Controversies In Cognition

1
Controversies In Cognition

• Neuropsychology and the evolutionary value of
  colour

2
Aims

• To identify whether colour vision fits easily
  into Milner and Goodales Two Visual Brains
  Theory .
• To look at research that suggests that processing
  of colour information provides us with more than
  just the conscious experience of colour.
• To look at the possible evolutionary benefits of
  colour vision.
• To present Akins (2000) position on the role of
  colour vision.

3
Objectives

• To be able to identify whether colour vision fits
  easily into Milner and Goodales Two Visual
  Brains Theory .
• To be able to describe and discuss research that
  suggests that processing of colour information
  provides us with more than just the conscious
  experience of colour.
• To be able to describe and discuss the possible
  evolutionary benefits of colour vision.
• To be able to discuss Akins (2000) position on
  the role of colour vision.

4
Last week

• We looked at Milner and Goodales Two Visual
  Brains Theory.
• Neuropsychological evidence provides strong
  support for their theory of visual perception.
• Using this evidence it should be possible to see
  the way in which the two streams differ in how
  they deal with incoming visual information.
• However, despite strong support for this theory
  there are still gaps that need clarification.
• One such gap relates to the role that colour
  processing plays in the ventral and dorsal
  framework.

5
Where does colour fit in?

• Colour is clearly a key physical property of the
  visual world so it must have a place in Milner
  and Goodales framework.
• There is some good evidence that colour
  information enters the ventral stream via input
  from the parvocellular (P) channel, which is,
  unlike the magnocellular (M) channel, mainly
  responsible for colour (Livingstone and Hubel,
  1987).
• Further evidence for the link between colour and
  the ventral stream is that cells in this stream
  are often sensitive to colour (e.g. Zeki, 1980).
• In addition, visual area V4 has often been
  regarded as the primate brains colour centre
  (e.g. Zeki, 1993) and, as this area is part of
  the ventral stream, this would be good grounds
  for assuming that the ventral stream provides the
  sole basis for the visual experience of colour.

6
Is V4 the primate brains colour centre?

• However, much evidence casts doubt on the idea of
  V4 being the primate brains colour centre.
• Although Zeki (1973) reported that every one of
  the 77 cells he sampled in region V4 showed
  selectivity to wavelength, more recent evidence
  suggests that many cells in the region are
  strongly receptive to orientation and are even
  indifferent to wavelength (Schein, Marrocco and
  De Monasterio, 1982 Desimone and Schein, 1987).
• It was also believed that monkeys with surgically
  induced lesions to V4 show major impairments to
  colour discrimination.
• However, when monkeys with V4 lesions were
  tested, the results suggested only a mild
  impairment (Heywood and Cowey, 1987) or none at
  all (Dean, 1979).

7
Is V4 the primate brains colour centre?

• Furthermore, when luminance and chromaticity were
  carefully controlled on a display monitor for
  monkeys with V4 completely removed, there was no
  evidence of any impairment in colour oddity tasks
  (Heywood, Gadotti and Cowey, 1992).
• Wild, Butler, Carden and Kulikowski (1985)
  monkeys with damage to V4 retain their colour
  vision (e.g. they can learn to pick up a yellow
  object) but lose colour constancy (the
  phenomenon whereby an objects colour is
  perceived as unvarying or fairly stable despite
  considerable variation in the spectral
  composition of the illuminant (Heywood and
  Cowey, 1999, p.29)) (e.g. they cannot find the
  same yellow object under blue overhead lighting).
  
• For humans with damage to an area that straddles
  the temporal and parietal cortices, which may
  correspond to the monkey area V4, there is a
  similar loss of colour constancy though colour
  recognition remains intact (Ruttiger, Braun,
  Gegenfurtner, Petersen, Schonle and Sharpe,
  1999).

8
Is V4 the primate brains colour centre?

• It now seems reasonable to say that V4 is either
  responsible for, or has a crucial role in, colour
  constancy. Its also reasonable to suggest that
  another region (for example, V8) or, more likely,
  several regions in combination, beyond V4 are
  responsible for constructing colour.
• It still seems likely, though, that the colour
  centre is based within the ventral and not the
  dorsal stream.
• The link between colour and the ventral stream
  may be an intuitive one because, in the
  terminology of Milner and Goodale, we commonly
  use colour as part of our internal representation
  of the world, plus we are consciously aware of
  it, i.e. we can report on our experience of any
  given colour we see, we can name it and compare
  it with, or distinguish it from, other colours.

9
Is there any neuropsychological evidence for this
assumption?

• If colour perception could fit easily into
  Goodales and Milners framework then one would
  expect there to be evidence of neuropsychological
  dissociations like those mentioned last week.
• There are many cases of individuals with impaired
  colour perception resulting from brain damage.
• Example 1 cerebral achromatopsia.

10
What achromatopsics can do with colour

• Although they are without an awareness of colour,
  these patients can still somehow use differences
  in wavelength to make certain distinctions of
  form and movement.
• E.g. patient MS is densely achromatopsic and
  yet can do some tasks involving colours that abut
  each other (Heywood, Cowey, and Newcombe, 1991,
  1994). His verbal replies suggest that he is
  successful in these tasks because he can detect
  edges between stimuli, even though the stimuli
  appear to him to be identical.

