (HR) practices were measured through hypotheses; adequate number of teachers, substantively recruited critical staff, appraisal of headteachers and LG staff, continuous teacher development as well as appraisal of teachers and LG staff as predictor of change in PLE pass rate. Similarly, Timely submission and communication of UPE capitation grant releases to schools, and finally, schools’ inspections, support to schools to develop improvement plans, support supervision and reporting as predictors of change in PLE pass rate. The results of the study, controlling for household income and minimum infrastructural/facilities standards, suggest that (1) recruitment of adequate number of teachers, substantive recruitment of critical LG staff, appraisal of teachers and LG staff, continuous teacher/professional development, timely submission and communication of UPE capitation grants as well as schools inspection, support supervision and reporting are not predictors of change in education outcomes, measured as change in PLE pass rate by primary schools in LGs. The findings of this study are informative for policy makers in teacher development and the overall improvement of the education outcomes in local governments, with particular focus on those areas that will promote holistic improvement in learning outcomes.Includes bibliographical references (pages 113-125
A deep convolutional network approach with attention mechanism for green sea urchin detection and localization
No full textGreen sea urchin, Strongylocentrotus droebachiensis, exerts considerable influence on
marine benthic habitats in Arctic and sub-Arctic regions, including kelp forests. Additionally, the species’ gonads (roe) are a highly prized delicacy on Asian markets. To
assess and monitor ecological conditions in coastal regions due to sea urchin overgrazing or to establish aquaculture of green sea urchins, computer-assisted autonomous
detection might be desirable. However, the accuracy of underwater identification of
green sea urchins is affected by a number of factors, including low picture quality,
scattering and absorption of light, overlap or occlusion of underwater species, differences in the sizes of the species, and the presence of background objects. In this work,
we present a multi-step process for the autonomous identification of green sea urchins
in natural habitats using an underwater image dataset consisting of 2,400 images.
The process includes augmentation, color correction and enhancement based on the
state-of-the-art YOLOv7 object detector. The results of the experiments demonstrate
that the proposed framework is capable of accurately recognizing green sea urchins
in various underwater scenes and for varying distances between the camera and the
target with a mean Average Precision (mAP) of 83.6%.Includes bibliographical references (pages 60-70