Development of algorithms for the extraction and evaluation of Respiratory Rate from ECG and PPG signals
Donatella Vecchione, Annalisa Letizia, Vincenza Tufano, Francesco de Nola
The thesis work derives from my internship with Teoresi group Engineering S.p.A company, which allowed me to understand how the working environment is set up and to apply the concepts learnt during university courses. The project aims to propose an alternative method to extract respiratory information from ECG) and PPG signals.
In particular, it focused on developing robust and universal methods for estimating accurate respiratory rate (RR) regardless of the intra and inter-individual variability that affects ECG and PPG features, in order to monitor respiratory activity in a cheap and non-invasive way and allowing to substitute the current and invasive respiration-monitoring techniques which may interfere with natural breathing and which are unmanageable in some applications such as daily monitoring (stress test or sleep studies).
Continuous monitoring of respiration plays a key role in the detection and management of different conditions, such as stress and sleep disorders. Biomarkers like respiratory rate, breathing phases, and tidal volume are relevant for the detection of mental stress, anxiety, and sleep apnea events. In addition, the coupling between respiration and heart rate has been used as a key parameter for the aforementioned conditions as well as for the analysis of the interactions between the cardiac and respiratory systems. Breathing rate (BR) is widely used in clinical setting for diagnosis and prognosis. On general hospital wards it is usually measured by manually counting chest wall movements. This practice shows some drawbacks, in fact it is time consuming, inaccurate, and poorly carried out. An alternative approach may be to estimate BR from ECG and PPG signal, which are already routinely measured in a wide range of clinical contexts and moreover they are easy to record. A recent study in healthy population showed that BR can be estimated from the ECG and PPG signals with a similar precision to Impedance Pneumography, the current clinical standard for electronic BR measurement and this represents a promising area of research in clinical routine.
Development of MatLab algorithms for the extraction and evaluation of Respiratory Rate from ECG and PPG signals.
IEEE, PUBMED, GOOGLE SCHOLAR, SCIENCE DIRECT
This work aims to show a series of techniques in order to estimate respiratory rate (RR) from ECG and PPG signals. The illustrated algorithms can be considered non-invasive and alternative methods to the standard ones used nowadays in both hospital and private settings. Among the most important advantages, the listed ones are included:
- low cost;
- efficiency both in terms of timing and data reliability;
- possibility to monitor simultaneously more clinical parameters;
- reduction of motion artifacts;
- computationally cheap;
The respiratory rate predicted by the implemented algorithms shows a good correlation with the reference respiratory signal acquired with direct methodology. This suggests that the features used are good indicators that can be exploited for the calculation of breath rate. It is also important to note that the results of the exploited approach are not degraded by different patient condition.
In conclusion, it can be stated that the results obtained are of remarkable clinical importance as they may be useful as an inexpensive and rapid method for the screening of possible pathologies or as indicators/biomarkers of an unhealthy lifestyle. It is worth noting that the information of different EDR or PDR signals can be fused in order to increase the accuracy of the estimation of the respiratory information. Several fusion methodologies have been proposed in the literature. Such fusion has been shown to be more effective when combining EDR signals containing complementary (i.e., non-redundant) respiratory information. Future work could focus on evaluating whether the fusion of the best-performing EDR and PDR signals in this study could result in an increased performance or if an improvement could be reached by fusing the worst-performing ones.
It is still possible to make enhancements and optimizations in order to validate the algorithms on other datasets which include subjects affected by pathologies investigating the effects of a range of technical (sampling frequency, measurement site, signal acquisition equipment) and physiological (age, gender, breath rate) factors on the quality of extracted signal.
Possible future works may concern the telemedicine field, in particular the development of respiratory monitoring contactless techniques so as to further reduce the invasiveness on patients; another important future work may concern the automotive field: PPG signal, together with its extracted Respiratory Rate could be used to assess real time driving drowsiness (or in general to trace the psycho-physical state of the driver) to with non-invasive methods.