Aug 27 2009

Humans Do Not Have Multi-Core Processing

Despite the siren call our our increasingly multimedia, multitasking society, humans seem to behave like old computers with a single-core processor, rather than the latest PCs with multiprocessing.

This is not literally true – our brains are massively parallel, and they can process multiple kinds of information simultaneously. This is why I wrote above about human behavior and not brain processing. To clarify the distinction – our brains can parallel process, but we can only pay attention to one thing at a time. So while multiple subconscious processes are going on at the same time, that part of our brains that pays attention, that manipulates information in our working memories, can only focus on one thing at a time. This stream of attention has peripheral awareness of other things going on, and information on the periphery of our awareness can grab our attention, of course. But there is a certain threshold for this, and it is higher than one might naively think.

But a number of independent lines of research all lead to the same conclusion – we can’t actually pay attention to multiple things at the same time. We may be able to switch back and forth, but the attention any one item gets will suffer.

And yet, people who multitask generally think they are good at it. They are deceiving themselves, it turns out.

Researchers at Stanford University defined two groups of subjects – those who frequently multitask and those who do not. They wanted to find out what skills the multitaskers have the non-multitaskers do not. They tested the two groups with three standard psychological tests of attention. The first test involved showing subjects blue and red rectangles and then having them determine on a subsequent showing if the red rectangles changed orientation – this tests the ability to ignore irrelevant information. The second task involved showing subjects a series of letters and then hitting the button when they see a letter they had seen exactly three letters previously – a test of memory organization. And the third task involved being shown letter/number pairs and then being asked at random to either determine if the letter was a vowel or if the number were odd/even – a test of the ability to rapidly switch between tasks.

This turned out to be one of those wonderful experiments where the experimenters find exactly the opposite of what they expected – the high multitaskers performed significantly worse on all three tasks. This was especially surprising because the researchers chose these tasks based upon the hypothesis that they involved skills useful for multitasking. Further, the high multitaskers thought they were good at multitasking while the low multitaskers thought they were poor at it.

This study reinforces the conclusion that people easily delude themselves. It does not show we are bad at multitasking – it was not designed to test that. Rather, it shows that people who multitask either do so because they think they are good at it when they are not, or they do it because they have to or want to and then rationalize their multitasking by deluding themselves into thinking that they are good at it.

But why are high multitaskers worse at standard tests of attention? This needs further exploration. Perhaps multitasking dulls our mental abilities – we lose the ability to focus, to filter out extraneous information, and to quickly shift our attention to a necessary task. Or – perhaps high multitaskers do so because they naturally have poor attention skills, or don’t recognize the negative impact multitasking has on their performance. Maybe low multitaskers are better able to detect the deleterious effects multitasking has on attention and performance.

Perhaps this should not have been as surprising at it was. We know, for example, that even something as simple as talking on a cell phone has a significantly negative impact on driving performance. It also turns out that this is not related to having one hand on the cell phone, but the fact that the conversation diverts our attention from the road.

We also know that our attention is much more limited than we naively feel. There is, for example, a dramatic phenomenon known as inattentional blindness. Here, for example, is Richard Wiseman’s color changing card trick. There is also this video. Take a look (prior to reading the comments).

The lesson is – don’t multitask. Do not text while you are walking on the street or (ack) driving your car (yes, people do this). Don’t read your e-mail while having a conversation (I admit, I am guilty of this one.) Or (if you are George Costanza) don’t eat, watch TV, and have sex at the same time. If you need to perform a task well, give it your full attention.

Share

16 responses so far

16 Responses to “Humans Do Not Have Multi-Core Processing”

  1. jonny_ehon 27 Aug 2009 at 10:25 am

    Whenever the issue of multi-tasking and cell phone use while driving comes up, I always wonder what it is about cell phone use that is so dangerous as opposed to chatting with a passenger. From reading this article, I now wonder if it’s because with a passenger, it’s easier to switch attention from the conversation back to the road more easily since they see that you’re busy driving and understand if you ‘leave’ the conversation for even a moment.

    Has there been a study on people’s driving behaviour while chatting with a passenger? There’s been lots on cell phones, but how do drivers act with a passenger there? How often does the conversation halt? Whenever something like a close call happens, does the driver point it out to the passenger?

  2. mccorvicon 27 Aug 2009 at 10:33 am

    This also reminds me of those people who say, “Oh, it’s ok. I actually drive better when I’m drunk because I know I’m drunk so pay more attention.” I thought I was the only one who heard someone say this, but I’ve even heard people on TV make jokes about these people.

    It’s that self-delusion thing. People are so silly.

  3. Zelockaon 27 Aug 2009 at 10:49 am

    Well we do multitask to a point in that you brain regulated and does a number of functions at one but all of these are done without direct mental supervisions. For tasks you pay attention to directly no there is no multitasking. It’s the same thing as rubbing your belly and patting your head. Unless you do it enough to for it to no longer need direct runtime you will mess up the patting or the rubbing. Luckily if you do muscle actions enough they tend to fall out of the direct runtime category (or else we could never do anything while walking for example and even then you will sometimes trip).

    For the most part when people talk about multitasking (at least it being successfully done) they are really talking about series tasking where a person moves from one task to the next but only attempts actions on one task at a time.

  4. banyanon 27 Aug 2009 at 10:52 am

    I always hear that women are better at multi-tasking than men. Is it safe to say that that’s nonsense now?

  5. Zelockaon 27 Aug 2009 at 11:06 am

    Well officially the study is too general to give a full response to this question. The following factors may eventually be proved to have different conclusions

    Tests tended to be short term memory based. For most tasks that you don’t really need to pay attention too you don’t really need that much memory of them after the fact. This study addresses some of that issue.

    http://www.livescience.com/health/080428-working-memory.html

    Study was too focused on display items (text and shapes). Test requires direct attention to translate so other mediums (sound, physical interaction) could both produce different results.

    Study was focused on people that identified themselves as muti or single tasking. Would have been better controlled if that was determined from outside observation rather than personal reporting.

    Unless I miss something there were no direct controls for age or sex. Both may affect outcome.

  6. TBRechon 27 Aug 2009 at 12:18 pm

    jonny_eh, I wondered the same thing and asked this on comment board dealing with the brain. I was not completely satisfied with the answer.

    I rarely use my hands free cell phone while driving, but do talk to a passenger. As a skeptic, I would like to see a study looking at the effect this kind of conversation has on driving concentration also. Now, I am talking about normal everyday conversation with an adult. Consider so many people drive around with children. What kind effect do children as passengers have on driving? Let the researchers compare any impairment with that of hands free cell phone conversation.

    Continuing on the critical thinking about this subject, if multitasking is not truly possible, then, how do surgeons, EMTs, paramedics and other trauma providers function in an environment when attention is quite divided? How can any given officer of the law drive a patrol car, talk on the radio, and use the in car computer safely? I suggest to those who have not had the opportunity to do a ride along with your local law enforcement agency yet, that you look into doing so. I believe you will be impressed by how people do multitask in real life.

    So, more studies are needed exploring this subject as far as this skeptic goes. Of course funding can be an issue, but maybe an undergrad student or grad student could find a creative source or do the leg work needed to gather enough stats to be statistically significant. I suspect most law enforcement agencies will have crash data for their own staff. They might consent to sharing this information and even work with the academic investigators to gain insight to the safety of their officers.

  7. DevoutCatalyston 27 Aug 2009 at 1:06 pm

    “…It’s the same thing as rubbing your belly and patting your head. Unless you do it enough to for it to no longer need direct runtime you will mess up the patting or the rubbing…”

    Is this the same as playing the piano and singing at the same time? Practice, practice, practice finally allows perfect? It’s a difficult thing to have both hands doing different lines, and then sing on top of that, add spice, and do all three with feeling.

    Might explain why so many merely strum a guitar (or pretend to) and sing — much less demanding.

  8. HHCon 27 Aug 2009 at 3:47 pm

    As the last reference was to an old Seinfeld episode, you can’t multitask as suggested if your sex play involves constriction of the air passage. In fact, when your “computer lights” go out so does the processing.

  9. artfulDon 27 Aug 2009 at 7:48 pm

    Sometimes the lights are still on but the bulb is just too dim.

  10. weingon 27 Aug 2009 at 8:27 pm

    I can listen to music and read without a problem, but it has to be without lyrics. I find lectures very irritating when the lecturer shows slides and reads them, but enjoy it when he displays a picture and describes what I’m seeing. It seems that parallel processing is easier only with inputs from different senses.

  11. Bryceon 27 Aug 2009 at 8:49 pm

    The appropriate comparison would be to contrast brains with computers using channel I/O (http://en.wikipedia.org/wiki/Channel_I/O). There are parts of the brain with more specific functions that can be relied on to perform properly without “attention”. This could be used as a blunt rule to determine if some tasks can be done in parallel: if you can do it without thinking, you can probably do it when thinking about something else. Of course, this would take some self-honesty to evaluate when your doing a task poorly, something apparently lacking with many people who are on the phone and driving dangerously at the same time.

