Cognitive Daily

i-eca0cf2af9fc3ac4445c7dff7d8aab70-research.gif

i-7d5ea495a1663fa97bec6b4dac4a0da7-melcherdeer.jpgMemory is a curious thing, and visual memory is even more curious. In some ways, we don’t remember much about the scene that’s right in front of us. As countless change blindness studies have shown, we often don’t notice even obvious changes taking place in a scene. Other studies have concluded that visual short term memory has a capacity of just three or four objects.

Yet I have vivid visual memories of scenes I have only glimpsed for a few seconds: A deer below the rim of the Grand Canyon; Michael Jordan draining a three-pointer to win the NBA championships; the standing ovation our daughter received at the school play. If I could only retain three or four of the items in those scenes in short-term memory, why are my long-term memories so vivid? How could they have ever become long-term memories if my short-term memory couldn’t even contain them?

Some researchers have suggested that any additional details in those memories are filled in by verbal descriptions — the “gist” of the scenes, instead of the actual visual imagery. But this explanation doesn’t match up with the rich visual memories I have of some places. Could some other process be responsible for the richness of visual memories?

There is some evidence that visual memory for scenes increases with longer viewing time. But can we remember more than the four items supposedly retainable in short-term visual memory?

David Melcher showed photos of scenes to volunteers for periods ranging from 5 to 20 seconds. But he also added a twist:

Sometimes viewers saw the images uninterrupted, for 5, 10, or 20 seconds. Then they were quizzed on the contents of the images (“What color were the flowers in the pots?” “What was in the view in the background?”). Other times, after either a 5- or 10-second interval, there was no quiz, and the next image was presented, followed by its quiz. Then later in the test, the unquizzed image reappeared for another 5 or 10 seconds, followed by the quiz for that image. So this image had been viewed for either 10 or 15 seconds in total, but in two separate viewings. Here are the results:

i-a7587d3cd1edcbefba3a99a69a1f0ef0-melcher1.jpg

Accuracy seems to be directly related to the total viewing time. For each increase in viewing time, accuracy increased significantly. And look at the 10-second point, where some images were seen in a single viewing, and some were seen in two sessions. There was no significant difference in the results!

On the basis of this experiment, Melcher could calculate the average number of objects retained in memory. After 5 seconds, about 5.1 objects were recalled, compared to 7.0 objects after 20 seconds. And while central objects in a scene were recalled more frequently than peripheral objects, the same pattern held for both types of objects — the longer the scene was observed, the more likely viewers were to remember objects.

But how much time can elapse between the first and second viewings of a picture while preserving the effect? In his second experiment, Melcher systematically varied the time between the first and second viewing. This time, instead of having viewers rate other pictures during the break, they either performed a reading or a visual short-term memory task. There was either a 0-, 10-, or 60-second delay between the first (10-second) and second (1-second) presentation of the picture. Here are those results:

i-f62b38eb6ff7bde7574b857d8e4c0e1e-melcher2.gif

As you can see, there was no significant difference in the results, regardless of the interval between the first and second presentation of the picture or what viewers were doing during the interval — even when they completed a task designed to occupy visual short-term memory. As a control, sometimes pictures were rated after just a single 1-second viewing, and accuracy on the test was significantly lower.

So far from finding a fixed limit to short-term visual memory for scenes, Melcher found that the memory improves after longer exposure to a scene. In previous experiments, Melcher had found that scene memories diminish after a 24-hour period, so it doesn’t appear that they have simply moved into long-term memory.

Melcher suggests that there might be something like a “proto-long-term memory,” not as robust as true long-term memories, but with a larger capacity than short-term visual memory.

Melcher, D. (2006). Accumulation and persistence of memory for natural scenes. Journal of Vision, 6, 8-17.

Comments

  1. #1 Timothy
    August 23, 2007

    Is proposing an intermediate storage or proto-long-term-memory between short term memory and long term memory the best possible explanation? It seems like this effect could be explained without hypothesizing another modular memory storage . For example, memories could be integrated into long-term memory rather quickly but due to lack of consolidation or some other process they are forgotten.

  2. #2 Nikhil
    August 23, 2007

    Memories are the results of human feelings.When a person has an interest in that particular image/scene or is enthusiastic about that then the image memory is said to increase.A pure hypothetic proto-long-term-memory may not be the perfect answer.

  3. #3 Steve
    August 23, 2007

    “Memories are the results of human feelings”
    This is not entirely true. Emotions do interact with memory and improve recall, etc. But memory can and does occur without any sort of emotional (amygdala) involvement.

