Predicting Human Errors From Brain Activity

Hugh Pickens writes "Researchers report that brain activity can be used to predict the likelihood of someone making an error about six seconds in advance, with gradual changes starting as much as 30 seconds ahead of time. The team used an imaging machine to scan the brains of a group of volunteers who performed a task in the presence of distracting information. When performing correctly the volunteers' brains showed increased levels of activity in those parts associated with cognitive effort, as would be expected. However, these areas gradually became less active before errors were made and at the same time another set of regions in the brain became more active. These regions are part of a so-called "default mode network" and show increased use when people are resting or asleep [PDF]. While imaging machines are far too big and complex to be used in workplaces to monitor the brain activity of people engaged in important tasks, the team hopes to correlate errors to changes in electrical activity in the brain with electroencephalography (EEG), using electrodes placed on the scalp. If EEG features can be found that correspond to the change in brain activity, then a hat that gives warning of an imminent mistake might one day become reality. We've previously discussed similar studies of brain activity."

I don't believe it

I've seen some stupid research in my time, but actually believing you can predict errors in advance is like saying you can predict which leaf will fall off of a tree in the next 6 seconds.

I agree

The parent post makes a very good point. I agree with him wholeheartedly. You may be able to predict actions, but how can you predict whether an action will be an error?

Mod parent up!

Re:I agree

Like the error of forgetting to click "Post Anonymously" when making a post agreeing with yourself.

thats a Bad Analogy, Guy!

oh wait.

Genius, he has it

No no, you see, he was proving that you can't predict mistakes before they happen. If someone could, why would they allow him to agree with himself? He's... taking one for the team, if you will.

Re:I agree

Parent makes sense intuitively, but there are parts of the brain that are very sensitive to conflict; the anterior cingulate cortex (ACC) is one of these.

If you have ever take part of a Stroop test, your ACC has been activated. In a Stroop test, the word for a color is printed in a different color i.e. the word green is shown in the color red. A participant is asked to say either the word or the color. As the speed of doing these discriminations increases, so do errors; interestingly, cognizance of errors is nearly instantaneous, however. You know that you made an error, even before the neural circuitry committed to speaking the words has finished forming the words.

The ACC becomes more active in Stroop tests because Stroop tests cause conflicts in two neural circuits. The ACC arbitrates these circuits. Therefore, an increase in ACC activity (which will happen in advance of the error occurring) correlates with an increase in likelihood of mistakes...more in-depth research and some algorithms (I haven't RTFA) means that an error can be predicted, but of course, not with 100% success.

Re:I agree

Therefore, an increase in ACC activity (which will happen in advance of the error occurring) correlates with an increase in likelihood of mistakes.

I get distracted and make mistakes when I'm watching basketball games too.

Re:I agree

As someone who has taken too many of these tests, the problem ends up being your mind gets bored with the whole exercise withing minutes... these test take hours. After about 10 minutes at MOST, your brain starts to wander. You can NOT focus on the test no matter how much mental effort you put into these. It simulates situations like driving a car where actions and attention become so repetitive that there is an autopilot mode that kicks in. Bit in the same regard it can't be applied logically to situations like being in a boxing ring, because in those circumstances your brain is constantly adjusting to vectors that can't possibly be predicted, therefor always being up to paying attention at the task at hand.

Re:

But there are some subset of errors that are known to be errors when they are performed. Your analysis is for a certain type of error that, realistically, can't be predicted. So why bother. Here's the usage scenario for this type of "error" prediction..

If you're falling asleep...

If you're falling asleep on the road, you're more likely to make the error of driving to one side. Similarly, if you're getting distracted by a stray thought, your concentration can suffer and are more likely to make errors in your primary task. You may

Re:I don't believe it

I don't know. This research is more like predicting whether leaves will fall in six seconds... when it can tell a big breeze is six seconds away.

From TFS, it sounds like people are getting distracted and bored doing stupid mind-numbing tasks and when they do so, they make errors. As such, they have invented a bulky and expensive way to tell when you're drifting off (and that is fairly well correlated with making errors.)

Re:

Exactly. The problem with getting bored / distracted isn't so much that you make errors, but that you don't usually notice until after you start making mistakes -- but the distraction is already clearly present, and I see no reason it shouldn't be detecta

Re:

Basically this particular task (the "flankers task") is so easy that people only make errors when they stop paying attention. You see something like SSSHHSSS or HHHSSHHH and you have to respond to the central letters while filtering out the outside letter

Put electrodes on your scalp to detect errors?

