2016 Habilitation à diriger les recherches
2011 to date Associate Professor of Human Factors and neuroergonomics at ISAE (Institut de l’Aéronautique et de l’Espace) in Toulouse
2010 Ph.D. in Neurosciences (ISAE-SUPAERO "best thesis award")
2006 Master of Neuropsychology ("very good")
2004 Master of Human Factors ("good")
The way that the brain dynamically allocates mental resources according to the task demand is crucial from a human factors point of view. A strong increase of mental workload, or on the contrary, a long episode of very low mental workload, can increase the likelihood of error.
Monitoring mental workload
The recording of the brain activity can give indications on the amount of mental effort than an individual put in a given task. Light brain imaging techniques such as fNIRS (functional near infra-red spectroscopy) are field deployable, contrary to much more cumbersome system such as fMRI.
Fnirs allows to measure concentration changes of oxygenated hemoglobin (HbO2) and deoxygenated hemoglobin (HHb). As an active brain region is enriched in HbO2, this device gives clues about the level of mental effort-via the amplitude of change in the HbO2 concentration-and about the particular cognitive functions engaged in the task, thanks to the identification of the active brain areas. For example, the figure below shows that fNIRS measurements can discriminate various mental efforts, during an easy vs. complex landing scenario in flight simulator.
Negative consequences of a high mental workload
Context of high mental workload can consume a large part of attentional resources that can no longer be allocated to the processing of other critical stimuli, such as auditory alarms in the cockpit. This phenomenon called “inattentional deafness” can be studied thanks to EEG/ERP. The lack of attentional resources is indexed by a decline of the brain reaction to the auditory stimuli, such as in the example below, the “P300” ERP amplitude is progressively disrupted by increased mental workload in a simplified piloting task (low and high load) in comparison to a control condition with no mental workload. This alteration of the brain reaction to the auditory stimuli is correlated with the probability to “miss” auditory stimuli.
Stress and emotion
The analysis of aviation events shows that stress and emotion are important contributors to the occurrence of inappropriate behavior or erroneous decision-making.
Emotion and decision-making in aviation
It has been found that some crews showed a trend in flying through convective storms instead of deviating around them, or persisted in conducting the landing procedure despite meteorological condition that would require a go-around. This latter behavior seems favored by various negative emotional consequences attached to the go-around maneuver such as an increased uncertainty and the financial cost that it generates (passenger delay, kerosene consumption…). We can simulate such negative emotional consequence with a financial system (reward and punishment) in fMRI, as illustrated below.
The results of an experiment revealed that risky decision-makers activate less the dorsolateral prefrontal cortex, a brain region known to be involved in executive functions and decision-making processes
Emotion and cognition interacts in the prefrontal cortex
Such effects of emotion are encountered in everyday life. In our anxiogenic and stressful world, the maintenance of an optimal cognitive performance is a constant challenge. It is particularly true in complex working environments (e.g. flight deck, air traffic control tower), where individuals have sometimes to cope with a high mental workload and stressful situations.
It is believed that stress can reduce human cognitive efficiency, even in the absence of a visible impact on the task performance. Performance may be protected under stress thanks to compensatory effort (e.g. coping), but only at the expense of a cognitive cost. Such psychophysiological cost, invisible at the naked eye, may be indexed via the analysis of the prefrontal cortex activity and the heart rate.
We developed a novel task called “Toulouse N-back Task” that intensively engaged both working memory and mental calculation. During the task, our participants were summited to the threat (or not) of unpredictable aversive sounds. Results revealed that task difficulty (mental effort) was successfully indexed by prefrontal activity measures with fNIRS. Also, the threat of unpredictable aversive sounds provoked an increase brain activity, suggesting a reduced efficiency because of the deleterious effect of the stress. As illustrated below, these effects were also supported by the analysis of the cardiovascular activity, with an increased heart rate in response to the stress and the cognitive load.
Another important factor of performance variations concerns aging. The effects of aging on cognitive performance must be better understood, especially to protect those who participate in risky activities such as aviation.
The impact of aging on the brain functioning of general aviation pilots
We assessed aging effects in private pilots, when flying or when engaged in neuropsychological tasks evoking important intellectual abilities. For example below, sixty-one highly educated pilots performed spatial working memory or planning and reasoning task and we evaluated their performance and cerebral activity.
