1722 research outputs found
Sort by
Exploring the Effect of Demographics Inclusion on Subject-independent Emotion Recognition
No full textElectroencephalography (EEG) can capture electrical activity associated with human emotion processing from the scalp. The electrical activity can be processed using deep learning models to predict emotional states. Two approaches can be employed to develop these deep learning models: subject-dependent and subject-independent. The subject-independent approach is more practical as it trains the model on data from some individuals and tests it on entirely different individuals, ensuring it generalizes well to new users. However, because of the high variability of EEG across individuals, the subject-independent approach tends to yield low performance. Recent studies suggest incorporating demographic information along with EEG signals is one way to overcome this issue. By using the subject-independent approach, this study investigates different demographics factors such as age, biological sex and cultural factors impact emotion prediction. Moreover, this thesis delineates the development of a deep learning models dedicated to emotion recognition on five different datasets. To find the impact of age and biological sex a logistic regression model was used to correlate the output of a deep learning model with subjects’ age and sex, thereby evaluating whether these factors impact emotion prediction. Our analysis indicates that the ‘sex’ variable significantly influenced the predictions of the deep learning model in three out of five emotions, whereas ‘age’ does not have any effect. These findings suggest that sex is a factor that needs to be considered when designing EEG-based emotion recognition models. Furthermore, attention network layers were used to identify brain areas more involved in predicting emotions. Additionally, an odds ratio analysis was conducted using logistic regression to evaluate the impact of sex on emotion prediction. Our findings reveal that cortical activation patterns elicited by emotional audio-visual stimuli differ between females and males, with females showing more neural activation in the left hemisphere and males showing more in the right hemisphere. Moreover, when the output probabilities of the deep learning models are further postprocessing with the subject’s sex, the odds of correctly predicting emotions increase. These findings suggest that sex differences can lead to more robust subject-independent emotion recognition models. Additionally, this study also investigates how cultural factors impact emotion prediction. Specifically, we used a stacking model that combines deep learning with multinomial logistic regression to predict positive, neutral, and negative emotions among 15 Chinese, 8 French, and 8 German subjects. Our approach achieved accuracies of 77.3% for Chinese subjects, 73% for French subjects, and 65% for German subjects, which are comparable to or exceed accuracies reported by previous studies. Our approach highlighted that incorporating cultural information increases the likelihood of predicting positive emotions for Chinese participants and negative emotions for Europeans. Moreover, French and German subjects exhibited similar neural patterns across all emotions, suggesting a more common cultural sharing between those subjects. Overall, our findings emphasize the importance of integrating demographics information considerations into emotion recognition models. This inclusion not only improves emotion prediction accuracy for subject-independent approaches but also promotes inclusivity and ethical practices in emotion recognition systems. Which could lead to more robust subject- independent models with potential applications in areas such as healthcare, education, and marketing.Master of Science in Applied Computer Scienc
Examination of customized questioned digital documents
No full textWith the increasing trend of digitization of business processes and personal communication across the globe, digital documents of intrinsic value continue to be created. Whereas the questioned document examination (QDE) field of forensic science deals with the examination of “physical” documents potentially disputed in a court of law, there are no developed approaches for handling questioned digital documents (QDDs). Although techniques that address related problems such as identifying document types and image forensics exist, concrete strategies for analyzing questioned “digital” documents still need to be developed. This paper focuses on developing methods to examine QDDs that are customized from a database, due to the versatile use of customized documents in many areas. As a basis for our approach, we make the case for the need to develop analysis techniques for a digital counterpart of QDE which we term Questioned Digital Document Examination (QDDE). We posit that there is a benefit in considering digital aspects of forensic science disciplines where the questions answered by the discipline are clear, from a digital perspective. The paper describes some of the aspects that can be considered in the domain of question digital document examination. In designing methods for QDDE, we discuss the process of document recreation and describe the feasibility of our recreation process in different scenarios. Our experiments show that an alternative approach of considering digital aspects from a well-defined physical domain is worthwhile. It also supports the practical application of our approach in examining documents customized from a database.University of Winnipeg, Grant/Award Number: 16673, 16792 and 19685.https://onlinelibrary.wiley.com/doi/10.1111/1556-4029.1570