Given what it's supposed to tell about people, let me be the first to dub this device "the asshat".

And I certainly hope it never hits the market.

Re:

I want one when talking to women.

an idea...

Will they ever be able to make a hat that lets slashdot users know if they will be modded '+1 funny' vs '-1 flamebait?' 6 seconds before clicking 'submit?'

That's good news!

While imaging machines are far too big and complex to be used in workplaces to monitor the brain activity of people engaged in important tasks...

Thank God!

W7

"You are about to make a mistake. Cancel or Allow?"

The "error detection" hat may be misinterpreted.

The descriptions of "the error-detecting hat" look like the intent is to train the brain to stay alert and not make errors - or to refocus it when it wanders. (DING DING DING! HEY! WAKE UP! PAY ATTENTION!) The hat may be useful, but that use of the feedback may be the wrong approach.

The signature they're describing corresponds, not just to a lack of alertness, but specifically a lack of alertness because the brain is going into a resting state. Seems to me that might be because all this decision-making has made the working regions of the brain tired and the brain is trying to clear them out so they'll operate properly again. So the problem is not the lack of alertness, but the attempt to continue to make decisions during the resting cycle.

Given that, a better use of the feedback might be to tell the wearer that it's time to stop making important decisions and take a break, rather than trying to overuse a "mental muscle" that's exhausted - and perhaps train him to recognize the mental state himself so he can then dispense with the hat.

The "break times" in working days were set up when studies showed that taking breaks, despite the "work time lost", resulted in more and better work in the work time remaining. This looks like a way to optimize the process, rather than running breaks on a clock.

Over-Reaching a Bit

I have to agree with other people who've posted that this is over-reaching a little bit. This is nothing more than a correllational study: all they can say is that these systems, in this particular series of trials, showed increased activity shortly before an error took place.

The main problem as I see it is that since they can't determine causaltiy, and only conducted this experiment with a small sample population, and with a specific task, is it could have been the task itself causing the particular regions to become active after a certain period of time. I just gave the article a quick look through, but I'd be curious to see if the errors came in distinct, set intervals. It could be simply the nature of the task that caused the activity. Furthermore, what about left handed participants? What about age groups outside of the twenties (which are a particular cohort, and can be expected to have similiar results/activity as such)? It seems like they failed to counterbalance either their participants or their trials in any meaningful way.

Also, I'm not familiar with this journal or whatever it is, but I've never seen one where the methods section came last, which is a little strange. That's almost always the first thing I go to after the abstract.

The Head of the Microsoft Vista development group

was wearing an in-house version of this hat before he decided to release Vista. Unfortunately, the hat was running Vista.

I Can Do This, Too, Sometimes

I've noticed that before I make certain mistakes, or completely forget what I was doing, I will experience a spike in confusion. Sometimes, the ramp up the spike is slow, taking most of an hour; other times, it comes on me suddenly, taking only a minute. I feel confused and doing simple tasks are difficult. Then, it passes. Often, I will have completely forgotten what it was I was attempting to do, until much later when something reminds me.

EEG, Prediciting and Probability

"Prediction" is not accurate because that implies an absolute. The activity correlates with an increased probability of making a mistake. The study relies on statistics throughout, from the analysis of the fMRI data on, and so can only deal in probabilities.

An Israeli team found that an increase in degree of synchronization of midline frontal theta EEG varied inversely with the probability of making a mistake. Such theta synchronization occurs over spans of 10 to 30 seconds. They also found that when a response occurs during the rising or falling slope of the synchronized theta (as opposed to near a peak), the person was more likely to make a mistake. The latter probably is the source of the evoked potential called the Error Related Negativity; it is the brain preparing to notice the error. The former seems to indicate a lagging in attention, which is when errors are most likely to occur. The two are related, meaning the brain "knows" when it is starting to droop and is more likely to make a mistake, and tells itself to get ready to notice a mistake if it happens.

Re:

It sounds to me that the key finding is that they can detect when a person is distracted by monitoring their brain waves. A much more interesting finding than that distracted people make errors.

Re:Errors?

The device actually predicts when you are not concentrating on your work, and that is what people would be told if they ever were to use such a thing in production. I think the notice that you are losing concentration would probably be enough to get you concentrating again on your work. Especially if the workers were penalized for time spent not concentrating.