Towards a neuroergonomics approach to understand inattentional deafness in aviation

Abstract

Aviation safety analyses often reveal that pilots persist in erroneous decision-making despite auditory warning. Classically, the lack of response to alarms is attributed to pilots ignoring consciously such warnings due to decision biases or risk taking. However these explanations do not fully account for the misperception of critical auditory alarms as observed in flight simulators (Dehais, Tessier, Christophe, & Reuzeau, 2009). A relevant approach is to consider the concept of inattentional deafness, that is the absence of reaction to auditory stimuli (Macdonald & Lavie, 2011). Since high mental demand is a key element of flying, it is likely that the important information processing of instruments could interfere with the concurrent appraisal of unexpected auditory alarms and leads to inattentional deafness. To investigate this hypothesis, we adopted a neuroergonomics approach (Parasuraman, & Rizzo, 2007) that promotes the use of brain imaging techniques to understand the neural mechanisms of human error. Indeed, we recorded electrophysiological measurements (EEG) while participants were supervising an automated landing sequence by considering both visual and auditory signals. Our results revealed evidence of an early visual-to-auditory gating mechanism that occurs when visual parameters (“land”) were contradictory to the auditory alarm (“go around”). This mechanism attenuates early auditory processing (N100) and could partly explain the “inattentional deafness” in aeronautics (Scannella, Causse, Chauveau, Pastor, & Dehais, 2013). In a second experiment, pilots were required to perform a landing decision task based on the analysis of visual indicators while continuous EEG measurements were performed under different load conditions. During the task, a tone was presented, either standard, which participants were told to ignore, or deviant (“the alarm”) which participants were told to report. Preliminary behavioral results showed that up to 30% of deviant sounds were not detected during the high load condition. The analysis of the event related potential showed that a drastic diminution of the late auditory component (P300) amplitude was concomitant with the occurrence of inattentional deafness (Giraudet, St-Louis, & Causse, 2012). This shows that some sounds may remain unnoticed despite the fact that they are perfectly audible. These protocols were adapted and replicated in a motion flight simulator and showed that critical task load may totally impair auditory alarm perception (Dehais, Causse, Vachon, Regis, Menant, & Tremblay, 2013). These three recent experiments were the very firsts to demonstrate the existence of inattentional deafness to aircraft alarms and provide a methodology as well as objective measurements to assess the efficiency of future auditory alarms.