Equivalent circuit models for a biomembrane impedance sensor and analysis of electrochemical impedance spectra based on support vector regression.
Journal:
Medical & biological engineering & computing
PMID:
30941674
Abstract
In this study, an electrochemical impedance biosensor was developed as a simple and fast method for real-time monitoring of biofilm binding properties via continuous impedance spectroscopy. To prepare the sensing membrane, cells were immobilized onto gold electrodes with nitrocellulose membranes. Different cell growth features were measured with the impedance instrument and analyzed using an equivalent model for data fitting and support vector regression (SVR) for data processing. The collected impedance spectra revealed that the binding attachment areas of cells differ depending on the cell density. Our results demonstrate the usefulness and feasibility of training our impedance-based sensor with a small amount of data to predict the effective area of different biofilms (GE, NGE, and CNGE), with a prediction error of 9.8%. Graphical abstract.