The Spread of Beliefs in Partially Modularized Communities

Many life-influencing social networks are characterized by considerable informational isolation. People within a community are far more likely to share beliefs than people who are part of different communities. The spread of useful information across communities is impeded by echo chambers (far greater connectivity within than between communities) and filter bubbles (more influence of beliefs by connected neighbors within than between communities). We apply the tools of network analysis to organize our understanding of the spread of beliefs across modularized communities and to predict the effect of individual and group parameters on the dynamics and distribution of beliefs. In our Spread of Beliefs in Modularized Communities (SBMC) framework, a stochastic block model generates social networks with variable degrees of modularity, beliefs have different observable utilities, individuals change their beliefs on the basis of summed or average evidence (or intermediate decision rules), and parameterized stochasticity introduces randomness into decisions. SBMC simulations show surprising patterns; for example, increasing out-group connectivity does not always improve group performance, adding randomness to decisions can promote performance, and decision rules that sum rather than average evidence can improve group performance, as measured by the average utility of beliefs that the agents adopt. Overall, the results suggest that intermediate degrees of belief exploration are beneficial for the spread of useful beliefs in a community, and so parameters that pull in opposite directions on an explore-exploit continuum are usefully paired.

Misinformation and Public Health Messaging in the Early Stages of the Mpox Outbreak: Mapping the Twitter Narrative with Deep Learning

BACKGROUND

Shortly after the worst of the COVID-19 pandemic, an outbreak of Mpox introduced another critical public health emergency. Like the COVID-19 pandemic, the Mpox outbreak was characterized by a rising prevalence of public health misinformation on social media through which many US adults receive and engage with news. Digital misinformation continues to challenge the efforts of public health officials in providing accurate and timely information to the public. We examine the evolving topic distributions of social media narratives during the Mpox outbreak to map the tension between rapidly diffusing misinformation and public health communication.

OBJECTIVE

To observe topical themes occuring in a large-scale collection oftweets about Mpox using deep-learning.

METHODS

We leveraged a dataset comprised of all MPox related tweets that were posted between May 7, 2022 and July 23, 2022. We then applied Sentence Bi-directional Encoder from Transformers (S-BERT) to the content of each tweet to generate a representation of its content in high-dimensional vector space where semantically similar tweets will be located closely together. We project the set of tweet embeddings to a two-dimensional map by applying Principal Component Analysis (PCA) and Uniform Manifold Approximation Projection (UMAP). Finally, we group these datapoints into 7 topical clusters using k-means clustering and analyze each cluster to determine its dominant topics. We analyze the prevalence of each cluster over time to evaluate longitudinal thematic changes.

RESULTS

Our deep learning pipeline revealed 7 distinct clusters of content: (1) Cynicism, (2) Exasperation, (3) COVID-19, (4) MSM, (5) Case Reports, (6) Vaccination, (7) WHO. Clusters that largely communicated erroneous or irrelevant information began earlier and grew faster, reaching a wider audience than later communications by official instances and health officials.

CONCLUSIONS

Within a few weeks of the first reported Mpox cases, an avalanche of mostly false, misleading, irrelevant, or damaging information started to circulate on social media. Official institutions, including the World Health Organization (WHO), acted promptly providing case reports and accurate information within weeks, but were overshadowed by rapidly spreading social media chatter. Our results point to the need for real-time monitoring of social media content to optimize responses to public health emergencies.

CLINICALTRIAL

Deep learning for topical trend discovery in online discourse about Pre-Exposure Prophylaxis (PrEP)

Pre-Exposure Prophylaxis (PrEP) interventions are increasingly prevalent on social media. These data can be mined for insights about PrEP that may not be as apparent in surveys including personal musings about PrEP and barriers/facilitators to PrEP uptake. This study explores online discourse about PrEP using an interdisciplinary public health and computational informatics approach. We collected (N = 4,020) tweets using Twitter's Application Programming Interface (API). These data underwent a three-step neural network/deep learning process to identify clusters within these tweets and relative similarity/dissimilarity between clusters. We identified 25 distinct clusters from our original collection of tweets. These clusters represent general information about PrEP, how PrEP is communicated among diverse groups, and potential pockets of misinformation and disinformation regarding PrEP. Specific clusters of interest include discussions of medication side effects, social perception of PrEP usage, and concerns with costs and barriers to access of PrEP interventions. Our approach revealed diverse ways PrEP is contextualized online. Importantly this information can be leveraged to identify points of possible intervention for disinformation and misinformation about PrEP.

Role of basic residues in the subgroup-determining region of the subgroup A avian sarcoma and leukosis virus envelope in receptor binding and infection

