Tracking Mental Workload by Multimodal Measurements During Minimally Invasive Surgery Training

Luc Chatty¹, Ahmet Omurtag², Raphaelle Roy³, Marc Garbey¹. ¹Houston Methodist Hospital, ²University of Houston, ³ISAE Supaero

Introduction: Surgery requires abundant attentional resources which can deteriorate in the face of excessive mental workload. This is particularly the case in minimally invasive surgery, where the lack of direct vision of the operation presents a challenge. To improve the care in such operations, a better understanding of stress in the OR as well as the development of training tools for stress management is necessary. Mental workload and stress have been well described in the literature, however, very few have studied a multimodal approach. The first objective is to be able to differentiate between the exercises use as fundamentals for Laparoscopic surgery, then between different levels of expertise. Finally, we want to develop a multimodal classifier allowing surgeons to compare the quality of their stress management to their pairs.

Methods: In order to monitor the stress of the operator, we use a multimodal system consisting of:

• cognionics quick 20 eeg device, The EEG records the Evoked Response Potentials (ERPs) induced by a sound probe (every 3s, 100 ms sound, 6 frequencies randomized (750-2000 Hz)).
• the eye tribe eye tracker, the eye tracker gives the pupil dilation, a well known marker of stress
• Hexoskin biometric shirt, which records heart activity and respiration

A matlab software has been developed to synchronize and record the data from those different devices. Minimally invasive surgeons train using exercises of different difficulty level. Of those, we chose three: peg transfer, circle cut and string pass. The operator is also recorded at rest to obtain a baseline.

Results: The preliminary results obtained with a dozen current subjects indicates that the following parameters will allow for a multimodal classification. Pupil dilation levels show a clear difference in amplitude between the task judged the hardest and the easiest (all subjects included). The heart rate variation high frequencies proportion is well known marker of stress. The lower it is, the more stressed the operator is. This indicate a lower stress for trained surgeons during the tasks (all tasks included). The ERP (Cz channel) levels show that for a short session of exercise, the workload increases significantly.

( p<0,05 *, p<0,01 **)

Conclusions: Our preliminary results indicate that tracking minimally invasive surgeons mental status using multimodal techniques can provide a basis to assess the stress management of surgeons. This will lead the way for a study that will observe the correlation between task performance and stress.

---

Presented at the SAGES 2017 Annual Meeting in Houston, TX.

Abstract ID: 80534

Program Number: P672

Presentation Session: Poster (Non CME)

Presentation Type: Poster