Plasma microRNA Profiles of Myotis lucifugus from a White-Nose Syndrome-affected Population
No full textWhite-nose syndrome (WNS), caused by the fungus Pseudogymnoascus destructans (Pd), has devastated bat populations across North America by disrupting torpor, accelerating fat depletion, and causing severe winter mortality. Surviving populations of little brown bats (Myotis lucifugus) exhibit altered fat storage and adaptive physiological responses, suggesting potential mechanisms for WNS resistance or tolerance. MicroRNAs (miRNAs) are small, non-coding RNA molecules regulating gene expression that play critical roles in metabolic and immune pathways essential for hibernation physiology and pathogen defense. My thesis integrates insilico analyses and experimental validation to evaluate the role of miRNAs in hibernation physiology to establish a novel, non-lethal method for monitoring bat health. Using DIANA miRPath and a targeted literature review, I identified four miRNAs (miR-543, miR-27a, miR-92b, and miR-328) implicated in metabolic and immune pathways relevant to WNS, including lipogenesis, insulin signaling, and FOXO-mediated stress response. I quantified the presence and seasonal expression patterns of selected miRNAs using reverse transcription quantitative real-time PCR in plasma samples collected from WNS-affected bats during fall pre-hibernation and spring emergence. miR-27a-5p and miR-92b-5p showed increased expression in spring compared to fall, and miR-27a-5p correlated positively with Pd fungal load, suggesting its potential as a biomarker for WNS severity. miR-26a-5p was consistently stable across seasons and conditions and was used as a robust endogenous control for plasma-based miRNA studies. This study is the first to demonstrate stable detection and seasonal variation of circulating miRNAs in plasma from free-ranging little brown bats, and one of only a handful to quantify plasma miRNA levels in any of the nearly 1500 bat species, establishing a novel, non-lethal method for monitoring bat health. Future studies should validate gene targets and assess how miRNA expression varies with host physiological state and hibernation conditions. Importantly, this approach could guide targeted management interventions by enabling early identification of vulnerable populations or individuals. Ultimately, the ability to monitor bat health non-lethally, using plasma miRNAs, offers significant potential to enhance wildlife disease surveillance, guide conservation strategies, and contribute to the broader effort of mitigating WNS impacts across North America.Natural Sciences and Engineering Research Council (NSERC, Canada) Discovery Grant to Dr. Craig Willis; Research Manitoba MSc Studentship; University of Winnipeg Graduate Studies ScholarshipMaster of Science in Bioscience, Technology, and Public Polic
Gene Expression Strategies for Developing Blood-Based Screening Diagnostics for Colorectal Cancer
No full textThe N-myristoyltransferase isozymes, NMT1 and NMT2, have attracted growing interest in disease research, especially in the context of cancer. They are responsible for catalyzing the lipidic modification of N-terminal glycine and lysine residues of target proteins, altering their function, localization, and molecular interactions. Recent studies, including those from our laboratory have established NMT1 and NMT2 as key molecules in oncogenesis. In fact, our previous work has demonstrated the overexpression of NMT2 in the peripheral blood mononuclear cells (PBMCs) of colorectal cancer (CRC) patients and individuals with adenomatous polyps (AP). CRC is classified as one of the most treatable cancers, however, it remains a leading cause of cancer related deaths. Without effective screening tests, CRC is often able to progress into advanced stages. While the stool-based tests and colonoscopy remain the primary methods for CRC detection, they often have low adherence rates due to inconvenience and the invasiveness of the colonoscopy. Therefore, a convenient, more accurate, and cost-effective screening test for CRC holds the potential to improve both the quality of life and survival outcomes for patients. Building on our previous studies focused on developing an NMT2-based blood test, this study aimed to assess the gene expression of the NMT isozymes and their upstream target, MetAP2, to build upon previous IHC findings. To ensure accurate gene expression normalization, I validated four candidate reference genes in PBMCs collected from individuals with various colorectal pathologies and no evidence of disease. Using these reference genes, I assessed expression differences among pathophysiological groups and observed meaningful group differences based on effect size. Importantly, the choice of normalization strategy influenced the results, highlighting the importance of accurate reference gene selection.Canadian Cancer SocietyMaster of Science in Bioscience, Technology and Public Polic
