Our brain orchestrates breathing through a balance of afferent sensory information and efferent volitional information. An anomaly in this balance triggers painful sensations regrouped under the term of dyspnea. Dyspnea represents a physical pain and also a psychological distress. Therapeutical opportunities to correct dyspnea respiratory abnormalities exist, however they are not always available and sometimes not sufficient. In these cases, the dyspnea is called persistent and need new therapeutical leads to be alleviated.
In this context, an olfactory stimulation (OS) could represent a good candidate. OS has been identified to yield beneficial clinical effects in non-respiratory pathologies such as depression or dementia but also in respiratory pathologies. Some studies showed that using menthol could reduce the sensation of respiratory discomfort.
In clinic, to adapt to the omnipresent aspect of dyspnea, an OS should be able to be delivered at any moments. Thus, we propose to develop a device that identifies respiratory stress (RS) and subsequently diffuses an OS. The first step is to develop an algorithm capable of identifying RS and decide whether or not deliver an odor. To do so, we used the autonomous nervous system sympathetic branch increase during a stress event that we can capture with the measure of Heart Rate Variability (HRV) metrics computed from the Electrocardiogram. We used a machine learning algorithm to decide to deliver the OS or not as a function of the HRV metrics. To test our algorithm, we induced experimental RS in 30 healthy subjects while diffusing a pleasant, unpleasant and without OS while recording their ECG.
Our results show that our algorithm manages to identify RS and deliver odor during stress. Besides, we measured that with an OS, specifically chosen by the subject as being pleasant, the algorithm no longer identify stress HRV profile present without OS, suggesting that OS alleviate respiratory discomfort.