Assessment of Startle and Surprise Responses in Pilots Using ECG Measures

Abstract

In aviation, a startling and surprising scenario can negatively impact pilot cognition and aircraft control. When unexpected in-flight events occur, these responses can degrade performance and increase the risk of fatal accidents in commercial aviation. Reliable physiological indicators of startle and surprise (S&S) would therefore improve pilot-state monitoring. This study investigated whether the designed scenarios and cardiac measures predict subjective S&S, and whether cardiac features can discriminate between high- and low-intensity responses in pilots. Twenty-eight pilots participated in the experiment in the SIMONA Research Simulator. The study included two baseline (non-event) scenarios and four scenarios designed to induce S&S responses. Engine Failure (EF) and a Primary Flight Display Failure (PFDF) were designed to mainly induce surprise, while a False Stall warning (FS) and a Lightning Strike (LS) were designed to mainly induce startle. During the experiment, electrocardiography and eye data were collected. Post-scenario, participants completed subjective startle and surprise questionnaires. Cardiac measures were corrected for pilot control input activity to isolate the signal attributable to S&S responses. EF and PFDF produced the two highest surprise ratings, and LS the highest startle rating, as designed. In contrast, the startle-designed FS showed high scores on both startle and surprise. The scenario type predicted subjective responses, but the heart rate (HR) window means added nothing beyond it. Physiological features did not discriminate high- from low intensity trials beyond scenario. Explanatory ocular
data analysis showed the same pattern, with scenario type being significant. Event-locked HR trajectories nevertheless showed distinct scenario-driven temporal clusters. Hence, scenario type reliably drove both the subjective response and the event-locked cardiac trajectory. The mean HR change during the response windows did not predict the intensity with which individual pilots experienced those events. Neither the cardiac-derived nor the exploratory ocular features discriminated between high- and low-intensity events. These findings suggest that the cardiac response is more
sensitive to the scenario that produced it than to the degree of S&S pilots reported experiencing.