  4. #4 ThursdayMan
    August 23, 2007

    Hello Dave, Interesting post! I’d just want to point out that your impression of having a vivid memory of a scene could be quite distinct from what you actually remember — that is, the feeling of ‘vividness’ could be an illusion. I can think of a few states of affairs that I think I remember vividly, but if I were asked to draw them from memory, I’d do quite poorly. And people are sometimes quite certain that experimentally suggested ‘false memories’ are in fact real — so they could have a vivid feel as well. Examples in perception abound, such as the feeling that we vividly experience an entire wall of Marilyn Monroe images (Warhol style), whereas in fact only a few are falling on the cones in the retina (and the feeling the rest are ‘in focus’ is..well.. a feeling).

    All right, thanks for the post!

  5. #5 Janne
    August 23, 2007

    There’s some studies (http://www.lucs.lu.se/People/Jana.Holsanova/index.html) that seem to show that when we recall a scene we tend to follow the scan order of the eyes as we looked at the scene in the first place. It’s based on having subjects look at a complex scene, wearing an eyetracker, having a break, then describing the scene they saw. They will tend to follow the same order, and their eyes will go to the corresponding places as they try to recall the scene.

    That would neatly fit with the number of objects retained as a function of time; we tend to return over and over again to the most salient bits in an image, so the number of new objects retained should follow a log distribution, and the most salient objects should be clearer and more detailedly recalled.

    So, no specific memory mechanism is really needed, just a “normal” memory, together with the attentional mechanisms determining gaze direction over time.

  6. #6 avc
    August 25, 2007

    There are other possible explanations for more “complete” visual memories. There is evidence that your brain is processing a lot of the visual information coming in and then it just disregards it when its not relevant, but nonetheless aware of your surroundings.

    Its seems that when something is “tagged” as highly important for your brain like Michael Jordan draining a three-pointer to win the NBA championships or the standing ovation your daughter received at the school play, you tend to focus your attention only on that thing and stop processing the rest of the visual info surrounding you, leaving more “visual memory” open for details.

    This has evolutionary roots. Imagine 50,000 yrs ago when one of our ancestors walking by trying to find a prey suddenly spots another human on the horizon, he needs to focus all his attention to pay attention to details so that he can discern whether the other human is a friend or foe, being able to recognize the facial expressions, the things that person is holding, if there are people coming with him, etc will be essential to make the right choice.

  7. #7 chizadek
    August 26, 2007

    I don’t form vivid memories. The only time I can ‘see’ something that’s not there is when I dream. Why might that be?

  8. #8 bjorn
    August 28, 2007

    Could it be attributed to an evolutionary advantage of not going in circles? The data could be explained by visual data being stored in a (linear) spatial manner, and re-exposure bringing the memory to the “front-of-the-line”. I.e. think of walking along a path and seeing the scenery at point p1. As you keep walking the memory of p1 fades “behind” p2, p3, etc. If you then see p1 again, the memory would come to the front and be augmented by the new exposure. This is different from the regular working/short/long-term memory model and could be described as a proto-long-term-memory. Evolutionary speaking, we’ve been living stationary for a very short time, so it would make sense for our visual memory to be ordered along a time/spatial dimension…

  9. #9 Nigel
    August 29, 2007

    [Sorry this is such a long post (half of it is references, though), but you have got me on my hobby-horse here]

    chizadek:

    The short answer to your question is “nobody knows;” longer answer follows:

    Since Francis Galton published the findings from his questionnaire on mental imagery (Galton, 1880, 1883), psychologists have known that some small minority of otherwise normal people report that they experience little or no visual mental imagery, at least while they are awake. (I say “otherwise normal” because there are quite a few reports in the neurology literature of people who seem to have lost their mental imagery after brain damage. However, Galton’s respondents showed no signs of brain damage – many were distinguished and still active scientists – and I hope I am safe in assuming that chizadek‘s brain is also intact.) Unfortunately, however, although since Galton’s time there has been extensive, if not always very fruitful, research on individual differences in the subjective vividness of people’s imagery, there has been virtually no research into why some healthy people apparently have little or no visual imagery (except, as they often concede, in dreams). Indeed, apart from a descriptive case study of a “non-imager” by Sommer (1978 ch, 7), and a historical case study (by me) of how J.B. Watson was able to convince himself that imagery does not exist (Thomas, 1989), I have not discovered (over about 25 years of being interested in the topic) any other serious published research on the topic of “non-imagers.” There are not even any reliable estimates of their incidence in the population: you will find various percentages occasionally quoted, but none of them appear to have any real basis (see Thomas, 2007 note 21). As to the causes (and symptoms, if any) of someone’s being a non-imager, even informed speculation is hard to come by, and I am pretty sure that the best you will be able to find anywhere is a fairly lengthy post that I made to the Psyche-D online discussion list a few years ago (Thomas, 2001).
    [Incidentally, having mentioned (and linked to) Galton’s famous study, I perhaps ought to mention that one of its best known findings – viz. that scientists tend to have less vivid imagery than other people, and that the “non-imagers” are to be found particularly amongst scientists – has recently been quite persuasively refuted by Brewer & Schommer-Aikins, (2006). Not only has it proven impossible to replicate the original result, but Galton’s analysis of his own data appears to have been faulty.]

