Chapter 10 Visual Awareness

1
Chapter 10Visual Awareness

• The neural bases of visual awareness as revealed
  by dissociations between vision and awareness of
  vision

2
Visual Awareness Conscious of What?

• Conscious
• Product (content) of seeing
• Text
• Faces
• Attended input
• Activated memory

• Not Conscious
• Process of seeing
• How we read
• How we see faces
• Unattended input
• Non-activated memory

3
Conscious Visual Awareness

• Subjective Experience
• Qualia (redness, roundness)
• P(henomenal)-consciousness (experiential state)
• What it feels like to be a bat or a person
• Behaviour
• We can act on available visual information
  intentionally.
• We can communicate about available visual
  information.

4
Why are we conscious?

• Consciousness may provide some adaptive
  advantage.
• Frogs. Reflexive responses (Zoombie mode) to
• -- small round moving objects
• -- looming objects
• - Species with many systems (reflexive and
  other) that could be in conflict. Combine
  outputs of systems into a higher order complex
  representation and make it available to the
  relevant portions of the brain for a sufficient
  amount of time to support decision making and
  action.

5
Some Potential Functions of Consciousness

• Coping with novelty and unpredictability through
  learning, adapting, and problem solving
• Prioritizing information
• Controlling mental and physical actions
• Making decisions

6
Neural Correlates of Visual Awareness

• Six (five) syndromes with evidence of preserved
  visual function and impaired visual awareness
• Blind sight
• Agnosia Apperceptive and Associative
• Prosopagnosia
• Neglect Extinction
• Reading in Pure Alexia (awareness not impaired)
• Identifying the mechanism of awareness
• Brain function of people with normal
  awareness
• Brain function of people with impaired
  awareness
• Brain mechanism of awareness

7
Neural Correlates Three Different Accounts

• Privileged Role
• Schacter Dissociable Interactions and Conscious
  Experience Model (DICE)
• Gazzaniga Left-Hemisphere Interpretive
  Mechanisms
• Integration
• Kinsbourne Integrated Field Theory
• Crick Koch Binding Model
• Damasio Convergence Zones
• Quality of Representation
• Farah et al Degree of Impairment in Visual
  Perception

8
Consciousness The Privileged Role of Particular
Brain Systems

• Descartes (1600s) Patterns of brain activity
  impinging upon pineal gland were consciously
  experienced.
• Schacter (late 1980s) Dissociable Interactions
  and Conscious Experience (DICE) model
• Conscious Awareness System (CAS) A brain system
  separate from perception and cognition that gives
  rise to conscious awareness.
• Loss of awareness reflects a disconnection
  between the visual system and CAS.
• Gazzaniga (1980s) Left-Hemisphere
  Interpretation Mechanism.

9
Schacter, 1989
10
Roser Gazzaniga (2005, p. 58)

• The left-hemisphere interpreter may be
  responsible for our feeling that our conscious
  experience is unified. Generation of
  explanations about our perceptions, memories, and
  actions, and the relationships among them, leads
  to the construction of a personal narrative that
  ties together elements of our conscious
  experience into a coherent whole. The
  constructive nature of our consciousness is not
  apparent to us. The action of an interpretive
  system becomes observable only when the system
  can be tricked into making obvious errors by
  forcing it to work with an impoverished set of
  inputs, such as in the split brain or in lesion
  patients. But even in the damaged brain, this
  system lets us feel like us.
• It is becoming increasingly clear that
  consciousness involves disunited processes that
  are integrated in a dynamic context. It is
  assembled on the fly, as our brains respond to
  constantly changing inputs, calculate potential
  courses of action, and execute responses.

11
Consciousness A State of Integration among
Distinct Brain Systems

• Dynamic states of brain systems
• Kinsbourne (1988) Integrated Field Theory.
• Awareness as a state of the brain wherein all
  modality-specific perceptions, memories, current
  actions and action plans are mutually consistent.
• Vision without awareness disconnection of or
  damage to the visual system prevents vision from
  participating in integrated patterns of activity
  across the brain.
• Crick Koch (1990s) Binding Model.
• Visual awareness arises when visual properties
  (e.g., color) of stimulus are bound together via
  synchronized oscillations
• Damasio (1990s) Convergence Zones.
• Binding operates across, as well as within,
  modality-specific representations of an object.

