Understanding "Perception"...

Discussion in 'General Science & Technology' started by Time/02112, Nov 30, 2000.

Thread Status:
Not open for further replies.
  1. Time/02112 Senior Member Registered Senior Member

    Messages:
    298
    Understanding perception

    Saturday, November 25 2000 @ 04:22 PM EST
    Contributed by: Porfiry
    For most of us, seeing the world around us is such an effortless process that we tend not to give it a second thought. In fact, our vision reflects an incredibly complex feat of bioengineering that outperforms many computer-based systems.

    This process involves grouping together different features of an object to form the whole: nib, barrel and lid form a pen, for example. While it is clear that adults and older children can do this, until now, it has been difficult to determine when and how this vital skill develops in young infants--mainly because they can't tell you whether they see a pen or simply a collection of unrelated elements.

    Now, using a brain imaging system known as the Geodesic Sensor Net and a test made up of shapes like the "Pacman" from 1980s computer games, scientists at London's Birkbeck College have shown that eight-month-old babies do indeed bind attributes to form a whole object. Their results are published 24 November 2000 in the journal, Science.

    The test involved showing the infants a group of four Pacman shapes collectively known as the Kanizsa Square. When placed with the four "mouths" facing inwards, the Pacmen give the visual illusion of a square, which is not actually there. Previous work in adults has shown that the perception of the square correlates with a burst of brain activity known as gamma oscillation.

    When six-month-old babies were shown the square, the characteristic signals were not there. However, when eight-month-old babies were shown the same figure, gamma oscillations were seen.

    Lead Science author Gergely Csibra said: "Understanding how an infant brain develops is obviously fascinating and may have implications for the education and care of babies. This new work not only tells us that babies as young as eight months recognise complex objects in the same way an adult does, but also allows us to think of new studies into early infant development.

    "The difference between six- and eight-month-old babies is also intriguing and may show that there is an important development in how the brain organises information from the outside world at that age."

    In the study, 11 six-month-old and 11 eight-month-old infants were shown either the Kanizsa Square or another collection of Pacman shapes on a computer screen. The brain activation was detected by the Geodesic Sensor Net-which fits over the head like a shower cap and can detect the gamma oscillations if they are present. The Geodesic Sensor Net enables researchers to study the activity in babies brains using a safe and child-friendly method. The device gently rests a number of passive sensors on the child's head, which are able to detect the minute changes in electrical fields which happen as groups of nerve cells are active together in the brain. It is a variation on a method called EEG that has been used routinely in hospitals for several decades.

    Mark Johnson, director of the research laboratory said: "This research provides us with a powerful new tool for investigating how babies think when presented with everyday objects, toys or faces. While we have only studied healthy babies so far, it is also possible, but not proven, that we may also be able to study babies which are, unfortunately, not developing typically."

    In addition to Drs. Csibra and Johnson, the team included Greg Davis and Mike Spratling, all of whom work at the Centre for Brain and Cognitive Development, School of Psychology, Birkbeck, University of London. The research project was funded by the UK Medical Research Council.
    ------------------------------------

    Is it possible to alter your perception of time? .........
    Below, are additional links from archived posts that correlate to understanding the variences of "Perception" http://www.xone.net/tti/board/ubbhtml/Forum1/HTML/000316.html http://www.xone.net/tti/board/ubbhtml/Forum1/HTML/000256.html http://www.xone.net/tti/board/ubbhtml/Forum1/HTML/000251.html
     
  2. Google AdSense Guest Advertisement



    to hide all adverts.
  3. Crisp Gone 4ever Registered Senior Member

    Messages:
    1,339
    Hi Time/02112,

    Very interesting article indeed, but I find it curious that neural networks are not mentioned here. At the moment neural networks have been extensivily studied and they provide quite an accurate description of how the process of "perception" works (after all, perception is nothing more than remembering images and that's one of the tasks neural networks are capable of).

    Just a few weeks ago I did some research on very simple neural network models (as a student I am supposed to do this kind of stuff to earn some grades

    Please Register or Log in to view the hidden image!

    ). It is really amazing to see how even the simplest neural network models, such as the Hopfield model, can tell the difference between two images or plainly deduce an image from its distorted original. (The Hopfield model is a neural network where all nodes can either be "on" or "off", which is a serious simplication of the human neural system).

    Even the process of storing images in memory can be covered by the maths behind neural networks, and this is the reason why I am surprised they never get mentioned here.

    Anyway, if you find this story interesting, I highly recommend doing some reading on neural networks. Some good links for a quick tour might be:

    "What is an artificial neural network?"
    -> http://www.dtek.chalmers.se/~d95graag/ann_explain.html
    A very simple introduction to neural networks. Explains how neural networks can remember images.

    "Memory - The Hopfield model"
    -> http://suhep.phy.syr.edu/courses/modules/MM/Neural/neur_hopfield.html

    "Neural Nets"
    -> http://www.shef.ac.uk/psychology/gurney/notes/index.html
    A more mathematical approach to neural networks. Recommended if you have some experience in physics (it uses statistical mechanical methodes).

    Bye!

    Crisp



    ------------------
    "The best thing you can become in life is yourself" -- M. Eyskens.
     
  4. Google AdSense Guest Advertisement



    to hide all adverts.
  5. DaveW Registered Senior Member

    Messages:
    243
    Please don't copy articles from the main page. Provide a link if you have to.
     
  6. Google AdSense Guest Advertisement



    to hide all adverts.
  7. Time/02112 Senior Member Registered Senior Member

    Messages:
    298
    DAVE W.
    Well before we get into a heated debate about if we are, or not allowed to post specific things here in this forum, may I remind you that this type of censorship will brin a lot more unpleasent expressions from others who do th SAME THING?

    If your going to make rules like this to just one isolated posting, why not make it official, and tell everyone else here the same thing?
     
  8. DaveW Registered Senior Member

    Messages:
    243
    Censorship!!!

    This is ExoScience. It seems rather redundant to post an article from the ExoScience page onto the ExoScience message board, as if no one here had seen it. Just post a link in the future, please.
     
Thread Status:
Not open for further replies.

Share This Page