Stress is normal for the 5,500 scientists and engineers at the Jet Propulsion Laboratory. They know that whenever they make a decision, even the slightest error could have serious consequences.
Memories of 1999, after all, are still fresh. Eight years ago, when the Mars Polar Lander space probe entered the atmosphere of the red planet, radio contact was suddenly lost. The satellite simply disappeared from the screens at the control center. Four hundred million dollars had vanished into silence.
The two managers in charge of the project were convinced that they would be fired without further ado. "That's how we deal with errors in our culture," says Markus Ullsperger. But this time, the managers were spared, Ullsperger, a brain researcher at the Cologne-based Max Planck Institute for Neurological Research, recounts. "And it was a good decision," he says. "After all, millions had been invested in their training and education."
From the standpoint of neuropsychology, this was an excellent management decision. Errors, Ullsperger is convinced, are in fact one of the most valuable sources of knowledge. "A man's errors are his portals of discovery," Irish writer James Joyce once said, anticipating a conclusion modern neuroscience has now confirmed.
Ability to Detect its Own Errors
Ullsperger, like a dozen other research teams around the world, is currently studying how the brain tracks down and processes its own errors. "Our brain has the fascinating ability to detect errors and, if they have already occurred, to learn from the experience," he explains.
"Error-related negativity" (ERN) is a concept that has captivated the scientific world. It refers to a characteristic wave of voltage beneath the skullcap, which can be measured whenever the brain detects that an error has been made. Especially surprising is the fact the ERN signal already begins to flicker even before a person is aware of his error.
In the early 1990s, Michael Falkenstein, a neurophysiologist from the western German city of Dortmund, observed for the first time how voltage declines by at least 10 microvolts in a specific group of nerve cells, and that this occurs only 100 milliseconds after a person has made an error -- about the time it takes for your cursor to respond to a click of the mouse.
Falkenstein's discovery marked the beginning of a period of systematic study of the brain's fine-tuned error detector. It paved the way for fascinating new theories on questions such as why compulsive disorders occur or why some people hesitate while others make confident decisions. It also shines a new light on the development of addiction.
Suddenly it becomes clear why a person can often avoid making a certain mistake based purely on gut feeling. "The experiences of the error system provide precisely that subconscious knowledge on which intuition is based," explains Ullsperger.
Vaguely Uneasy Feeling
The error system acts in two ways. First, it intervenes in a corrective way when a person has committed an error. But it also has a warning capability. When it recognizes that an action may not lead to the desired outcome, this recognition is expressed in a vaguely uneasy feeling.
Ullsperger and his fellow researchers plan to find out exactly how this works, using a functional nuclear magnetic resonance (NMR) scanner. Research subjects lying in the NMR tube undergo simple tests, such as the Eriksen-Flanker task, a common and well-known tool of neuroscientists. In the test, rows of letters, like SSHSS, SSSSS or HHSHH flicker in front of the subjects' eyes. They are then asked to press one of two buttons: the left button if the letter in the middle is S and the right button if it is H.
This isn't as easy as seems. The letters to the right and left of the main letters confuse the observer. Especially when they are given only a limited amount of time to perform the task, the subjects frequently correct their answers a few moments later. "They behave the way we do when we misspeak, notice the error and then quickly correct our sentence," says Ullsperger.
Electrodes in a rubber cap on the subject's head measure the typical ERN waves flickering through the brain during this process. Meanwhile, the NMR scanner observes the area of the brain in which nerve cells are especially active.
Stops Producing Dopamine
The method makes it possible to replicate the anatomy of error detection. What it reveals is that immediately following the ERN wave, the midbrain suddenly stops producing dopamine. This neurochemical signal is transferred to the basal ganglia and thus into the limbic system, in which emotions are generated.
Researchers have also discovered another nerve cord involved in error detection. It leads to a deep section of the cortex, which then broadly distributes the signal in the cerebral cortex. "This cascade sends the following signals to the executive positions: Stop, something is going wrong here! Check again and, if necessary, correct immediately," explains Ullsperger.
Drinking the Inner Voice into Submission
The Cologne-based neurologist can also demonstrate that subjects who have made a mistake in the Flanker test take more time for their ensuing responses. "People change their decision-making strategy," he says. "They begin to learn from their errors."
But what does the drop in dopamine production cause? What triggers the entire chain of signals? Ullsperger's explanation is that whenever the brain decides to take a specific action, it simultaneously develops an idea of the expected consequences. If the desired result occurs, the brain rewards itself with the feel-good hormone dopamine. But if something unexpected happens, the reward is withheld -- a form of self-inflicted punishment.
Human perception is highly specialized to notice contradictions between expected and actual occurrences. An ensemble of at least 1,000 nerve cells appears to be responsible for this ability to compare desire and reality.
