WIP: Integration of Computational Modeling Active Learning Activities Within a Core Graduate Organ Systems Physiology Course

Abstract Work in Progress: Integration of Computational Modeling Active Learning Activities within a Core Graduate Organ Systems Physiology Course Biomedical Engineering graduate students at the University of Virginia must take an organ systems physiology course in the spring semester of their first-year core curriculum. This course covers systems physiology foundations as well as concepts from most of the major organ systems, including nerves, muscle, the heart, circulation, lungs, kidneys, and metabolism. Our graduate students come from a wide variety of undergraduate academic backgrounds and majors, however, which presents significant pedagogical challenges for any sort of a "one-size-fits-all" physiology course, especially at the graduate level. Some of our students already possess multiple semesters of college-level physiology background, whereas others come into our program with no physiology background whatsoever. A traditional didactic approach to teaching physiological concepts would either fail to engage the more experienced students (e.g. if targeting those with less experience in the subject matter), or would risk leaving those with no background in physiology behind (e.g. if the course were paced to accommodate those with at least some prior background in the subject). Our challenge has therefore been to deliver instruction that simultaneously engages the full breadth of graduate student backgrounds while also providing sufficient rigor for a graduate-level understanding of physiology. To address this challenge, we developed a partially flipped course that relied on directed reading assignments and the preparation of study sheets in response to pre-reading questions to ensure that all of the students obtained a baseline working knowledge of the fundamentals for each topic. In-class time consisted of short lectures clarifying points of confusion and covering specific topics in more depth, as well as team-based active learning workshops that usually focused on applying computational modeling to physiological examples. For many of the workshops, students implemented "classic" physiological models in MATLAB — e.g. Hodgkin & Huxley's excitable membrane model, Huxley's muscle contraction model, Suga & colleagues' time-varying elastance model of the left ventricle, etc. The goal of the modeling exercises was to reinforce physiological concepts while also providing students with the opportunity to apply engineering and mathematical approaches for predicting the behavior of these systems. Modeling assignments were evaluated as group homework problems, whereas physiological concepts were assessed through either a final exam (Spring 2019) or three midterms (Spring 2020). Student feedback was assessed at the end of each semester by both in-person discussions and anonymous surveys. A qualitative coding approach was used to evaluate free response survey questions. We analyzed data from the past two years, in which a total of 53 graduate students were enrolled (83% of whom were PhD students). Based on homework scores and student feedback, most students performed very well on the modeling aspect of the course and were comfortable with applying mathematical models. The results regarding understanding physiological concepts were more mixed, however, with mean scores on the exams at ~80%, and just over half of the students performing B- or worse on those assignments. Student feedback from the interactive discussions and the anonymous surveys also reflected that discrepancy, with many students stating that they felt that the course over-emphasized the modeling at the expense of the physiological concepts. Going forward, we are restructuring our core curriculum to focus on the physiology content earlier and providing peer teaching opportunities to address the discrepancy in student backgrounds. Modeling applications will be included in a revised modular follow-up course.