    As for Melcher’s findings discussed in the main post, my thoughts as I read the post were very similar to those expressed above by Janne (but then I too have some knowledge of Jana Holsanova’s work). Long ago, Noton & Stark (1971a,b,c; Stark & Ellis, 1981) showed via eye tracking experiments how taking in even a relatively simple visual stimulus, such as a line drawing, involves successively fixating salient points and developing a “scanpath” appropriate to exploring that particular stimulus. More recently, Holsanova and others (Brandt & Stark, 1997; Laeng & Teodorescu, 2002; Johansson, Holsanova & Holmqvist, 2006; and see also Demarais & Cohen, 1998; Spivey & Geng, 2001; de’Sperati, 2003; Bensafi et al., 2003) have shown that, as suggested by “enactive” or “perceptual activity” theories of imagery (Neisser, 1967; Thomas, 1999, 2007 §4.5; Bartolomeo, 2002; Blain, 2006), these scanpaths are re-enacted when someone later recalls a memory image of the stimulus. (For further discussion of this and other related work on eye movements and imagery see here.)

    Melcher’s results seem to fit right into this picture. We do not take in a visual scene all at once (as change blindness experiments also indicate), rather we scan its various details serially; then, when we recall it, we also recall the details serially, and the spatial relationships between these details are encoded in the form of the scanpath we originally established, during viewing, for shifting attention between them. We (or most of us) experience this re-enactment of the original pattern of attention shifting and detail recognition (we presumably also re-enact covert attentional acts, as well as overt eye movements) as mental imagery.

    It is important to recognize that normal visual perception (and mental imagery) is a process that is extended in time rather than happening at an instant. Even tachistoscopic vision seems to depend on a temporary analog representation, so called iconic memory (not the same thing as mental imagery) that, during the brief time it persists, is serially scanned by internal attentional shifts in order to produce the actual visual experience (Neisser, 1967). Vision is not a series of full detail snapshots (like the frames of a movie), but an active, temporal process of exploration of the visible envoronment. By the same token, a memory image is not a visual snapshot that has been kept for the album, but a (partial) re-enactment of some episode of perceptual exploration (Neisser, 1976; Thomas, 1999, 2007 §4.5.1; Bartolomeo, 2002; Blain, 2006).

    From this perspective, ThursdayMan, in privileging current foveal stimulation, has things backwards: we really do see a complete wall full of colored Marilyns (after all, that is not only what we seem to experience, but also what is actually there). The fact that, at any particular instant, light from only a small part of the wall is impinging on the fovea just goes to show that what is currently stimulating the fovea (or the whole retina, come to that) is not the same thing as what we are seeing. Foveal transduction of light quanta is not seeing as such. The eyes are constantly moving to explore the visual scene, and, indeed, if the image focused on the retina remains unchanging for more than a brief moment, we cease to see anything at all (Yarbus, 1967). Foveal (and, come to that, full retinal) transduction is only a small (albeit vital) part of the complex, active, and temporally extended process of visual perception.

    References

    Bartolomeo, P. (2002). The Relationship Between Visual perception and Visual Mental Imagery: A Reappraisal of the Neuropsychological Evidence. Cortex (38) 357-378.

    Bensafi, M., Porter, J., Pouliot, S., Mainland, J., Johnson, B., Zelano, C., Young, N., Bremner, E., Aframian, D., Kahn, R., & Sobel, N. (2003). Olfactomotor Activity During Imagery Mimics that During Perception. Nature Neuroscience (6) 1142-1144.

    Blain, P.J. (2006). A Computer Model of Creativity Based on Perceptual Activity Theory. Unpublished doctoral dissertation, Griffith University, Queensland, Australia.

    Brandt, S.A. & Stark, L.W. (1997). Spontaneous Eye Movements During Visual Imagery Reflect the Content of the Visual Scene. Journal of Cognitive Neuroscience (9) 27-38.

