Validation of an EEG – based stress index for surgeons’ stress assessment during real working activities.

Background: A recent review on surgical applications highlighted how multidisciplinary surgical teams must possess the ability to control distractors and mastering communication and stressful events for an effective teamwork in the operating room (OR) [Kurmann et al. 2012]. In this regard, from a neurophysiological perspective stress response in humans is mediated by a complex and interconnected neuroendocrine cellular and molecular infrastructure which constitute the stress system, and it is located in both the central (CNS) and autonomous nervous system (ANS) [Schmutz, et al. 2019]. For this reason, while conventional methods for operator’s stress assessment still rely on subjective measurements derived from external supervision, several studies tried to establish reliable and objective methods. Objective: The study aimed at developing and validating an EEG-based index for surgeons’ stress assessment by combining biological (cortisol, urine and blood samples), neurophysiological (Electroencephalogram - EEG, Electrodermal Activity - EDA and Electrocardiogram - ECG), behavioural, and subjective data. Methods: The study employed a multimodal approach to decode and identify the most relevant stress-related features. Successively, surgeon’s stress was evaluated with a deep analysis of cortisol measurements from blood, saliva and urine at resting state and post-operatively (immediately after the surgery). Results: The EEG-derived stress index allowed to discriminate between two different categories of surgeons in terms of stress experienced before (pre) and after (post) the surgery. The same results were derived from the Skin Conductance Level (SCL) and Adrenocorticotropic hormone (ACTH). More specifically, the SCL was significantly higher for the less experienced surgeons. Additionally, it was observed that the ACTH significantly increased along the surgery for the less experienced surgeons, while it significantly decreased for the experienced ones (Figure 1). High and significant correlations between the EEG-derived stress index and the SCL (R > 0.6) and ACTH (R > 0.5) were also found. Conclusions: The combination of the different data allowed the surgeons’ stress to be properly characterised and measured via their brain activity. This study underscores the clinical utility of objective and real-time stress assessment tools for identifying risks of stress-related burnout, mitigate mistakes and enhance safety.