We found that aging has significant effects on the intellectual performance in several domains, with the older group performing worse than young and middle aged individuals. Importantly, older individuals are more impacted when tasks are more complex. Yellow and red arrows show that the gap between the three age groups is larger when difficulty is higher (i.e. 10 and 12 boxes, the highest level of difficulty).
But there is hope ! When comparing low and high performers, older participants showed compensatory brain activation patterns, with high performers demonstrating greater recruitment in key brain regions during the planning and reasoning task. Thanks to enhanced effort and motivation, older individuals are able to increase their intellectual performance. In summary, compensatory cognitive strategies may counter cognitive deterioration due to aging.
Human machine interaction
An important aspect for neuroergonomics is the improvement of the coupling between human and technology. Upcoming technological breakthrough such as single pilots operations, 4D navigation, increased automatisms, ecological human machine interface (data fusion, synthetic presentation of information) or generalization of touch screens must be assessed in terms of impact on aviation operations.
Monitoring pilot’s visual circuits with eye tracking
A current work with Air France involves eye tracking technology to improve pilot’s monitoring strategies of the cockpit instruments. The idea is to define new training procedure by providing a personalized feedback to pilots with their own visual circuits. Experiments are conducted at Air France, in full flight simulator.
Eye tracking technology allows “monitoring the monitoring” and gives indications on the actual information processed by the crew. Red dot represents the current point of fixation of the pilot. Current results tend to show that gaze allocation of pilots who failed to perfectly perform an approach is sub-optimal compared to the most accurate pilots. In particular, the attention is too less or too much focused on a variety of instruments.
Sometimes, crews tend to be “automation addict”, which can generate complacent behavior and an excessive dependence on automatisms. These phenomena promote the occurrence of a particularly prominent typology of error, the failure of the crew to properly monitor the flight instruments. This lack of monitoring can be particularly hazardous during the final approach phase. In the figure below, the analysis of the pilot’s visual circuits emphasizes the trend of the pilots to monitor less often critical information such as the attitude indicator (AI) or the ECAM when full automatisms are engaged.
Examples of funded projects that I lead at ISAE
Dassault Aviation chair
Focused on neuroergonomics, highly autonomous systems and system engineering, the Dassault Aviation chair fosters synergy between several research groups at ISAE-SUPAERO (Neuroergonomics, Control & Decision, System engineering)
Research topics encompass :
- The online monitoring of the flight dynamics and flying performance via behavioral, physiological, and cerebral measurements
- The development of ecological human machine interfaces that provide synthetic information on the flight dynamics, thanks to a comprehensive data fusion incoming from various sensors of the aircraft. Ecological interface allows intuitive decision-making and reduces mental effort and cognitive bias
- The definition of architectures for rule-based systems oriented toward the human machine interaction performance, the compliance with the required safety level, and the mastering of complexity
In the chair, current works in neuroergonomics specifically concern the improvement of the human operator coupling via the monitoring of the visual attention of the pilots. The several steps are :
- Definition of « gold standards » describing ideal airline pilots’ visual information taking for each flight phase
- Prototyping of a system allowing the online fusion of current ocular data and the database with « gold standards »
- Offer the possibility to trigger attention getter in the cockpit when strong deviations from the ideal visual circuit are detected, for example the blinking of an instrument that was neglected by the crew during a too long period, suggesting attention narrowing, attention tunneling, or distraction.
Pilots’ visual circuits recorded during flight simulator scenarios with a Tobii Pro Glasses 2
NEUROERGO project (NEUROlogical and physiological metrics of strEss and cognitive load for modeling and online monitoRinG of Operator) NEUROERGO project brings together several academic partners working in the neuroimaging and aviation fields. The overall objective is to reach a deep understanding of the respective effects of mental workload and emotion (in particular acute stress, or stress related to a sustained anxiety). In the project, participants are assessed in various experimental platforms such as fNIR or flight simulator.
During each experiment, the effect of mental workload (modulated by task difficulty) and emotion (with or without stressors) are assessed on behavioral performance, brain activity, pupil diameter, and heart rate. Investigating brain and peripheral activities at the same time help to dissociate the respective effects of cognitive and emotional factors.
A participant is equipped with cardiac electrodes before being placed in the fMRI scanner.
