Quantitative Analysis of Exponential Curve Fitting for Biological Applications

Publisher Website
Google Scholar
Abstract
A preliminary quantitative assessment of the ability of two computer programs for fitting the sum of two exponential functions is presented with reference to the errors and number of data points normally encountered in biological investigations. Results suggest that, even in a typical biological investigation with good control of experimental errors, the errors in quantities formed by compounding all the parameters of the function may well exceed ±50% in certain circumstances. These results cast doubt on the usefulness of many compartmental analyses of biological systems where good accuracy is not obtainable and a simple model unacceptable. Error values are given for specific models. For example, the estimates of the errors in calculating grey and white matter blood flow in the brain are 5% and 13% respectively in a typical experiment; the error in estimating bone formation rate by curve analysis could be ±25%.

This publication has 5 references indexed in Scilit: