Artificial Minds Ian Fairholm

1
Artificial MindsIan Fairholm

• Neural Correlates of Consciousness

2
Aims

• To define what neural correlates of consciousness
  (NCCs) are and explain why we might expect to
  find them.
• To identify the methods used in looking for NCCs.
• To present Milner and Goodales Two Visual
  Brains Theory and the evidence for it.
• To identify how close we are to finding the NCC
  for visual consciousness.

3
Objectives

• To be able to define NCCS and explain why we
  might expect to find them.
• To be able to describe the methods used in
  looking for NCCs.
• To be able to describe Milner and Goodales Two
  Visual Brains Theory and the evidence for it.
• To be able to come to a conclusion as to how
  close we are to finding the NCC for visual
  consciousness.

4
Why do we expect to find NCCs?

• If consciousness is a product of the brain then
  we should expect to find that conscious
  experience, or particular elements of it,
  correlates with activity in the brain.
• Hence, the definition a specific pattern of
  brain activity that correlates with particular
  conscious experiences (Rees et al., 2002)
• Important to remember that a correlation isnt
  the same as a cause.

5

• In attention research, the focus is on visual
  attention.
• Perhaps we need top focus on visual consciousness.

6

• Prosopagnosia
• Implicit recognition

7
Conscious and Unconscious Vision

• Because much of our conscious experience is of a
  visual nature, most of the work on NCCs has
  focused on visual consciousness.
• Cases of blindsight suggest a kind of vision
  exists that is different from conscious vision.
• Individuals like DB, with damage to the primary
  visual cortex (area V1) in one hemisphere, lack
  visual awareness in the contralateral part of the
  visual field (their blindfield).

8
Blindsight

• But, these individuals have been shown to have
  residual, unconscious, visual function in their
  blindfield - blindsight.
• When given certain visual forced choice tasks,
  involving a stimulus presented in their
  blindfield, these patients are unable to
  consciously see the stimulus at all.
• But when forced to guess the right answer, they
  are correct far more often than would be expected
  by chance, sometimes as much as 90-95 of the
  time.

9
The Basis Of Blindsight

• How is blindsight possible? In normal primate
  vision, signals travel from the retina to the
  thalamus and then to V1.
• This is the main pathway from the eyes to the
  brain and provides the fine-grained information
  needed for the identification and analysis of
  visual images.
• However, there are several other pathways.

10
The Basis Of Blindsight (2)

• A particularly important one is the
  evolutionarily older subcortical pathway that
  carries information from the retina to the
  superior colliculus.
• This visual pathway is fast and unconscious. It
  provides only rough information about the
  location identity of stimuli it cannot resolve
  fine detail.
• Its also particularly sensitive to stimuli
  falling on the periphery of the retina (i.e. the
  corner of the eye).

11
The Basis Of Blindsight (3)

• Movement or stimuli in the periphery can
  automatically capture attention, and the
  colliculus directs a quick reflex eye movement to
  the stimulus.
• This rapid response is thought to have evolved to
  support predator avoidance or prey capture.
• This collicular pathway may underlie blindsight.
• Blindsight is paradoxical only if one regards
  vision as a unitary process (Milner and Goodale,
  1995, p. 86).

12
Milner and Goodales Two Visual Brains

• Blindsight suggests a dissociation between fast
  motor reactions and conscious perception.
• Milner and Goodale (1995) offer their explanation
  for this, that there are two vision systems
  visual perception and visuomotor control. Only
  the former is involved in conscious awareness.
• They map onto the two neural streams in the
  primate visual system the ventral and dorsal
  streams.

13
The Dorsal Stream

• Provides automatic, unconscious vision for
  action, e.g. returning a tennis ball
  instinctively that has been served to us.
• Responsible for visually guiding online
  movements that are made in the here and now.
• Ends in the posterior parietal cortex, and is the
  evolutionarily older of the two cortical
  pathways.

14
The Ventral Stream

• Supports conscious visual experience, enabling
  recognition of visual objects and selection of
  appropriate behaviour towards them, e.g. knowing
  what a tennis ball is and what to do with it.
• Linked with learning and memory, it underlies
  visual planning and decision making.
• Ends in the inferotemporal cortex, and is the
  evolutionarily newer of the two cortical
  pathways.

