Dynamical Systems Seminar: David Albers
Population physiology: an information theory perspective
Ìý
Ìý
Date and time:Ìý
Thursday, February 3, 2011 - 4:30pm
Abstract:Ìý
In this talk I will introduce the notion of population physiology, or human physiology in the context where both: (i) medium to long time scale dynamics are the focus; and (ii) a population of humans is required to construct or discover a signal. (Metaphorically, population physiology is to physiology as climatology is to weather.) More concretely, I will outline progress in constructing an interface between electronic health record (EHR) temporal data analysis using information theoretic techniques, which typically contains time-scales of hours to years (whose origin lies in the field of biomedical informatics), and the modeling of human physiological systems. In the end, by linking the data-driven signal with physiological modeling, I demonstrate that feeding patterns in different environments can be deduced based on EHR glucose measurements. In the process of this discussion, I will also demonstrate how to apply time-delay mutual information to a complex population of time-series.