The proliferation of new media has created significant changes to how students interact with cultural diversity in the twenty-first century. No longer limited to interactions available only in their local contexts, many students connect online to highly diverse places, peoples, and texts. This dissertation describes these interactions as twenty-first century multicultural learning. Although multiculturalism is typically defined locally according to a nation’s populace, this empirical case considers the increased salience of transnationalism and the growing importance of linguistic and cultural flexibility in an era of globalization. While recognizing seminal multicultural education scholarship that is foundational to this study’s evolution (Banks, 2009; 2013; Hernández-Sheets, 2003; Ladson-Billings, 1994; Lee, 1992), this study proffers a reexamination of multicultural learning in a rapidly changing communication landscape.

This study features DramaCrazy.net, a free website on which people posted, watched, and discussed Asian dramas. Data collection focused on DramaCrazy’s Korean dramas forum and spanned two months of almost daily visits to the site from May through June of 2013. Data included peripheral observations, Web traffic estimation data, over 200 screenshots, and eight printed discussion threads. Content analysis of publicly available data (Brett, 2010; Farnel, 2015; Hine, 2014; Hughes, 2012; Thelwall, 2008) allowed for close analysis of “archives of naturally occurring, non-reactive, online interactions” (Hewson & Laurent, 2012, p. 183). The forum’s majority was adolescents who lived in diverse locations outside of Korea. A central finding was that these youth pursued multicultural learning and identity construction through their reading and composition of digitally mediated texts. They generated knowledge about Korean language, history, and contemporary culture, as well as information about their local contexts. Their literacy and language practices also involved digitally mediated self-representations that signified belonging to diverse locations and multiple affiliations.

Through a semiotic approach to examining meaning-making practices wherein “choice is the basis of and the expression of meaning” (Kress and Selander, 2012, p. 267), I discuss these adolescents’ literacy and language choices as meaningful to their understanding of themselves and others. Their participation in the forum involved imagining, analyzing, articulating, and remixing their cultural identities, social worlds, and connections to global others. They composed transcultural self-representations that disrupt a fixed categorical logic of identity that persists in school policies and practices (Kim, G.M., 2016a; 2016b). These complex cultural identities highlight a need for more student-centered definitions of culture in education research and practice. As culturally relevant (Ladson-Billings, 1994) and culturally responsive (Gay, 2000) pedagogies have become the established approach of many teacher education programs, these asset-based pedagogies require updating to include ways that cultures are learned and lived by students through digitally mediated, transnational, literate practices. In a spirit similar to Paris and Alim’s (2014) “loving critique forward” this study aims to advance multicultural education research in general, and literacy and language studies in particular, through its documentation of transcultural digital literacies and languages.

The loss of hair cells (HCs) and spiral ganglion neurons (SGNs) in the inner ear can lead to irreversible hearing loss in mammals. Because these cells are few in number and difficult to access, studies on their development, degeneration, and reprogramming capabilities in response to injury have been limited. Previous studies have analyzed the gene expression profiles of HCs and SGNs in mice through bulk and single-cell RNA sequencing, with some identifying transcription factor (TF) genes, such as Atoh1, Pou4f3, Gfi1, Neurog1, and Neurod1 as significant for HC and SGN development and maturation. However, it is not known whether these findings apply to the human system. Thus, a reliable and accurate in vitro model of human HCs and SGNs has yet to be generated to study their development and injury responses. In this study, we overexpressed two unique sets of TF gene cassettes in IMR90.4 cells with an endogenous POU4F3-p2A-eGFP reporter: (1) TetO-ATOH1-p2A-POU4F3-p2A-GFI1 (APG1) and (2) TetO-NEUROG1-p2A-ATOH1-p2A-NEUROD1 (NAN1) for generation of HC- and SGN-like cells, respectively. These modified hiPSCs were differentiated and collected for bulk RNA-sequencing after 6 and 14 days. Sequencing analysis revealed that after six days, APG1 induced cell gene expression was enriched for auditory processes, but this enrichment was less significant by day 14. Though day 6 NAN1 induced cells appear more neuronal than day 6 APG1 cells, possibly leading to an SGN-like cell fate, they do not take on a definitive cell type identity, instead expressing some HC- and neuronal genes at days 6 and 14.