12
A Neural Correlate of Consciousness

• Crick Koch (1990)
• Our basic idea is that consciousness depends
  crucially on some form of rather short-term
  memory and also on some form of serial
  attentional mechanism. This attentional mechanism
  helps sets of relevant neurons to fire in a
  coherent semi-oscillatory way, probably at a
  frequency in the 40-70 Hz range, so that a
  temporary global unity is imposed on neurons in
  many different parts of the brain. These
  oscillations then activate short-term (working)
  memory. p. 277 of Block et al (Eds., 1997

13
Consciousness A Graded Property (Quality) of
Neural Information Processing

• Representation of information in neural systems
  is graded, not all-or-none.
• Information may be partially represented
• Normal subjects Using impoverished input (e.g.,
  Marcels studies of subliminal perception and
  masking studies)
• Patients Damage to the visual network.
• Farah, OReilly Vecera (1993). Impairment in
  visual perception model.
• Quality of perceptual representation correlated
  with probability of conscious awareness
• Consciousness may be associated only with high
  quality representations.

14
Conscious and unconscious perception (Marcel,
1983a, b)

• Accuracy in Experiment 1 (1983a).
• Semantic similarity gt Graphic Similarity gt
  Detection.
• (1983a). Five experiments challenge the view
  that representations yielded by perceptual
  analysis are identical to and directly reflected
  by phenomenal percepts. (p. 197)
• (1983b). In other words, conscious perception is
  constructed from perceptual representations it
  is not identical to or a direct reflection of
  representations yielded by perceptual processes.
  (p. 238)

• A typical trial (1983a)
• Present fixation point
• Present word briefly (too short to be aware of
  seeing it)
• Present pattern mask
• Participant makes decision about
• (a) presence/absence of word
• (b) graphic shape or meaning of the presented
  word. For these decisions, the subject selected
  the item that was most similar to the target word
  from a pair of test words.

15
The Six Syndromes

• Covered by Rita Anderson
• Prosopagnosia
• Pure Alexia (in passing not relevant)
• Covered by Carolyn Harley
• Blindsight
• Agnosias
• Neglect Extinction

16
Prosopagnosia Evidence for implicit (covert)
face recognition

• Tranel Damasio (1985).
• Task Rate familiarity of faces of family
  (herself, family, friends), famous people
  (actors, politicians) interspersed with
  unfamiliar faces.
• Despite inability to recognize familiar faces,
  two females with prosopagnoia generated larger
  and more frequent skin conductance responses
  (SCR) to familiar than to unfamiliar faces.
• Demonstrates total dissociation between overt and
  covert face recognition, despite normal visual
  perception and associated memories.

• S1
• Mean Rated familiarity
• (1very familiar, 6 very unfamiliar)
• Family 6.0
• Famous 6.0
• Unfamiliar 6.0
• SCR (mean amplitude)
• Family .934
• Nonfamily .048
• Famous .731
• Nonfamous .012

17
Prosopagnosia A case study of face recognition
without awareness

• de Haan, Young, Newcombe (1987)
• PH Unable to recognize familiar faces visually,
  but can recognize them from their names. Knows
  that he is looking at a face, can comment on age,
  sex, hairstyle. Performance on the following
  three tasks similar to normal people, but he
  feels like he is guessing.
• Matching Task (Identity) See two photos,
  Same/Different person?). PH is slower than
  normal and makes more errors, but he matched
  familiar faces faster than unfamiliar faces (same
  pattern for controls)
• Error Rate Mean
  RT (ms)
• Familiar Unfamiliar
  Familiar Unfamiliar
• PH 18.7 16.4 2550 2762
• Controls 1.6 3.9 977
  1045

18
PH case study cont.

• Interference Task. Classify name as pop star or
  politician when name presented alone or
  accompanied by a photo.
• RT slowest when name and photo are unrelated with
  respect to occupation than in other conditions.
  PH must be processing facial identity to show
  this occupational interference effect.
• Learning task. Learn correct/incorrect pairings
  of names/faces from before and after the
  accident.
• PH learned correct name/face pairings faster than
  incorrect pairings, suggesting that he can make
  relatively fine perceptual face discriminations.
  

Same person 1502 ms
Unrelated 1714 ms
Related 1560 ms
From deHaan et al (1987)
19
Semantic priming PH and Controls Young,
Hellawell de Haan (1988)

• Three types of priming conditions (using faces or
  names) for target names
• Related John Lennon (prime) for Paul McCartney
  (target name)
• Neutral Unfamiliar person
• Unrelated Two familiar people not related to one
  another.
• Task Classify name as familiar or not measure
  RT
• Typical trial
• Fixation
• See Prime (450 ms for PH, 250ms for controls)
• See Printed Name

• Mean RT (ms) to familiar and unfamiliar target
  names preceded by face or name primes
• E4 E1
• Familiar PH Control
• Related 981 695
• Neutral 1056 776
• Unrelated 1083 815
• Unfamiliar 1373 883

PH slower, but shows facilitation from related
face or word primes
20
What type of model can best account for the basic
results?

