Automatic Detection of Sympathovagal Response Using HRV Analysis. Case Study: Resident Surgeons During Training and Their First Laparoscopic Surgery

Monitoring the physiological response during training can provide valuable information about how surgeons cope with the stress associated with this activity. Heart Rate Variability (HRV) analysis is a widely used method of assessing the autonomic nervous system (ANS) response, providing important knowledge on the stress state of subjects and how the ANS intervenes in the response to such stress. During a stressful task, the ANS releases hormones such as cortisol and adrenaline into the bloodstream, triggering an increase in heart rate, increased perspiration, sharper concentration and other physiological processes. In this study, 6 resident surgeons were monitored at rest, during training with a surgical simulator, and during their first surgery. Correlations between HRV features and surgeon performance were investigated. Six Machine Learning (ML) models were trained to detect mental stress using a public dataset comprising 84 5-min ECG recordings labelled as stress/rest. Subsequently, these models were tested on data collected from the resident surgeons. In conclusion, pre-trained ML models demonstrated significant efficacy, recognising mental stress in surgeons with a specificity and accuracy exceeding 80%.