On the top, the screen displays the image of the pupil, recorded continuously during the fMRI acquisition
Cognitive and emotional influences are also studies in flight simulator in general aviation pilots. Psychosocial stressors are administered to the participant during piloting (a camera is placed in front of the participant and they are instructed that their performances will be compared to other participants. Likewise than during the fMRI protocol, multiple measurements are performed such as brain activity (as measured with fNIRS) and heart rate. The mental workload is manipulated via an additional task, performed in parallel with the piloting, and displayed on a tablet computer on the right of the pilot.
A participant equipped with the NIRScout fNIRS (nirx), during the navigation, attention has to be also paid on the task displayed on the tablet touch in order to increase mental workload.
AIRTIUS is a multidisciplinary project with fundamental and applied components. The objective is to respond to industrials needs in terms of technological evolution of the cockpit’s human machine interface, while ensuring a high level of flight safety. The consortium involves five partners : 2 academics (ENAC, ISAE) and 3 innovative companies of the Midi-Pyrénées/Languedoc-Roussillon (Astrolab, Ingenuity i/o and Intactile Design) The overall objective is to anticipate and accompany the disappearance of a number of tangible interactors due to the future introduction of large touch screens in the cockpit. The challenge for the industrials is to cope with the increasing system complexity with a higher flexibility and reduce costs. Nevertheless, these evolutions have notable limitations for an operational use. Contrary to the current physical control devices (buttons, selectors, thrust lever, pull switch…), whose perception and manipulation is also made through the sense of touch and proprioception, touch user interfaces are complex to use without the gaze or when the plane encounters some turbulence. It is yet crucial to guarantee usage in degraded context : smoke in the cockpit, stress, fatigue, high mental workload… Moreover, compared to full usage of the physical space, touch screens does not promote the building of shared situation awareness. A mixed approach, integrating touch screens, tangible objects and interaction in the whole physical space could better take into account perceptive and sensorimotor capacities of the pilots.
Working facility for the development and test of new cockpit interactors.
Real flight experiments will be undertaken to assess the usability of the new cockpit interactors in ecological conditions.
Increasing demand in air traffic control efficiency implies higher levels of automation. In particular, airports require automated tools and assistance to ensure an always growing throughput and capacity. Currently, arrival and departure assistance are in operation on several airports, but pushing automation further is promising in many aspects (time and fuel savings, improved overall capacity...). However, if ideas and algorithms exist for automating taxiing or introducing robots on the apron, one remaining challenge to be addressed is the transition period that current concepts cannot deal with. The purpose of this project is to propose a ground control interface which would not only enable full-automation, but also support the transition period during which the human operator would be able to gradually take advantage of the algorithms and automation in a non-intrusive manner. This transition-centred design therefore grants acceptance and overall efficiency. In order to achieve this result, we propose a new and very human -centred HMI for Air Traffic Controllers Officers (ATCO). The initial idea is to use a ground radar image with taxiing aircraft that would also capture ATCO’s intentions and display the path proposed by promising multi-agents distributed algorithms. The user can then either reject, accept the suggestion, or interact with it. ATCO/algorithms interactions are carefully designed y to make them intuitive, quick, and straightforward. Touch sensitive technology insure that input can be done on the flow, at no cost and comprehensively. Amongst many other important benefits, this project makes it possible to evaluate auto-taxiing benefits in real conditions.
The results of the entire experiment campaign show that the Modern Taxiing platform can increase the overall performance of ground taxiing, with greater thoroughput and less time in the ground sector. The use of the tugs appears to reduce the technology gains, with the greatest performance occurring when using only the interface. However, the advantages due to technology also come at a price, with an increased in perceived workload although the physiological response does not significantly vary. The technology is still currently too immature for accepted use by the air traffic controllers, but comments made during debriefing suggests that with improvement, the participants would be accepting of this new technology in an operational context. The technology also appears to assist participants during some operational events, namely, in managing the impact of a towed aircraft, a change in configuration, and a pilot error.
Despite encouraging results (e.g. reduced fuel consumption), the confidence in the system has to be improved, as showed by a lower SATI score in the “mota” condition.
10 avenue Edouard Belin
Tel : 00 33 (0)5 61 33 81 28
mail : mickael.causseATisae.fr