Measuring the Coupling of Hormone Concentration Time Series Using Polynomial Transfer Functions

Abstract

This chapter discusses use of polynomial transfer functions for measuring the coupling of hormone concentration time series. A method for estimating the strength of coupling and the relationship between two hormones is outlined. The performance of polynomial transfer functions is reported as a tool to reconstruct the coupling between various combinations of hormone concentration–time series data derived from a known mathematical construct that describes certain hypothetical endocrine feedback networks. The use of synthetic data allows comparison between method-predicted coupling and known (model-defined) hormone interactions. The measurement of the percent of the variance accounted for by the polynomial transfer function was consistent. The measurement is affected by the number of hormones (or inputs) involved in the control of a given hormone. It is shown that polynomial transfer functions can consistently describe the strength of the relationship between two hormones, and that the ability to measure this coupling decreases with increasing data noise and with increasing error in the estimated hormone half-lives. This technique is expected to be useful for the investigation of normal hormonal physiology and pathophysiology.