Neural network for photoplethysmographic respiratory rate monitoring

The reflection mode photoplethysmographic (PPG) signal was studied with the aim of determining respiratory rate. The PPG signal includes respiratory synchronous components, seen as frequency modulation of the heart rate (respiratory,sinus arrhythmia), amplitude modulation of the cardiac pulse and respiratory-induced intensity variations (RIIVs) in the PPG baseline. PPG signals were recorded from the foreheads of 15 healthy subjects. From these signals, the systolic waveform, diastolic waveform, respiratory sinus arrhythmia, pulse amplitude and RIIVs were extracted. Using basic algorithms, the rates of false positive and false negative detection of breaths were calculated separately for each of the five components. Furthermore, a neural network was assessed in a combined pattern recognition approach. The error rates (sum of false positive and false negative breath detections) for the basic algorithms ranged from 9.7% (pulse amplitude) to 14.5% (systolic waveform). The corresponding values for the neural network analysis were 9.5-9.6%. These results suggest the use of a combined PPG system for simultaneous monitoring of respiratory rate and arterial oxygen saturation (pulse oximetry).