The application of detrended fluctuation analysis to assess physical characteristics of the human cell line ECV304 following toxic challenges

Abstract

In this paper we present a non-contact, impedance-based sensor system capable of characterizing the toxic response of cells to three different types of toxin. ECV304 cells were treated with 1 mM Hydrogen peroxide, 5% Dimethyl Sulfoxide, and 10 μg/ml saponin. Impedance spectroscopy was performed over a 2 h period on the cells within a commercial cell growth chamber, positioned on a pair of measurement electrodes, at frequencies between 200 and 830 kHz at 10 kHz intervals. Analysis of the impedance data was undertaken using the feature-extraction technique, Detrended Fluctuation Analysis (DFA). DFA scales the autocorrelation of a non-stationary signal, such as those generated using impedance spectroscopy for cytotoxicity testing. The correlation between the average fluctuation of the signal, F(n) (and scaling exponent, α) and a measurement of the cell size from image analysis was evaluated. The results showed that F(n) and α were strongly related to the changes of the morphological size of the cells. The results demonstrated that non-contact impedance spectroscopy, coupled with DFA can be used to monitor cell size in real time.