Undergraduates can complete an honors thesis in the lab, working with the PI on a structured research project related to maternal or postpartum health.

Graduate students can complete a thesis in the lab, working closely with the PI on an original research project related to perinatal sleep, maternal health, or postpartum health behaviors.

Master of Public Health students in epidemiology can complete required capstone or projects using the lab’s data, methods, or ongoing studies.

Master of Science students in epidemiology can complete thesis-based research projects focused on study design, data analysis, or maternal health outcomes.

This study uses a statistical method called latent class analysis to identify patterns in mid-pregnancy food intake within a representative perinatal sample, based on when people eat throughout the day. It then examines how these eating patterns relate to depression during pregnancy.

Although prenatal depression accounts for a significant portion of postpartum depression and affects broader health outcomes, it remains understudied in the current literature. More broadly, this study aims to better understand eating behaviors during pregnancy in order to improve perinatal wellness and inform future interventions that support maternal health.

By Christina Personette (with contributions from David Phan and Namhyun Kim)

This study develops a composite sleep health score integrating modifiable dimensions of multidimensional sleep health, sleep regularity, timing, and duration, and uses latent class growth analysis (LCGA) to identify distinct sleep health trajectories from mid-pregnancy through one year postpartum. We then examine the associations between these trajectory groups and postpartum weight retention.

Few studies have characterized sleep health longitudinally across the pregnancy–postpartum transition, particularly using ecological momentary assessment (EMA) data. Understanding modifiable sleep trajectory patterns during this critical period may provide novel targets for behavioral interventions aimed at reducing postpartum weight retention.

By Namhyun Kim

In the postpartum period, disrupted sleep-wake cycles often lead to irregular chrononutrition patterns, such as late-night snacking and compressed fasting windows. These patterns can contribute to weight retention and metabolic risk. This project is a secondary analysis of the Sleep GOALS (Goal-focused Online Access to Lifestyle Support) pilot randomized controlled trial. We are investigating whether a behavioral health intervention targeting sleep, diet, and physical activity can improve eating timing.

By Rya Clifton

This study aims to apply compositional data analysis for outcomes. Compositional Data Analysis (CoDA) is a newly developed method to examine components that sum up to a certain framework. For example, sleep, sedentary behavior, and physical activity change within a 24-hour framework, creating perfect multicollinearity when putting them into generalized linear models. Several methods have been proposed to address multicollinearity, but application for perinatal data is scarce.

We focus on Dirichlet regression to model compositional data for outcome, which is even more scarce in perinatal epidemiology. The SGOAL study is a pilot randomized controlled trial to investigate the efficacy of interventions for sleep and weight retention throughout pregnancy. We apply Dirichlet regression to investigate how sleep, sedentary behavior, and physical activity change at the same time.

By Takashi Yoshimasu

This study aims to apply compositional data analysis for exposure. The study is a cross-sectional study examining the association between antenatal depression and components of activity during the day, including sleep, sedentary behavior, and physical activity. Compositional data analysis (CoDA) was devised in 1982, but the application for perinatal data is recent. When these components are included together in a model, they create perfect multicollinearity. Several methods have been proposed to address this multicollinearity, including isotemporal substitution model or isometric log-ratio transformation. We aim to apply isometric log-ratio transformation for HABIT.

HABIT trial is a sequential randomized control trial to address gestational weight retention by offering an education program to promote sleep, weight management, and physical activity for pregnant populations with BMI >25. We use the data from baseline to examine the association between antenatal depression and components of activity.

By Takashi Yoshimasu

This study investigates the association between fasting interval and depressive symptoms. This is a secondary analysis of prospective cohort study with 230 pregnant populations. Fasting interval is defined as the interval from dinner to the next breakfast. Some research shows that the longer the interval, the better for depressive symptoms. However, these research studies are involved in non-pregnant population in experimental settings. We aim to investigate how fasting interval is associated with depressive symptoms within pregnant populations.

By Takashi Yoshimasu