    Brewer, W.F. & Schommer-Aikins, M. (2006). Scientists Are Not Deficient in Mental Imagery: Galton Revised. Review of General Psychology (10) 130-146.

    de’Sperati, C. (2003). Precise Oculomotor Correlates of Visuospatial Mental Rotation and Circular Motion Imagery. Journal of Cognitive Neuroscience (15) 1244-1259.

    Demarais, A.M & Cohen, B.H. (1998). Evidence for Image-Scanning Eye Movements during Transitive Inference. Biological Psychology (49) 229-247.

    Galton, F. (1880). Statistics of Mental Imagery. Mind (5) 301-318.

    Galton, F. (1883). Inquiries into Human Faculty and its Development. London: Macmillan.

    Johansson, R., Holsanova, J., & Holmqvist, K. (2006). Pictures and Spoken Descriptions Elicit Similar Eye Movements During Mental Imagery, Both in Light and in Complete Darkness. Cognitive Science (30) 1053-1079.

    Laeng, B. & Teodorescu, D.-S. (2002). Eye Scanpaths During Visual Imagery Reenact those of Perception of the Same Visual Scene. Cognitive Science (26) 207-231.

    Neisser, U. (1967). Cognitive Psychology. Englewood Cliffs, NJ: Prentice-Hall.

    Neisser, U. (1976). Cognition and Reality. San Francisco, CA: W.H. Freeman.

    Noton, D. & Stark, L. (1971a). Eye Movements and Visual Perception. Scientific American (224 – vi) 34-43.

    Noton, D. & Stark, L. (1971b). Scanpath in Eye Movements During Pattern Perception. Science (171) 308-311.

    Noton, D. & Stark, L. (1971c). Scanpath in Saccadic Eye Movements While Viewing and Recognizing Patterns. Vision Research (11) 929-942.

    Sommer, R. (1978). The Mind’s Eye. New York: Delacorte Press.

    Spivey, M.J. & Geng J.J. (2001). Oculomotor Mechanisms Activated by Imagery and Memory: Eye Movements to Absent Objects. Psychological Research (65) 235-241.

    Stark, L. & Ellis, S.R. (1981). Scanpaths Revisited: Cognitive Models Direct Active Looking. In D.F. Fisher, R.A. Monty & J.W. Senders (eds.). Eye Movements: Cognition and Visual Perception (pp. 193-226). Hillsdale, NJ: Erlbaum.

    Thomas, N.J.T. (1989). Experience and Theory as Determinants of Attitudes toward Mental Representation: The Case of Knight Dunlap and the Vanishing Images of J.B. Watson. American Journal of Psychology (102) 395-412.

    Thomas, N.J.T. (1999b). Are Theories of Imagery Theories of Imagination? An Active Perception Approach to Conscious Mental Content. Cognitive Science (23) 207-245.

    Thomas, N.J.T. (2001). Re: Mental Images Query. Psyche-D moderated listserv (Sat, 22 Dec 2001): http://listserv.uh.edu/cgi-bin/wa?A2=ind0112&L=psyche-d&F=&S=&P=999
    Unfortunately many of the links in this no longer work. Cited works by me (Nigel Thomas) can now be found at http://www.imagery-imagination.com/. Eric Schwitzgebel’s articles (now published) can be found at http://www.faculty.ucr.edu/~eschwitz/

    Thomas, N.J.T. (2007). Mental Imagery. In E.N. Zalta (ed.), The Stanford Encyclopedia of Philosophy (Fall 2007 Edition), Online publication: http://plato.stanford.edu/entries/mental-imagery/

    Yarbus, A.L. (1967). Eye Movements and Vision. New York: Plenum Press.

  10. #10 jbuscher
    October 30, 2007

    Interesting stuff. There’s a great podcast at Shrink Rap Radio (http://www.shrinkrapradio.com) improving memory by psychologist Thomas H. Crook, Ph.D. is the author of the new book, The Memory Advantage: Improve Your Memory, Mood, and Confidence Throughout Life. I thought it was a great interview. Here’s a link to the Mp3 of the show:
    http://www.shrinkrapradio.com/2007/02/28/79-understanding-and-improving-memory/#more-17

    Great website! Keep up the great work.

  11. #11 vanel
    January 22, 2009

    It looks like Russian sciantists have figured out the answer:

    http://www.pmemory.com/improve_memory-adr.php

  12. #12 vanel
    January 22, 2009

    Russian scientists have figured out the answer:

    http://www.pmemory.com/improve_memory-adr.php

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.