jump to navigation

Zeki on Colour&Motion Systems of the Brain and Visual Consciousness December 28, 2006

Posted by sumesh in Art, Cognitive Neuroscience, Cognitive Science, Colour, Consciousness, Talks, Vision, Zeki.


Prof. Semir Zeki’s inspiring talk (Nov21,2006. India), with many insightful quotes from the generally ignored footnotes of Kant’s Critique of Pure Reason(1781/1787), was about the colour and motion systems of the brain and the implications, of his research on the visual brain, for understanding visual consciousness.

Zeki is known for his work on anatomical and physiological studies on primate visual cortex and its functional organization. He has championed many different techniques, such as imaging techniques (e.g. fMRI & PET studies have demonstrated the existence of “colour centres” on each hemisphere of the brain and recently the micro-areas of V4 and V4a and motion area V5/MT), visual psychophysics (e.g. study of visual illusions), and single cell physiology (e.g. examining the contributions of individual neurons in vision) in the scientific exploration of the “functions and functioning” of the visual system.

Lately however, he extended his empirical studies to the flexy-fields of aesthetics, love, creativity and consciousness. I find Prof . Zeki’s attempt to relate art and neuroscience -naturalizing aesthetics- interesting, I take his thesis on love and creativity with a pinch of salt, and I disagree with his theory of consciousness. I admire his interpretations of philosophers such as Plato and Kant forward-looking than many philosophical works. I submit my ego before his mastery of the subject.

History -philosophy

Historically, one may place his theoretical position in the line of rationalists and Kant, and against associationists. His theory of vision is much more close to the theoretical underpinnings of Chomskyan view of language than meets the eye (see Donald Hoffman’s Visual Intelligence for an extreme defence this view). The philosophical base of his theory, I would say, has a long successful history and is consistent with the basic assumption of contemporary cognitive science that human behaviour is determined by internal mental mechanisms and not by the objective properties of the world.

The above point is striking as Zeki in his talk praised Kant for his valuable insights contained in the Critique. He said, it is remarkable that even”without conducting a single experiment” Kant got the ideas right. As you might have guessed it, Zeki’s quotes were about Kant’s idea of sensation, perception, understanding, binding, presentation. It seems to me that his interpretation of Kant was not Kantian, but Strawsonian!

History -neurology

The neurological side of history, however, shows a different picture. In the early nineteenth century neurologists held that vision involves exact representation of the objective properties of the world such as colour, motion etc. In the late nineteenth century experimentally it was shown that seeing (passive)and understanding (active) are different faculties of the brain. There were cases of people with a particular brain damage who were blind, whereas people with brain damage in a different area couldn’t understand the what they saw, they were mind-blind/agnosic.

Research works on human and animal visual systems of the last three or four decades and the computer simulations of the models proposed have shown a more complex picture of the brain. The brain has specialised areas for processing different elements of the visual environment, such as colour, shape, depth, motion, and so on. And different properties of the visual stimuli are processed at different time scales, for instance, colour is processed faster than form and motion. From this empirical studies Zeki argues that visual system is not a passive recorder of the objective properties of the world but is an active construction site where different properties (visual sub tasks)are processed in different parts of it at different processing times. The functional specialisation of the visual brain make possible a coherent picture of the ever-changing world by selecting the relevant visual properties from the environment. He has shown that a particular neuron, say a red-sensitive neuron, would respond only when the corresponding colour and not the specific wavelength of that colour, was in the receptive field.