11
What achromatopsics can do with colour

• This ability of MS, to detect chromatically
  generated contours (Heywood, Kentridge, and
  Cowey, 1998a), has also been shown in other
  achromatopsic patients (Barbur, Harlow and Plant,
  1994).
• MS, and other achromatopsics, can somehow detect
  chromatic information without experiencing the
  colours themselves. These findings suggest that
  at some level, their colour processing remains
  intact. They have no phenomenal experience of
  colour, and yet retain some colour abilities but
  not others.
• However, cerebral achromatopsia is not
  blindsight for colour (Heywood, Kentridge, and
  Cowey, 1998b).

12
Example 2 Colour Blindsight

• The blindsight phenomenon suggests that the
  subcortical pathways can carry considerable
  amounts of visual information. Some evidence
  suggests that this may include information about
  colour.
• Bender and Krieger (1951) Perenin, Ruel, and
  Hecaen (1980) Blythe, Kennard, and Ruddock,
  (1987) Stoerig, (1987) Stoerig and Cowey (1989,
  1992) all reported that wavelength discrimination
  was possible in the blind fields of patients.
• However, the evidence for colour blindsight is
  still not clear-cut.

13
Colour blindsight (2)

• E.g. two out of the 10 patients who Stoerig
  (1987) tested failed to reach the .05 level of
  statistical significance, suggesting that this
  residual chromatic processing is not always
  demonstrable and that even when it is it is
  seriously impoverished.
• Until there is stronger evidence for the reality
  of colour blindsight it may be that it cannot
  be reliably distinguished from standard cases
  of achromatic blindsight.

14
Unconscious Processing of Colour in Normal
Observers

• There is other evidence for the unconscious
  processing of colour (colour processing without
  actual experience of colour). Some of this work
  has moved away from focusing only on patients
  with neuropsychological deficits and instead has
  concentrated on normal observers however this
  work has been slow in coming and is somewhat
  controversial.
• Vision science has long assumed that colour is
  a term that refers only to something that can be
  consciously experienced, perhaps because the idea
  of non-conscious perception of colour might
  seem like a contradiction in terms (Schmidt,
  2000).

15
Unconscious Processing of Colour in Normal
Observers (2)

• Its important to be cautious, as some of this
  evidence comes from studies of subliminal
  perception which have often been shown to be
  unreliable and susceptible to response bias
  (Holender, 1986). Many experiments have been
  criticised for their failure to establish that
  the stimulus was not consciously perceived during
  presentation (Schmidt, 2000 Eriksen, 1960
  Holender, 1986).
• Nevertheless, during a stimulus-detection study,
  Rollman and Nachmias (1972) presented their
  participants with chromatic discs. Whenever the
  participants incorrectly reported the stimulus as
  absent they were asked to guess what colour the
  disc would have been if it had actually been
  presented. These guesses proved correct
  significantly more often than would have been
  expected by chance.

16
Colour Processing is not just about seeing colour

• Taken as a whole these findings suggest that
  colour processing need not actually lead to the
  experience of seeing colour (Schmidt, 2000).
• Colour vision is assumed to be a single entity,
  but it may in fact be made up of (at least) two
  separate components.
• Colour in fact seems to be used in several, quite
  distinct ways, to extract information from the
  visual scene.
• What we typically call colour is actually
  surface colour, which gives us information about
  the surface properties of an object (and of
  gases, liquids, etc.).
• This is colour as we normally come to think of
  it, what we would call the qualitative experience
  associated with colour perception, or colour
  qualia.

17
Colour Boundaries

• The other type of information is information
  about form and this is obtained via colour
  boundaries, i.e. chromatic signals create spatial
  structures.
• It seems likely that it is this detection of
  coloured boundaries that is intact in
  achromatopsics and it is this that enables them
  to make certain colour distinctions without being
  able to see the colours themselves.
• Heywood, Cowey, and Newcombe (1994) describing
  the pattern of abilities of patient MS He
  canuse wavelength to extract form but his
  performance on tasks of colour discrimination
  suggest he has no phenomenal experience of colour
  itself. (p.253)
• Presumably achromatopsics have damage to brain
  regions that are indispensable for the production
  of object colour. The regions that enable
  chromatic information to signal form, however,
  are still intact.

18
Research on infants

• It has long been known that infants as young as
  two months can make chromatic discriminations.
• But what mediates these discriminations? Either
  infants have normal colour vision in the ordinary
  sense, i.e. preserving information about
  wavelength, or they use edge cues as the basis
  for their response discriminations.
• Thomasson and Teller (2000) manipulated the
  juxtaposed stimulus fields of earlier
  experiments, eliminating the sharp chromatic
  edges to see whether infants (16-week-olds) could
  still make the chromatic discriminations without
  them.
• In all cases, manipulating the edge led to poorer
  performances, but performance still remained
  reliably above chance for most cases of chromatic
  stimuli. This suggests that infants as young
  16-weeks can make chromatic discriminations in
  the absence of sharp chromatically defined edges.
  However, as performance was affected by the lack
  of edges this suggests that they still play an
  important part in human colour discrimination.