Wild Ungulate Detections Using RPAS and Satellite Imagery in Manitoba
No full textEffective wildlife monitoring is essential for the sustainable management of animal populations and their habitats, given the ecological and societal significance of wildlife. Traditional aerial surveys using helicopters and fixed wing aircraft remain the predominant method for tracking wild ungulates. However, they present certain limitations due to their high operational cost, safety risks to crews and animal behavioral impacts. This thesis investigates the potential of remotely piloted aircraft systems (RPAS), commonly known as drones, and satellite imagery as alternative technologies for wildlife monitoring in Manitoba. To evaluate the feasibility of these methods, satellite and RPAS imagery were collected across four Game Hunting Areas (GHAs). GHAs are designated areas used by the province of Manitoba to manage wildlife populations and designate human hunting activities. Satellite imagery was manually reviewed for animal detection, while RPAS imagery was analyzed using thermal thresholding techniques. This study successfully detected farmed cattle in satellite imagery and deer in RPAS imagery. While satellite imagery reliably detected animal groups, RPAS imagery proved more effective overall by enabling the identification of individual animals with greater accuracy and detail. Given the substantial volume of data generated, further advancements in automation for wildlife detection and enumeration are necessary to enhance efficiency and scalability. This research contributes to the ongoing development of innovative monitoring solutions, offering insights into the integration of remote sensing technologies for improved wildlife conservation and management strategies.The Fish and Wildlife Enhancement Fund, Research ManitobaMaster of Science in Environmental and Social Chang
The Blameless and Blameworthy: Missing White and Indigenous Women and Girls' Social Construction on Winnipeg Police Service's Facebook Page
No full textRecently, police agencies have harnessed social media platforms, like Facebook, to communicate with the public regarding missing persons cases. I argue that the Winnipeg Police Service (WPS) is mostly absent from the social construction dynamics of missing women and girls. Instead, missing women and girls are socially constructed primarily through the comments and claims of Facebook users who draw on racialized stereotypes to imply these females’ responsibility and blame. Applying Valverde’s (2006) social semiotic template, I analyzed a purposive sample of 20 WPS Facebook posts about missing women and girls from 2019 to 2023 focusing on the selection of images, descriptive text, user comments and reactions. Results revealed that missing Indigenous women were constructed as most blameworthy for their disappearances, while missing Indigenous girls were constructed as less blameworthy, but not without some level of responsibility for their situation. In contrast, missing White women and girls were socially constructed as blameless ideal missing persons worthy of rescue. I conclude by reflecting on the theoretical and methodological implications of my study and offering directions for WPS social media policies to prevent the continued promotion of racial stereotyping and victim blaming.Master of Arts in Criminal JusticeMaster of Arts in Criminal Justic
Barriers to Systemic Therapy Services Among Transgender and Nonbinary Adults in Canada
No full textSystemic therapies, including relational, couple, marriage, and family therapy services, play a key role in supporting transgender and nonbinary people who experience mental health and relational challenges. However, limited research studies have examined this population's barriers to accessing these therapy services. This study aimed to explore these barriers to enhance the cultural competence of relational and family therapists. This study involved 12 TNB adult participants in Canada. Regarding methodology, the semi-structured interviews were conducted via Zoom and recorded, and the interpretative phenomenological analysis was applied. The results included the following: (1) Challenges in finding competent and affirming therapists; (2) Geographic limitations; (3) Financial concerns; and (4) Systemic and institutional barriers to accessing RCMFT services. The results can be shared in teaching, supervision, workshops, and conference presentations to improve cultural competence and reduce the barriers."This study was funded by the University of Winnipeg Major Research Grant (#20642) and the University of Winnipeg Research Start Up Grant."https://onlinelibrary.wiley.com/doi/10.1111/jmft.7006