One Experiment

In one of the contemporary classic experiment on colour vision Edwin Land, famous for his invention of the Polaroid camera, has illustrated that the stimulus for perceiving colour of an image in natural settings is not the wavelength composition of the radiant energy reaching the eye from that stimulus. To put it in the words of Zeki (1973), “the brain is no mere passive chronicler of the external physical reality but an active participant in generating the visual image, according to its own rules and programmes“. (my emphasis). Zeki’s findings make explicit some of the rules and programmes of the brain

Land’s experiments used “mondrians”, after the Dutch artist Piet Mondrian (1872-1944), collage of paper rectangles with varied colours and sizes( see the example below)

“Mondrian-like collages were key elements of Edwin Land’s experiments on color vision. The Mondrians were illuminated by the three projectors at the bottom with light comprising various proportions of short, medium, and long wavelengths. A telephotometer, seen on a tripod to the right, measured the wavelength composition of the light reflected from a given colored patch to the eye. Land thereby showed that the perceived color of the patch is not determined by the wavelength composition of the light reflected from it. (Photograph by J. Scarpetti, courtesy of the Rowland Institute, Cambridge, Mass.)” —(Image and description courtesy Physics Today July 2002. http://www.aip.org/pt/vol-55/iss-7/p43.html)

By adjusting the illumination of the Mondrians, Land showed that the patches maintained a constant rank order of perceived lightness, even though a patch that appeared dark might be sending much more light to the eye than one that appeared light. Land’s “retinex” theory of color vision “recognizes lightness (that is, reflectance) as the fundamental stimulus of color, and it emphasizes the importance of boundaries, which allow the eye to estimate lightness by seeking out singularities in the ratio of energy flux from closely spaced points”.(ibid)

In the talk, Prof Zeki mentioned the experiment by Land which is explained in detail in his books, to demonstrate the law of constancy. Colour constancy is a function of the visual brain to acquire knowledge of the constant, enduring and essential properties of the world, when the information reaching it is never constant and enduring. We see the background of this page as white even when viewed from different angles and distance, under different lighting conditions. For instance in day light, in incandescent light, in candle light, in winter, in summer, when you shut the door, open the window, put the curtain and so on the wavelength composition of the light reflected from the screen/page changes, yet we see the screen/page having the same colour white. This is, to put it simply, because our brain “erase” the continuing changes and maintaining the colours and categorise the object accordingly. This is true even for the most elementary kind of vision, such as seeing this line _____________. Vision there fore is an active process of registering vast information from the constantly changing environment, of selecting the constant and essential properties of the object, of discounting the bombarding information that is not necessary for the acquisition of knowledge of the object, of comparing the selected information with the relevant database of the past visual information. It is through these processes that we see an object, categorise it as a particular kind.

The following example makes explicit a feature of the visual brain-colour constancy- that we employ regularly in perceiving objects and the world.

The experiment is this. A Mondrian ( a multicolour image) is illuminated by a red, green and a blue light emitting projectors. Experimental subjects with normal vision and cognitive ability are asked to view a particular square, a green one for instance, of the Mondrian which is made to reflect 30 units of red light, 60 units of green and 10 units of blue. When all the three projectors are switched on , which colour will the subjects see as the colour of the patch? All of subjects see the colour as green. This is obvious as the green square is illuminated by green colour 30 units more than red.

In the second experiment the subjects are asked to view the same square, but this time it is made to reflect 60 units of red, 30 units of blue and 10 units of blue. The amount of red and green light is reversed this time. Which colour this time the subjects will see? Even when the red light is double, this time also the subjects report the viewed colour as green. That is, a green sensitive neuron would continue to respond to a green stimulus even when it was illuminated mainly by red light. The brain discounts the changes in the lighting conditions to categorise the object, here the green square.

In certain pathological conditions such as carbon monoxide poisoning, however, patients didn’t see colour like other subjects. In one effectively similar case a patient has been examined in detail. He was severely blind but could see colours. His colour perception however, has a curious feature. He was able to see the correct colour only if the surface reflected the wavelength it is associated with than any other wavelength. To mention, he was able to see a green square as green only if it reflected more green light as in the experimental condition 1 mentioned earlier. If the green square is made to reflect more red light than green as mentioned in the experimental condition 2nd above, then he saw the square as red, unlike other subjects. The discounting mechanism was absent in his visual brain.