19
Does this make sense?

• If colour is defined as two distinct things then
  this conflicts with our phenomenological
  experience of the world, that the only role of
  colour vision is to enable us to see various
  media as coloured.
• But as Akins (2000) points out, we should not
  necessarily be asking questions about colour only
  as it exists in our standard phenomenology, where
  the emphasis is on seeing coloured surfaces, we
  may need to ask broadly what value spectral
  information has for the visual system.
• This may involve considering the original
  evolutionary benefits of colour vision (akin to
  what Pinker, 1997, calls the logic of reverse
  engineering using evolutionary theory as a
  guide to gain insight into how the visual system
  and the mind works), something well discuss next.

20
Evolutionary benefits of colour vision

• We need to question common sense notions about
  the use of colour vision and instead look at its
  original evolutionary benefits. Can we account
  for both the experience of surface colour and the
  detection of boundaries through colour?
• Given that there not only appear to be
  specialised mechanisms that process
  colour/wavelength information but that there may
  also be separate mechanisms for surface and edge
  colour it is not unreasonable to reconsider the
  functional role of colour in everyday life.
• Theories of colour vision need to have some sense
  of biological utility they should focus on
  what the organism in its visual environment may
  actually need the components of colour vision for.

21
Evolutionary benefits of colour vision (2)

• Other mammals have dichromatic or monochromatic
  colour vision but only primates have the three
  types of cone receptors that provide trichomatic
  colour vision. What advantage do primates have
  over these animals?
• The detection of form through colour might offer
  considerable benefits for survival. As much of
  human vision involves spotting stimuli quickly
  and accurately in cluttered environments this
  additional method of detecting form would be
  advantageous in a number of ways for example, it
  would be useful for the avoidance of predators,
  for distinguishing ripe fruit amongst the leaves
  of trees, etc.

22
Evolutionary benefits of colour vision (3)

• Regan, et al. (1998) and Sumner and Mollon (2000)
  found that primate photopigments are ideal for
  differentiating edible fruits and young leaves
  against a background of mature leaves.
• This sort of research has mainly looked at the
  optimality of red and green cone responses to the
  spectral properties of red objects (ripe fruit,
  flowers, young leaves) amongst green objects and
  backgrounds (unripe fruit and mature leaves), the
  assumption being that the red/green system helps
  to distinguish the edible from the inedible
  (Dominy and Lucas, 2001).

23
Evolutionary benefits of colour vision (4)

• Questions remain about the role of the
  yellow/blue system and why it developed.
• Troscianko, Parraga and Tolhurst (2002) suggest
  that not all monkeys eat a diet that consisting
  only of red fruit and leaves, they also eat green
  leaves from certain trees but not others.
• Perhaps the red/green and yellow/blue systems
  work together to help distinguish the
  different types of leaves, particularly under
  different levels of illumination.
• Troscianko et al. found that, as well as being
  able to pick out red fruit and leaves from green
  foliage, the red/green system is indeed also good
  at discriminating between different tree types,
  though found no compelling evidence that the
  yellow/blue system is particularly suited to make
  these discriminations.
• This does not rule out the possibility that both
  the red/green and yellow/blue system provided an
  evolutionary advantage by helping to detect
  colour edges, just that it is at present unclear
  what the yellow/blue system's contribution is.

24
Evolutionary benefits of colour vision (5)

• The experience of surface colour, or colour
  qualia, on the other hand, has very different
  advantages, including being an essential feature
  for long-range planning, communication and other
  cognitive activities (Goodale and Humphrey,
  1998, p.193).
• This aspect of colour is more clearly linked to
  our internal representation of the world and to
  consciousness, which might suggest that it is, in
  evolutionary terms, the more recent development
  of the two aspects of colour perception. We take
  for granted that the primary or even the sole
  function of colour vision is to enable us to see
  the world in colour, but in fact it may be a
  secondary function or even a bi-product (Akins,
  2000). This might also explain why there are
  apparent dissociations between conscious and
  non-conscious aspects of colour processing that
  relate to surface colour and boundary colour
  respectively.

25
Recommended Reading

• Akins, K. A. (2000). More Than Mere Colouring A
  Dialogue Between Philosophy and Neuroscience on
  the Nature of Spectral Vision. Essay submitted to
  the Mcdonnell Foundation in application for the
  Mcdonnell Centennial Fellowship. Available at
• http//www.stu.ca./neurophilosophy/members/akins/a
  ppessay.htm
• Heywood, C. A., and Cowey, A. (1999). Cerebral
  achromatopsia. In G. W. Humphreys (Ed.)
  Neuropsychology of Vision. Hove Psychology Press
  Ltd.
• Schmidt, T. (2000). Visual Perception Without
  Awareness Priming Responses by Color. In T.
  Metzinger (ed.) Neural Correlates of
  Consciousness empirical and conceptual
  questions. Cambridge, Mass. London MIT Press.