Graduate Students Seminar

Speaker 1: Neha Agarwala

Title

Physical activity can predict mortality for a decade in us adults: The NHANES study

Abstract

Physical activity is very important for maintaining human health. However, the use of questionnaire data for assessing physical activity is subject to substantial recall bias. With the recent development of accelerometry data, several literatures have studied the association of physical activity with health outcomes and mortality. We focus on two important accelerometry derived measures of physical activity, namely total activity count (TAC) and active to sedentary transition probability (ASTP). Our goal is to study how far into the future does physical activity predict mortality. We use a Cox proportional hazard model with time varying coefficient for TAC and fit it using penalized cubic spline. We ascertained that higher TAC is linked with lower risk to mortality. Moreover, TAC is found to be significant for predicting mortality for almost a decade in the future. While the effect of TAC is constant for age 50-70 cohort, we see a declining effect of TAC for age group 70-84. Similar findings were obtained for physical activity fragmentation measure, ASTP.

Speaker 2: Janita Patwardhan

Title

Flipping the Switch: Islet Desynchronization through Cell Silencing

Abstract

Characterized by elevated blood glucose levels, diabetes jeopardizes the health of over 23 million Americans and many more worldwide. Pancreatic β cells, located in clusters of cells called the islets of Langerhans, secrete insulin into the bloodstream, assisting with the cell’s uptake of glucose for energy. Due to gap junctions connecting neighboring cells, β cells within islets are synchronized in their oscillatory release of insulin. Unfortunately, this coordinated activity is often lost in diabetic patients. Recent experimental work has shown that silencing special hub cells can lead to a disruption in the synchronous behavior, calling into question the democratic paradigm of islet insulin secretion with more or less equal input from each β cell. By applying a mechanistic model of the electrical and calcium dynamics of β cells during secretion to a network of nodes (cells) and edges (gap junctions), we have tested this hub cell hypothesis using functional connectivity to determine the special hub cells. We began by searching for an islet with certain network properties that experienced a loss of connectivity and desynchronization with hub cell silencing, found through functional connectivity. Instead, we discovered switch cells, whose silencing leads to the loss of activity within the islet, but has no relation to functional connectivity.