"It's truly astonishing, but the brain performs these difficult calculations online, that is, constantly, while dealing with many other things at the same time," says Richard Ridderinkhof, a neuroscientist at the University of Amsterdam. He compares this process with the actions of a driver whose vehicle is gradually veering off course. "Without giving it much thought, the autopilot in the driver's head corrects the vehicle's direction of travel."
Ridderinkhof is convinced that these discoveries could also provide valuable information in the field of disaster research. Airplane crashes, for example, are usually attributable to human error. The meltdown at the nuclear power plant in Chernobyl revealed, in a horrific way, how susceptible the human cognitive network is. The crash of the Space Shuttle Columbia, perhaps the most thoroughly studied accident in the age of high-tech, was the result of an even greater failure -- that of an entire institution.
So far scientists have inventoried, categorized and analyzed errors. They have found that anyone who places too much confidence in technology is at risk of failure. Another cause of error, scientists conclude, is a combination of poor preparation and stress. Unresolved organizational issues, such as the ones that doomed Robert Falcon Scott's legendary expedition to the South Pole, can also spell the downfall of a mission.
In many cases there is a fine line between a disaster and the discovery of an error. The biggest accident in the history of civil aviation is a case in point. In March 1977, two jumbo jets collided on the runway at the airport in Tenerife, one of the Canary Islands. The cockpit voice recorder precisely documents the seconds leading up to the crash.
Sense of Foreboding
A Boeing 747 operated by Dutch airline KLM was standing on the tarmac, ready for takeoff, while a Pan Am jumbo jet blocked the runway. Dense fog made visual contact impossible.
The tower had assigned a specific runway path to the KLM jet, but the impatient captain misinterpreted the Air Traffic Control instructions to mean that he was cleared for takeoff. As accident investigators listening to the voice recorder later learned, Willem Schreuder, the flight engineer on the Dutch jumbo, asked the captain: "Is he not clear (of the runway) then, that Pan American?" A sense of foreboding had apparently flickered through his consciousness. But the captain brushed aside the engineer's timidly expressed suspicion and began the takeoff. When he saw the other Boeing appear in the fog, it was too late to correct his fatal error, and 583 people paid the price.
The engineer's inner voice could have saved their lives. He probably wasn't completely aware of the error they were about to make. In any event, he was unable to articulate his hunch clearly enough. "We should, in fact, often give more credence to intuition," says psychologist Ridderinkhof.
An experiment revealed to Ridderinkhof why this is so. In the experiment, a bright light would periodically appear on a monitor, sometimes on the left side of the screen and sometimes on the right. Ridderinkhof asked his subjects to always direct their gaze to the side where the light did not appear. During the experiment he measured their subjects' pupil movements to determine whether they were following the instructions.
Denied Making Mistakes
Ridderinkhof knew that the curiosity of the human brain is far too great to simply ignore a signal like the light in the experiment. In fact, the subjects kept making mistakes, but then they corrected them and improved their performance over the course of the experiment. As expected, the typical ERN wave traveled through the cerebral cortex.
But when they were questioned afterwards, the subjects denied having made any mistakes. In other words, their consciousness had not been informed that the brain had recognized and then corrected the errors. For Ridderinkhof, this suggests that a large part of error processing occurs in the subconscious. Like Ullsperger, he too suspects that he has tracked down the neuronal correlate of intuition -- that inner voice that protects people from errors.
In their experiments, researchers routinely notice that this correction system is set to varying levels of sensitivity in different subjects. Could it be that hesitant people are simply afraid of errors, while the confident have a relatively dull error warning system in their gray matter?
The pathological extremes at both ends of human decision-making behavior offer possible answers to this question. Max Planck researcher Ullsperger also performed his error experiments with people who wash themselves obsessively or have other forms of obsessive-compulsive disorder. His conclusion is that "their monitoring system is so powerful that they can hardly focus on anything else but to constantly monitor themselves."
Cocaine Helps Decision Making?
A similar picture emerges at the other end of the decisiveness scale. Ingmar Franken, neuropsychologist at Erasmus University in Rotterdam, had cocaine addicts who had been clean for at least one month take the Eriksen-Flanker test. "It wasn't just that they often made the wrong decision," says Franken, "they also didn't notice their errors and, more important, they didn't change their strategy."
Franken believes that this could explain why cocaine addicts are so blind to the negative consequences of their own addiction. "Besides, the attraction of cocaine could in fact be that it improves decision-making ability," he says.
Franken's Amsterdam colleague Ridderinkhof obtained similar results in experiments with alcoholics. "Once the alcohol has clouded the brain, the error wave is absent," he says.
Translated from the German by Christopher Sultan