15
Dorsal and Ventral Streams

• The two visual brains arent totally independent
  but act together in everyday life.
• Evidence for the two streams comes from patients
  with visual form agnosia (e.g. DF) and optic
  ataxia (Balints patient).
• Milner and Goodale hypothesize that the NCC in
  primates is in the ventral stream.

16
Testing the Milner and Goodale hypothesis

• To test their hypothesis more directly we need a
  method that allows us to differentiate conscious
  and unconscious brain activity for the same
  stimuli.
• We need a visual stimulus that doesnt change
  physically but appears to change while we
  consciously perceive it.
• Examples Ambiguous Figures and Binocular Rivalry.

17
Ambiguous Figures

• There are many ambiguous figures. They all give
  rise to bistable percepts
• A single pattern of retinal input gives rise to
  two alternating perceptual interpretations.
• Because these changes occur in the absence of any
  change in the stimulus itself, they are related
  directly to subjective conscious awareness.

18
Binocular Rivalry

• Found when different images are presented to the
  two eyes.
• First reported by the inventor of the
  stereoscope, Sir Charles Wheatstone, in 1838.
• He found that when two letters or patterns are
  presented in a stereoscope, so that only one can
  be seen by each eye, they do not fuse into one
  image but instead compete for visual awareness.
• Sometimes the image from one eye is consciously
  seen, sometimes its the other image, and
  sometimes the two images overlap.

19
Binocular Rivalry (2)

• Logothetis et al. have used binocular rivalry to
  look for the neural correlate of visual
  consciousness in monkeys.
• Monkeys trained to squeeze different levers when
  confronted with two different images at once.
• Electrodes inserted in various areas of their
  brains to record activity of single neurons.
• The goal to find the areas where activity
  corresponded to the changing perceptions reported
  by the monkeys behaviour.

20
Findings of Logothetis et al.

• Cells in early visual cortex, e.g. V1, responded
  to unchanging input, but their behaviour didnt
  change when the monkeys perception changed.
• Further along the visual pathway, e.g. in MT and
  V4, some cells responded to what the monkey
  reported seeing.
• In the inferotemporal cortex (IT), almost all of
  the cells responded to what the monkey reported
  seeing (percept-related neurons).

21
So does the NCC of visual consciousness lie in IT?

• Activity in this area does appear to correspond
  to what the monkey was consciously seeing, rather
  than what was in front of its eyes.
• IT is part of the ventral stream, providing
  support for Milner and Goodales hypothesis.
• However, this assumes that monkey perceptions are
  conscious in the same way that human perceptions
  are.

22
Further Research

• Lumer, Friston and Rees (1998) used fMRI to
  detect changes during binocular rivalry in human
  brains.
• Findings indicated that activity in the parietal
  and prefrontal cortices might be associated with
  visual awareness, areas already implicated in
  selective attention.
• Again, suggests that conscious visual experiences
  are correlated not with activity in V1 and other
  early parts of the sensory pathways but with more
  central areas.

23
However

• It seems that several areas are necessary for
  normal visual consciousness (you cant have
  consciousness without them when you are
  conscious this area is activated)
• But not sufficient for normal visual
  consciousness (in isolation it doesnt produce
  consciousness activation of this area isnt
  always accompanied by consciousness).
• For example, we know that V1 is necessary for
  conscious vision (e.g. cases of blindsight), but
  other work suggests that it isnt sufficient.

24
Conclusions?

• No single area, pathway or locality appears to be
  the NCC for visual consciousness.
• However, its widely agreed that the ventral
  pathway, interacting with areas of prefrontal and
  parietal cortex, is an important part of it.
• Unfortunately this fails to touch on the central
  mystery how is consciousness generated in the
  first place, whether by one brain area or many?
• Maybe the whole enterprise is misconceived
  (Blackmore, 2002 2003).

25
Recommended Reading

• Blackmore, S. (2003). Consciousness An
  Introduction. London Hodder and Stoughton.
  (Chapters 3 and 16)
• Metzinger, T. (ed.) (2000). Neural Correlates of
  Consciousness empirical and conceptual
  questions. Cambridge, Mass. London MIT Press.
• A précis of Milner, D. A. and Goodale, M. A.
  (1995). The Visual Brain In Action. Oxford
  Oxford University Press is available at
  http//psyche.cs.monash.edu.au/v4/psyche-4-12-miln
  er.html