ECG derived respiration: comparison of time-domain approaches and application to altered breathing patterns of patients with schizophrenia

In life-threatening diseases and in several clinical interventions, monitoring of vital parameters is essential to guarantee the safety of patients. Besides monitoring the electrocardiogram (ECG), it is helpful to assess respiratory activity. If the respiration signal itself is not recorded, it can be extracted from the ECG (i.e. ECG derived respiration, EDR). In the present paper, we compared six EDR approaches, namely RS-decline quantified by central moments, respiratory sinus arrhythmia (RSA), R-wave amplitude, QRS area, RS-distance and maximum RS-slope. In order to evaluate the performance of each approach, we applied each method to a database of ECGs and reference respiration signals of 41 healthy subjects. All considered methods revealed relatively small absolute mean errors of the breathing rate (BR) at rest (0.75-1.3 Bpm). The method based on higher order central moments revealed a minimum mean absolute error of 0.75 Bpm (4.40%) and a maximum correlation and concordance with the reference BR (r _p  =  0.97, r _c  =  0.97). Using this technique, we analyzed changes of respiration in patients suffering from acute schizophrenia. An increased respiration rate of about 4 Bpm was found. Additionally, alteration of respiratory ratio and reduced respiratory sinus arrhythmia was demonstrated. We conclude that a precise dynamic monitoring of breathing and the investigation of changes in breathing patterns is possible without recording respiration per se.