Receptor specificity in avian sarcoma and leukosis viruses (ASLV) maps to the central region of the envelope surface protein, SU. Two hypervariable regions, hr1 and hr2, within this region of SU are the principal determinants of receptor specificity. The cellular receptor for subgroup A ASLV, Tva, utilizes a 40-residue, acidic, cysteine-rich sequence for viral binding and entry. This domain in Tva is closely related to the ligand-binding domain of the low-density lipoprotein receptor (LDLR). Ligands bind to LDLR via the interaction of clustered basic residues in the ligand with the acidic cysteine-rich domains of the receptor. Analysis of the ASLV envelope sequences revealed a cluster of basic residues within hr2 that is unique to the subgroup A viruses, suggesting a possible role for these residues in receptor recognition. Therefore, the effects of altering these basic residues on subgroup A envelope expression, receptor binding, and infectivity were examined. Most of the mutant proteins were transported to the cell surface and processed normally. Receptor binding was diminished approximately 50% by alanine substitution at amino acid R213 or K227, whereas substitution by alanine at R210, R223, or R224 had no effect. However, when coupled with mutations at R213 or K227, changes at R223,R224 reduced envelope binding by 90%. Mutation of all five basic residues abrogated receptor binding. The effect of the hr2 mutations on ASLV envelope-mediated infection did not parallel the effect on receptor binding. Residues 210, 213, 223, and 224 were important for efficient infection, while mutations at residue 227 had little effect on infectivity. These results demonstrate that the basic residues in the ASLV envelope have roles in both receptor recognition and post-receptor binding events during viral entry.

Mycosis fungoides skin lesions contain CD8+ tumor-infiltrating lymphocytes expressing an activated, MHC-restricted cytotoxic T-lymphocyte phenotype

In prior studies, we showed that most CD8+ cells infiltrating skin lesions of CD3+CD4+ mycosis fungoides were CD3+ T-lineage tumor-infiltrating lymphocytes (TIL) whose overall phenotype was suggestive of MHC-restricted cytotoxic T lymphocytes (CTL). However, their lack of cytotoxic-associated granzyme A mRNA suggested that they might be unactivated CTL precursors. In this study, we used single- and double-label immunohistologic techniques to assess the expression of TIA-1-reactive protein and HLA-DR by these CD8+TIL. Monoclonal antibody TIA-1 recognizes a novel family of proteins expressed preferentially by cytotoxic cells, including some that lack granzyme A. HLA-DR is a marker of T-cell activation. Single-label studies of 32 cases showed that CD8+TIL and TIA-1+ cells constituted a variable minority of the total cellular infiltrate and had a similar distribution. Double-label studies of 14 cases showed that in most instances the aggregate phenotype of the majority of CD8+TIL was CD3+TIA-1+HLA-DR+CD56-CD57-. These findings suggest that many of the CD8+TIL within skin lesions of CD3+CD4+ mycosis fungoides are activated, MHC-restricted CTL.

Most CD8+ cells in skin lesions of CD3+ CD4+ mycosis fungoides are CD3+ T cells that lack CD11b, CD16, CD56, CD57, and human Hanukah factor mRNA

To define further the characteristics of CD8+ cells in skin lesions of CD3+ CD4+ mycosis fungoides (MF), the authors used single- and double-label immunohistologic techniques and in situ hybridization to detect antigens and transcripts associated with certain types of cytotoxic or suppressor function. The cytotoxic markers included CD16, CD56, CD57, and an anti-sense probe for human Hanukah factor (HuHf) mRNA. Analysis of 23 cases demonstrated that lesional CD8+ cells were CD3+ T cells that generally lacked expression of any of the cytotoxic markers studied. Analysis of another 10 cases confirmed the CD3+ T-cell lineage of lesional CD8+ cells and demonstrated that these cells also lacked expression of the suppressor-associated marker, CD11b. In aggregate, these results indicate that most CD8+ cells in CD3+ CD4+ MF skin lesions are of T-cell rather than NK-cell differentiation. Their overall phenotype suggests that they may be major histocompatibility complex (MHC)-restricted cytotoxic T cells lacking appreciable levels of HuHF serine protease. Because the induction of CD8+ suppressor T cells is mediated by CD4+ T cells expressing the CD45RA+ RO- phenotype, CD45 epitope expression was studied in 15 MF cases. The vast majority (13/15) contained CD3+ CD4+ tumor cells that were CD45+ RA- RB+ RO+ 2B11+. This phenotype is consistent with memory T cells rather than suppressor-inducer T cells, and correlates with the paucity of phenotypically defined suppressor T cells in CD3+ CD4+ MF skin lesions.

CD45 epitope mapping of human CD1a+ dendritic cells and peripheral blood dendritic cells

The authors studied the pattern of leukocyte common antigen (CD45) epitope expression on dendritic cells in sections of human epidermis, tonsillar epithelium, dermatopathic lymph nodes, and in isolates from blood. The monoclonal antibodies (MAb) used were specific for all known CD45 epitopes, including the seven different CD45 common epitopes as well as the four known CD45R epitopes (two CD45RA, one CD45RB, and one CD45RO). Dendritic cells in all sites were uniformly reactive for the CD45 common epitopes tested except 2B11, which may recognize a CD45R rather than CD45 epitope. By single-label immunoperoxidase and double-label immunofluorescence epitope mapping of CD1a+ dendritic cells in tissue sections, it was generally difficult or impossible to detect expression of CD45RA, CD45RB, CD45RO, or 2B11. In blood dendritic cells, however, low levels of these CD45R epitopes were detected consistently using single-label immunoperoxidase staining of cytocentrifuge preparations. Monocytes were similar to blood dendritic cells except that the staining with MAb to CD45RO and 2B11 was slightly stronger. The authors conclude that dendritic cells differ from most subpopulations of lymphocytes in that CD45 common epitopes are readily detectable but the existing RA, RB, and RO epitopes are either undetectable or expressed at relatively low levels. These studies raise the possibility that CD1a+ dendritic cells may express a novel dominant CD45 isoform.