Use of Single and Blended Soil Chemical Amendments to Reduce Phosphorus Loss from Soil
No full textExcessive phosphorus (P) from agricultural soils poses a significant environmental threat due to its contribution to eutrophication in water bodies. Soil amendments have been proposed to reduce soil P solubility and decrease losses in runoff, but their effectiveness, especially as blended amendments, and underlying mechanisms remain underexplored. This thesis investigated the effects of single and blended applications of alum (AlK(SO4)2·12H₂O), ferric chloride (FeCl3), gypsum (CaSO4·2H2O), and magnesium sulphate (MgSO4) on soil P dynamics and transformations in six agricultural soils from Manitoba, and separately examined the effects of single and blended gypsum and ferric chloride in a simplified model soil (artificial soil) system composed of sand, silt, clay, humic acid, and 1000 mg kg⁻¹ total P. In both experiments, soils were incubated for up to 84 days at 22 ±1°C with periodic measurements of water-extractable P (WEP) concentrations and Olsen P concentrations to evaluate potential P loss and available P. Sequential P fractionation was used in both studies after 84 days to identify shifts in P pools. Scanning electron microscopy with energy-dispersive X-ray spectroscopy (SEM-EDX) was conducted only in the natural soil study to identify elemental co-localization. In the natural soil study, all amendments significantly reduced WEP concentrations relative to unamended controls except in soil 1. In soil 1, only a few treatments were effective in significantly reduced the WEP concentrations on one or more sampling days. The blended treatments, particularly gypsum or magnesium sulphate combined with ferric chloride, produced the greatest reductions (up to 85%). Decreases in Olsen P were comparatively modest (average of 9.5%), indicating that treatments reduced labile P without substantially lowering agronomically available P. Sequential fractionation revealed that amendments increased recalcitrant P forms and decreased NaHCO3-extracted P and NH4Ac-extracted P. In the controlled model soil experiment, all amendments significantly reduced WEP concentrations compared to the unamended control, with the gypsum + ferric chloride blend showing the most substantial decrease (47.6–58.9%). Olsen P concentrations initially increased in all amended treatments, but by 84 days, only soils amended with ferric chloride or its blend maintained higher Olsen P than the control. Sequential P fractionation revealed a shift from labile to more stable P pools, indicating increased P retention in the soil matrix. Collectively, these findings demonstrate that blended amendments, especially combinations with ferric chloride, enhance P retention by promoting P stabilization in agricultural and model soils.Master of Science in Environmental and Social Chang
How Can Expropriation Be Used To Solve the Housing Crisis?
No full textThis paper considers application of the law of taking, called expropriation in Canada, to the “Housing Crisis,” as it is often characterized. It examines four options: The royal prerogative of taking, procedural reform, compensation reform, and public-sector directed surges in supply. Drawing on recent events (a rash of vacant house fires) in Winnipeg in particular, it is submitted that procedural reform is the most promising option, but more research is necessary to determine how best to do this, balancing speed with procedural fairness.This publication was funded by the Institute of Urban Studies
Dataset Optimization Using Image Processing
No full textThe dataset plays a vital role in model training. It is commonly believed that larger datasets improve accuracy. However, if we cannot ensure the quality of the data, it not only consumes resources but can also lead to over-fitting. To address this issue, this thesis proposes eight methods on two datasets, which range from image similarity algorithms to clustering CNN features, to create the smallest possible subsets of data. We evaluated different scenarios for each method and compared the results with those obtained using the corresponding full dataset and random removal, determining which data should be retained and which discarded. The empirically observed generalization gap resulting from dataset pruning is substantially consistent with our theoretical expectations. The proposed method can reduce data from both datasets by 20% with almost no loss in accuracy. In fact, a 2.3% increase in accuracy is observed for dataset A even with the 20% removal. The method effectively reduces the smaller dataset by 60% and the larger dataset by 40%, while maintaining a drop in accuracy of less than 2%. Additionally, if a decrease in test accuracy of 4.6% for the smaller dataset and 4.8% for the larger dataset is deemed acceptable, it is possible to reduce the data from both datasets by 70%.Natural Sciences and Engineering Research Council of Canada - NSERC (Application No: RGPIN-2024-05056) and Research Manitoba(Application No: 6189)Master of Science in Applied Computer Science and Societ