Background: Prevention of the proliferation and use of nuclear weapons is of urgent importance to public health. However, persisting trends demonstrate a lack of both public engagement and consistent public discourse regarding issues related to nuclear policy and nuclear weapons. These factors suggest nuclear risk, once a dominant source of public concern, has faded from view and may not be well understood by the general public. This study contemplates how we can make a nuclear risk issue that is abstract into something that is more concrete and accessible for the general public, ultimately mobilizing individuals to take action. This dissertation examines how a storyline in a popular television show can be used to educate and persuade viewers about nuclear risk. Although entertainment narratives have historically been used to depict traditional public health issues, such as cancer prevention and reproductive health, a growing base of research evidence on narrative persuasion suggests fictional narratives can provide an effective approach for promoting awareness and communication about individual, community and collective health risks, in this case nuclear weapons. However, most research examining the impact of entertainment narratives has largely overlooked the complex role of viewer characteristics, such as political ideologies and core beliefs, in narrative persuasion mechanisms, despite their theoretical importance. Methods: This study explores the impact of a nuclear risk storyline on the comprehension and acceptance of information conveyed via the popular television show Madam Secretary. This study implements a unique mixed-methods approach using a viewer pre/post survey (N=295) and social media data (N=3629). First, regular viewers of the show were surveyed before and after the finale episode with the nuclear risk storyline, with a total of 295 participants completing both surveys. Multivariate regression was used to analyze effects of narrative exposure on viewers’ knowledge, attitudes, and behaviors. Latent profile analysis and moderation were used to examine differential impact among subgroups of viewers defined by their political ideologies and beliefs. Among viewers of the episode, mechanisms of narrative persuasion were subsequently examined to determine whether narrative engagement, perceived realism, or emotion mediated the relationship between viewers’ political ideologies and beliefs (represented by latent viewer subgroups) and their nuclear knowledge, attitudes, and behaviors. Tweets from the social media platform Twitter were also collected to analyze tweets related to the show to expand on quantitative findings. Media content analysis of tweets aimed to examine viewer engagement through social media as a mechanism to explain audience impact. Results: Findings suggest that fictional television narratives can help educate viewers about nuclear risk, but did not affect all viewers equally. Latent profile analysis identified two distinctive latent viewer subgroups: one group that is more conservative, hawkish, less anti-nuclear and another group that is more liberal, dovish, and more anti-nuclear. Latent class moderation revealed that only those who were more liberal, dovish, and more anti-nuclear experienced an increase in nuclear knowledge. However, mediation analysis surprisingly found that viewers who were more conservative, hawkish, and less anti-nuclear experienced a more positive emotional response to the storyline, but learned less than the more liberal, dovish, and more anti-nuclear viewers. Thematic analysis of tweets revealed that most tweets expressed enjoyment of the show and its characters, followed by narrative presence and perceived realism. Viewers were not always reacting to the nuclear risk storyline, but content experts and organizations played an important role in clarifying and reinforcing the intended message of the storyline. Conclusion: Results suggested the potential for selectivity, where viewer subgroups engage or dismiss parts of the narrative that do not align with their beliefs. Tweets also revealed additional barriers to viewers’ narrative engagement, such as competing storylines. These findings shed light on unique challenges in communicating health messages that are potentially political controversial and provide a more nuanced understanding of narrative engagement that have implications that can advance the field of narrative persuasion.

Idiopathic pulmonary fibrosis (IPF) is one type of interstitial lung disease (ILD) of unknown causes. High-resolution computed tomography (HRCT) scans play a crucial role in distinguishing IPF from non-IPF among subjects with ILD. This radiological evaluation is an important yet difficult task. In clinical practice, making a correct and reliable IPF diagnosis is critical to ensure patients with different causes of pulmonary fibrosis be treated appropriately and patients with IPF be assessed for novel therapies and lung transplantation. Therefore, this dissertation aims to build an automated IPF diagnosis system for ILD subjects, using volumetric and non-contrast chest HRCT scans.

Supervised learning methods are a type of machine learning methods that require labels of training samples as ground truth to learn a mapping function between input and output labels. Depending on the type of labels provided, if only coarse-level labels are available, rather than fine-scale labels, this task is called a weakly supervised task. For our example, acquiring fine-scale information of certain radiological patterns can be helpful for building the diagnostic system, but fine-scale labels are expensive to obtain. On the other hand, coarse-level labels are usually easier to acquire, such as CT scan-level information. Since we only have labels at a CT scan level, our problem is a weakly supervised task. To tackle this challenge, this dissertation leverages domain knowledge acquired from previous studies, including IPF progression and quantification information, to provide more efficient, reliable, and explainable diagnostic support.

In project I, we used 2D deep learning models with IPF progression information and optimal design criterion to weigh HRCT samples differently. In project II, 3D deep learning models with multi-scale attention models were used with IPF quantification maps to achieve good model accuracy and explainability. Furthermore, we evaluated the robustness of these developed models under a different set of HRCT parameters, using paired HRCT scans. These proposed methodologies in projects I and II can be applied to other weakly supervised tasks, where domain knowledge is available. The method used in the robustness tests can be applied to evaluate model performance if paired medical images are available.

There are many challenging optimization problems in the health sciences. Problems in health sciences are increasingly complex, and frequently the most advanced optimization techniques are required to tackle them. Researchers thus need various types of flexible optimization tools that are easily accessible and efficient. In this dissertation, we utilize a stochastic optimization technique called particle swarm optimization (PSO) and demonstrate its usefulness and flexibility using two applications in the biomedical field. For the first application, we propose a sparse grid hybridized PSO (SGPSO) algorithm to find different types of optimal or highly efficient designs for various longitudinal models. In particular, we consider non-linear mixed effects models useful in pharmacokinetic/pharmacodynamic studies and show SGPSO is a powerful tool for finding optimal or efficient designs that were previously thought to be intractable. For the second application, we propose a random forest hybridized quantum PSO algorithm for predicting disease progression of idiopathic pulmonary fibrosis (IPF) using quantitative information on high-resolution computed tomography (HRCT) imaging. IPF is a fatal type of lung disease with unpredictable functional progression at the time of diagnosis. We leverage single time point HRCT scans to predict the 6 months to 1 year follow-up status of IPF subjects. Results show that the two hybridized PSO approaches tackle important biomedical optimization problems effectively, and since the proposed methodology is not problem specific, there is potential for further applications to solve other biomedical optimization problems.

Purpose

Methods

Results

Conclusions