  12. taustinon 27 Aug 2009 at 10:59 pm

    Studies have been done on talking to a passenger while driving, and it is distracting, but it turns out to be more complicated. The distraction is somewhat mitigated by the passenger providing a second set of peripheral awareness in the car, where a cell phone conversation doesn’t. In other words, if the driver is distracted by the conversation and missed the truck he’s about to rear-end, he’s distracted by paying attention _to the passenger_, who might not miss the truck. So the driver has two chances to avoid the collision; he might notice the truck, and he might notice the passenger bracing for the collision.

  13. LeeBon 28 Aug 2009 at 12:02 am

    Stats on teenager drivers indicate a significantly higher rate of accidents when other teenagers are in the car with them. I’ve never seen a car of quiet teenagers carefully attending to the road, so my guess is that they are talking and quite distracted in the process.

  14. Steven Novellaon 28 Aug 2009 at 8:34 am

    Regarding passengers – they are distracting, but not as much. You probably don’t have to pay as close attention to what is being said, because you have visual clues as well. But also, as someone pointed out, the passenger is another set of eyes and can shout “look out” if you miss something.

    Regarding the examples of multitasking (surgeons, EMTs) – they are multitasking – my premise was not that people do not or cannot multitask, but rather that people only have one attention – one set of working memory. If they divide their attention, their performance suffers.

    However, the focus of your attention can be more diffuse or more focused. And, you can get good at doing several things together by rapidly shifting your attention from one to the other – but your attention on any one thing will suffer.

    I also think we need to distinguish this from integrating multiple sensory modalities. Clearly we are capable of doing that, but then they combine to one overall experience.

    The bottom line is that we are not like Data from Star Trek – remember the episode where he said he was listening to a musical composition, reading an article, and working on a mathematical problem all at once. He is multi-core.

  15. HHCon 29 Aug 2009 at 12:15 am

    Multi-tasking may be effective in different modalities. For example, you could talk on the phone (requiring listening skills and semantic processing), pay bills/ write checks (repetitive hand motor skills), and watch t.v. (visual information processing) at the same time. One modality may take more of your attention and the others less. But, you have stimulated your mind to complete repetitive daily tasks. So more “Data” not less makes you a happy trekkie through life.

  16. jmoraon 05 Sep 2009 at 7:38 am

    I think I may have read somewhere long time ago that intellectually gifted people are capable of focusing on two things at the same time, this is, they are really multicore. Do you know something about this? I’m having a hard time trying to find about it.

    For sure multitasking affects attention and the hability to focus, but I think the tasks in the study are not correctly chosen (IMHO). It’s a little weird to conclude that multitaskers are worse at multitasking, isn’t it. I’m thinking a study with two different groups:
    - one group would perform IQ tests on paper (or similar problems that require some time to be solved) with questions of different types mixed, while a second program pops different questions with a time span to be answered.
    - other group would perform the test with questions ordered by type, and then the other questions with the same time span but appearing one after the other instead of popping randomly.

    The time span would force people to multitask (as when you have to reply to a mail, phone, colleague, etc) without forcing the context change at a too specific time, something that mails and phones don’t do, and good colleagues either. Multitaskers and non-multitaskers (unitaskers?, monotaskers?) would be distributed in both groups and you can think about the rest.

    @banyan I’ve heard the same thing, I would like to know the source of that claim though. Maybe it means that women are better at automating tasks and making them unconsciously, I don’t know.

    @weing the same happens to me, music helps to break the monotony of reading and having an extra sensory input helps to stay awake, or I think so, this is completely subjective. The same goes for writing and similar tasks.

    @Bryce I think the analogy of human consciousness with a processor should not be taken too literally, but if we are going to do so, the study is not about those processes that don’t require the processor, as DMA, but about how multitasking affects humans, since they are not multi-core, drawing (for me) the conclusion that preemption is bad: http://en.wikipedia.org/wiki/Preemption_(computing) and that interruptions that cause process changes should be avoided. I do not agree that multitasking without preemption and without interruptions is bad, although it should also be avoided when possible.

    On a side note, I stand astonished about how useful some computer science models may be to explain processes that happen in the real world, including the part of the real world that is inside our encephalon.

Trackback URI | Comments RSS

Leave a Reply

You must be logged in to post a comment.