From many studies like this Zeki outlines his theory of visual brain thus: Vision is an active process involving many sub-processes occurring at different dedicated areas of the visual cortex. Vision is distributed in space as it is the result of processes occurring at geographically distinct visual areas. “The acquisition of knowledge by registering the constant and essential characteristics of objects is the primordial function of the visual brain.” (Science 6 July 2001: Vol. 293. no. 5527, pp. 51 – 52). Different attributes of the visual stimuli takes different processing times. so vision is distributed in time also. There fore perception of different attributes is not unified,it is distributed.

[Up-to this point, I’m in company with Prof Zeki, on his boat. Just to mention, my research is also tries to capture the distributed active internal processing at the representational level, which I call derivations. I cast the net a bit wider to include human cognition. And I find Prof. Zeki’s empirical findings very supportive to my own claims. ]

From these empirically supported claims he goes on to argue for in the following manner. Since perceiving an attribute of a visual stimuli is being conscious of it and perception is distributed, there is not a single unified visual consciousness., instead there are many visual micro-consciousness. And the binding of these distributed elements in different areas happening at different times occurs post-consciously.

But his theory of micro-consciousness, it seems, is in need of credible experimental support from the wider context of brain dynamics and Prof. Zeki knows this better than others. Comments like “our knowledge of how the brain works is still only very sketchy”, brain processes are “only now beginning to be physiologically charted” appear in his writings on a regular basis.

It seems, at first glance, that Zeki’s theory of visual micro-consciousness is an experimentally rich yet micro, version of Dennett’s Multiple Draft Model of consciousness. I do not yet know Zeki’s theory well, to make a sound comparison with the MDM. Prof. Zeki has not put his new paper on micro-consciousness on his site, so far. So a detailed comment about it after getting a copy of his theory. Meanwhile, I sketch below a general disagreement that comes to my mind after hearing his talk on this theory.

My criticism of Zeki’s theory of micro-consciousness stems from the following points. It does not follow from the existence of the distributed perceptual processings in the brain, that consciousness is also distributed. For it is possible to characterise the distributed processes, for instance registering an attribute and perceiving it, as not having independent consciousness. It is possible to characterise the perceptual consciousness is a result of something like feedback loop effect. The feed back is recorded in a different part of the brain other than the place of attribute processing. That is the feed back loop is connected to some other part of the brain. (There are in fact some candidate locations in the brain where the feed back loop is registered)

Further there are many reported experimental cases where patients retain consciousness even after lesions in major parts of the brain. To put it in the present context that would mean even after losing micro-consciousness patients retain consciousness. So a theory of micro-consciousness fails to be a general theory of consciousness.

In the experiment we mentioned above normal subjects do not have any consciousness of the registering colours/wavelength; they see colour only after brain makes the necessary comparison using its discounting mechanism. And normal subjects are not conscious of the registered attributes. The point is even if the wavelength is registered and processed in the normal brain, it is not conscious, and not even micro-conscious of it. But in the case of the patient mentioned above the comparison mechanism was absent and he saw only the colour corresponding to the registered wavelength. The patient’s brain, however, registered the wavelength and processed it, and he was conscious of the wavelength. So in this latter case an absence of a mechanism results in visual micro-consciousness, whereas in the normal subjects the presence of the mechanism results in the absence of that visual micro-consciousness. How do we explain this?

It seems, if the feed back of the perceptual processes is registered, more or less simultaneous with registering process, somewhere else in the brain other than the spot where it is processed, then we can offer a simple explanation of the problem. I don’t wish to speculate more about this here. Let’s see how he explains his theory in his forthcoming paper.

Courtesy (Both his books are very well written and contains detailed discussion of many fascinating experiments. )

Zeki, S. (1993) A vision of the brain. Oxford: Blackwell Scientific;

Zeki, S. (1999). Inner Visions: An exploration of art and the brain. Oxford: Oxford University Press

Zeki, S., and A. Bartels, (1999) “Towards a Theory of Visual Consciousness” in Consciousness and Cognition

http://www.sciencemag.org/cgi/content/full/293/5527/51 http://www.btgjapan.org/links/008.html

<a href=”http://technorati.com/tag/consciousness&#8221; rel=”tag”>consciousness</a>