Research has shown transgender children respond similarly to categorical measures of gender identity as their gender-matched cisgender peers, rather than those who share their sex assigned at birth. However, categorical measures may be limiting responses and not encompassing the diversity of gender identities. We presented a continuous measure to 223 transgender children (socially-transitioned children to live and present as a gender opposite the one assigned at birth), 71 gender nonconforming children (children who have not socially transitioned but show characteristics stereotypically associated with a gender other than the one assigned at birth), 281 cisgender control participants, and 181 cisgender siblings of transgender and gender nonconforming participants. Provided a line, participants were asked to mark their gender identity, with the left-most end indicating feeling completely like a boy, the right-most end indicating feeling completely like a girl, and in between representing varying degrees of a mixture of both. Each participant’s mark was converted to a percentage, with 100% indicating feeling completely like the gender they present as (for transgender/gender nonconforming children, opposite the gender assigned at birth). A univariate ANOVA examining the effect of participant group on identity showed a significant main effect, F(3,752)= 38.72, p < .001. Post-hoc Tukey comparisons showed, consistent with previous research, transgender children (M=0.81, SD= .189) did not differ significantly from the cisgender controls (M=0.87, SD= .181), nor cisgender siblings (M=0.87, SD=.206), in their identification with their current gender. However, gender nonconforming participants (M= 0.59, SD = .309) differed significantly in their scores from each of the other groups (ps < .001). These results are consistent with previous research conducted with categorical measures. Additionally, as none of the groups showed 100% binary identification, this study also demonstrates the importance of continuous measures of gender identity within gender development research.

Self-disclosing personal stress is an effective way of managing stress. Previous research has shown that high school students disclose more when prompted by robot that discloses vulnerability. Thus, disclosure reciprocity in human-robot interaction (HRI) provides a promising way of offering stress intervention. The current study examines how different types of disclosure from a social robot affect reciprocal human disclosure and other user outcomes through a stress-disclosing activity. Using a between-subject design, 36 young adults were randomly assigned to converse and engage in a stress-disclosing activity with a robot that shared technical facts (technical condition), feelings from other users (by-proxy condition), or feelings from itself (emotional condition). We hypothesize that emotional robot disclosure will elicit longer and deeper participant disclosure, and higher robot attributional ratings (such as likeability, perceived safety, user satisfaction, and intention for future use) than technical disclosure or by-proxy disclosure. Participant self-disclosure was measured by word count and depth of response during the robot interaction. User satisfaction, intention for future use, robot likeability, and perceived safety were captured using 5-point Likert scales. Our multivariate generalized linear model revealed that participants who experienced the technical robot disclosure condition gave the robot significantly higher rating for perceived safety, but no significant differences in other variables, such as length and depth of participant disclosure among robot disclosure conditions. Surprisingly, we found that people with higher perceived stress significantly disclosed less in both breadth and depth, and rated the robot as less likeable, had lower future intention to use and user satisfaction. This study demonstrates the complexity of designing robot for stress-intervention and the importance of capturing perceived stress as a predictor for stress intervention in HRI. Future research will examine how high stress and low stress group might perceive and interact with a social robot differently in long-term repeated interactions.

T-cell immunotherapy is a treatment utilising the body’s own immune system to target cancer cells. Dr. Marie Bleakley’s immunology lab at Fred Hutchinson Cancer Research Center uses T-cell immunotherapies to treat hematologic malignancies — blood cancers which include acute myeloid leukemia and myelodysplastic syndrome (MDS). These treatments rely on the identification of neoantigens (cancer-specific gene mutations) which are potentially highly cancer-specific and are promising as targets for new immunotherapies. My project will be in the identification of these neoantigens, under the supervision and mentorship of Dr. Melinda Biernacki, a research associate in Dr. Bleakley’s lab. My project will focus on generating in silico and preliminary in vitro data for candidate neoantigens in high-risk MDS.
The project will have two parts. First, the identification of candidate MDS neoantigens using human leukocyte antigen (HLA)-binding prediction algorithms to identify mutated amino acid sequences (peptides) that have a high probability of binding HLA molecules. Because T-cell antigens consist of a peptide-HLA molecule complex, this step will filter out mutant peptides that are unlikely to be recognized by T cells. Second, in vitro assays will determine which candidate neoantigens are immunogenic (can actually be recognized by the body’s T-cells). HLA-binding predictions do not reliably predict immunogenicity of candidate neoantigens, so in vitro testing is needed to assess T-cell recognition. The end product will be the identification of potential T-cell targets from mutations in various genes that are shared in subgroups of MDS patients that could be used to develop immunotherapies for MDS.

In human-robot interaction, virtual reality (VR) provides strong potential to aid in the design of social robots. In order to understand teens' perceptions of, and interactions with social robots, we utilized a human-centered design methodology to develop a collaborative VR game for teens. The design process involved interaction studies with a local high school in three phases: (1) Brainstorming and Sketching environments on Paper, (2) Card sort game play for exploring robot design options (3) Developing and usability testing the low-fidelity VR game play. The design goals were to create an inviting game context for teens to engage in collaboration and to elicit and capture the details of teen-teen and teen-robot interactions within VR. From the design process, we found that (1) the human-centered design process engaged teens in the design process (2) teens expressed positive emotions during collaboration and (3) the asymmetric design of the VR game allowed for not only strong engagement, but also unexpected role play as teens identified themselves as the robot during gameplay. Through the design and development of this VR game, we recognized the effectiveness of a VR platform as not only a design tool, but also as an engaging collaborative platform for gathering interaction data from teens. Given this potenial, we anticipate several future applications for the study of teen-teen interaction through VR collaboration.