Researchers Study Future, Safer Flight Decks
The secret to addressing human factors is adapting the cockpit to the way the human brain works best, say researchers.
Functional near-infrared spectroscopy enables a better understanding of brain activity. The pre-frontal cortex is where rational decision-making takes place.

A research laboratory in Toulouse, part of the ISAE SupAéro engineering school and funded by insurance firm Axa, is making strides toward improving safety by designing flight decks adapted to the way the human brain works best. The research in “neuroergonomics” uses the latest developments in neuroscience to cross the threshold of the pilot’s brain.


“At stake is resolving the old debate about whether the crew or the aircraft’s design caused an accident. We know enough now to create strategies for the aircraft to help the pilot,” Godefroy Beauvallet, head of the Axa research fund, told AIN during a visit to the laboratory in March.


In the crosshairs are two categories of accident–loss of control in flight (LOC-I) and controlled flight into terrain (CFIT)–where human factors account for 80 percent of the problem, said Mickaël Causse, associate professor. In these instances, the crew persists with an irrational decision despite alarms from the aircraft, he explained. An example is the fatal Aeroperú crash in 1996, when an incomplete maintenance procedure left airspeed and static pressure probes blocked. After takeoff, the crew argued and focused on altitude and speed but never came up with the correct procedure.


Assistance with Decision-Making


Lack of mental flexibility and attention tunneling can be found in the behavior of otherwise proficient pilots. Such behavior, caused by intense stress, can be characterized as “pathologic” because some brain lesions have the same consequences, Dehais noted. Both stress and lesions can impair cognitive abilities.


Today’s flight decks cannot help crews facing such predicaments, and the research aims to change this. The team, led by professor Frédéric Dehais, is combining neuroscience, human factor developments, signal processing and computer science to understand “neural mechanisms.” The researchers measure how well or how poorly pilots make decisions under stress, in laboratories ranging from an fMRI scanner, where exhaustive data can be collected, to a light aircraft cockpit in flight. The latter, while much more realistic, has limited room for accommodating equipment. Occupying the middle ground is a full-motion simulator developed in-house.


Dehais is keen to research how emotions, as opposed to rational thinking, influence decision making. For instance, a pilot can feel pressure not to abort an approach when so close to the destination. He might be under pressure, such as wanting to avoid a delay or not wanting to face questioning if he goes around. This is called hot reasoning.


Dehais and Causse have found a way to reproduce the negative emotional consequences associated with the go-around. An aviation task during the landing phase was performed in an fMRI scanner. The decision had to take into account various risks, so pilots were offered a monetary incentive for every successful landing and punished with a financial penalty for a go-around.


Most pilots made conservative choices when the incentive was nil and riskier ones when it was significant. “Money creates a basic emotion, comparable to an attraction for food or sex,” Dehais said. Researchers used the fMRI to measure activity in the pre-frontal cortex (the area of the brain where rational decision-making takes place) and found higher activity for the conservative courses of action and lower for the risky ones (because the brain was overwhelmed by emotions). In extreme instances of incapacitating stress, the pre-frontal area might shut down, Causse said.


Situational Deafness


A mystery in a number of accidents has been how a crew could ignore an aural alarm. When on final approach to Megève airport in the French Alps in 2009, a pilot and his instructor in a light piston single were fixated on the airplane landing ahead of them and the slow pace of its exit clear of the runway. Showing zero reaction to the piercing alarm that signaled their landing gear was not extended, they touched down and slid to a halt on the airplane’s belly.



Dehais re-created such deafness in an experiment on the ground. Pilots in the test had to make a decision about landing, with various levels of difficulty, the toughest being a strong crosswind. The researcher had an alarm sounding during the simulated approach. In 30 percent of the most difficult cases, the pilots did not hear the alarms.


Looking for answers, the researchers used an electroencephalogram (EEG) to measure activity in the area of the brain that responds to aural stimuli. In pilots who had not reacted to the alarm, it was inactive. “In 0.1 second, the visual channel can shut down the aural one,” Dehais commented.


What could be regarded as the baffling dismissal of a loud warning is probably the legacy of a primitive predator/prey survival mechanism, and researchers have already recognized that more strident alarms will not fix it. The team plans to install EEG equipment in a SupAĂ©ro-owned Robin DR400 piston four-seater for further studies, once the researchers have solved the problem of electromagnetic interference.


Researchers are considering several solutions to this situational deafness. One is to create an alarm that calls the pilot’s first name–the last thing the brain hears before complete deafness. Another would be to remove the information on which the pilot is focusing too intently and replace it with more relevant guidance. This feature could be activated when the flight computers perceive the pilot is not reacting to an alarm.


A short video loop (Vine-type) could pop up, showing an avatar performing the required action. For example, in a stall, a pilot’s intuition can command him to pull back the stick, something that happened in accidents such as AF447 (the A330 lost in the Atlantic). Seeing the avatar correctly pushing the stick forward would activate the pilot’s “mirror neurons,” which tend to direct the person to imitate what his eyes are seeing. Tests have proved that pilots react more quickly when shown the avatar.


Awareness of the pilot’s mental workload is another important area for progress. A wearable sensor allows Dehais and his team to measure oxygenation of the pre-frontal cortex. The technology–functional near-infrared spectroscopy–yields a representation of the cortex’s activity. A computer in a future cockpit could remove non-essential information when it senses the pilot is experiencing a high mental workload.


In the event cockpit designers want to avoid requiring the pilot to wear sensors, they could rely on seat-integrated heartbeat sensing to understand his emotional status at least. At times of high stress, the system would also use eye tracking to direct essential information to where the pilot is looking.


Much effort is going into alleviating pilot workload–but not by too much. Tests have shown the pilot can be easily distracted, especially via the aural channel, if his workload is too light. Ideally, the level of concentration required should be medium.


Axa has invested €1 million in the project so far. The team wants its headcount, now 15, to grow to 30 as the partnership extends over 20 years and more sources of funding are found.


Selecting Candidates with the Right Stuff


Professor Frédéric Dehais is also developing insights into new ways to select student pilots. Elementary computer tests can provide a fair prediction of how the student will perform in a cockpit. For example, what appears as a simplistic game with colored squares is in fact a measure of the spatial working memory that provides insight into how the person will perform in navigation.


French Engineering Schools Merging


ISAE SupAéro and ISAE Ensica, two well known engineering schools in the French aerospace industry, are completing a merger. “The industry has told us there is no longer a case for two different syllabi or two campuses,” ISAE CEO Olivier Lesbre said.


Starting this year, a single exam is offering 180 places for students. The Ensica name and campus will disappear and the ISAE-SupAéro campus in Toulouse is being refurbished and expanded. The ISAE group includes three more schools: Estaca, Ensma and Ecole de l’Air.


Nowadays, SupAéro students are encouraged to choose subjects such as onboard systems and production, which the industry is expected to need over the coming years for upgrading existing models and increasing production output, rather than aerodynamics–more in demand during fertile times when many new designs are being launched.