Cardiovascular disease (CVD) is the number one cause of death in humans. Arrhythmia induced by gene mutations, heart disease, or hERG K+ channel inhibitors is a serious CVD that can lead to sudden death or heart failure. Conventional cardiomyocyte-based biosensors can record extracellular potentials and mechanical beating signals. However, parameter extraction and examination by the naked eye are the traditional methods for analyzing arrhythmic beats, and it is difficult to achieve automated and efficient arrhythmic recognition with these methods. In this work, we developed a unique automated template matching (ATM) cardiomyocyte beating model to achieve arrhythmic recognition at the single beat level with an interdigitated electrode impedance detection system. The ATM model was established based on a rhythmic template with a data length that was dynamically adjusted to match the data length of the target beat by spline interpolation. The performance of the ATM model under long-term astemizole, droperidol, and sertindole treatment at different doses was determined. The results indicated that the ATM model based on a random rhythmic template of a signal segment obtained after astemizole treatment presented a higher recognition accuracy (100% for astemizole treatment and 99.14% for droperidol and sertindole treatment) than the ATM model based on arrhythmic multitemplates. We believe this highly specific ATM method based on a cardiomyocyte beating model has the potential to be used for arrhythmia screening in the fields of cardiology and pharmacology.
Read full abstract