• Basic Result
• Dissociation between overt and covert recognition
  of faces in some (but not all) prosopagnosics
• Privileged role and integration models assume
  that face processing is normal and that conscious
  access is post-perceptual.
• Privileged role
• de Haan, Bauer Greve (1992) Face recognition
  units (FRUs) disconnected from Conscious
  Awareness System (CAS) (p. 308).
• Bauer (1983). Ventral system (mediates identity
  and conscious awareness) damaged dorsal system
  (mediates affective response to faces) supports
  covert recognition (pp. 307-308).
• Integration
• Burton et al (1991) Partial disconnection of
  face recognition units (FRUs) from personal
  identity units, the major pathway to the rest of
  the system (p. 309).

21
Privileged Role account de Haan, Bauer Greve
(1992)

• Face processing module operates normally and
  hence, supports covert recognition of faces.
• Overt recognition of faces fails because the
  output of the face processing module cannot
  access awareness due to a lesion at Location 1.

From Farah (2000)
22
Integration Account. Burton et al (1991)
Interactive Activation and Competition (IAC)
model of face recognition

• Pools of units correspond to
• FRU Face recognition units
• NRU Name recognition units
• SIU Semantic information (about the person)
  units
• PIN Person identity node (amodal interface
  between input and semantic information)
• If reduce connection strength between FRUs and
  PINs sufficiently, activating a FRU will not
  activate the PIN beyond threshold and overt face
  recognition will not occur. But, the PIN is
  activated above resting level and can hence,
  support implicit effects (relearning, priming,
  etc).

From Farah (2000)
23
Quality of Representation Farah, OReilly
Vecera (1993). A simulation to test whether
dissociation effects in prosopagnosia could be
due to an impairment in visual recognition

• Both Privileged Role and Integration accounts
  assume that face recognition proceeds normally.
  Does it?
• PH makes more errors and is much slower than
  control subjects. No empirical data exist to
  determine whether the face module in
  prosopagnosia patients is normal.
• Is it necessary to postulate a separate mechanism
  for conscious awareness?
• If covert recognition tests are more sensitive
  than overt recognition tests, a person with a
  degraded representation that is below the
  threshold of awareness might show some residual
  recognition when tested implicitly.
• No empirical data are available. Can test the
  viability of the Quality of Representation
  hypothesis by running a simulation.

24
Farah et al (1993) The Neural Network Model

• Input Unit Pools Hidden Unit Pools
• 16 Face units 16 Name units
• Semantic Pool
• 18 units, one either politician or artist
• Distributed representation
• Faces (Names) random pattern of 5 units
  activated
• Semantic random activation of 6 units (one
  occupational)
• Network trained to associate an individuals face
  and name (via semantic units) using a Hebbian
  rule (i.e., units that fire together wire
  together).
• Learned 40 distinct individuals of which 10 were
  actors and 10 politicians

25
Farah et al (1993) Simulation 1a Overt
Recognition

• Lesion network by eliminating randomly chosen
  units from face input (or hidden) unit pools.
  Seven levels of lesion created by removing 2, 4,
  8, 10, 12, or 14 units from the 16 unit pools.
• Overt Recognition Percent correct name
  identification of faces in a 10-afc test (Chance
  10)
• Overt recognition significantly above chance
  until more than 50 of units removed.

26
Farah et al (1993) Simulation 1b Savings in
Relearning

• Savings in Relearning Face-Name Associations
• The lesioned network was retrained on the names
  and faces of 10 familiar actors and politicians
  paired correctly or incorrectly.
• Despite chance performance on overt test (at 75
  lesion), the system relearned correct pairings
  faster than incorrect pairings. This is similar
  to PH results.
• Why? System could build on residual knowledge to
  relearn the familiar pairings.

27
Farah et al (1993)Simulation 2 Perceptual Speed

• Presented 10 familiar and 10 unfamiliar faces
  (half actors and half politicians) to the
  lesioned network.
• Measured time for the network to settle (assesses
  speed of processing).
• Settling time faster for familiar than unfamiliar
  faces, even at lesion levels that yield chance
  overt recognition. This is similar to pattern
  seen for PH when matching faces.

28
Farah et al (1993) Simulation 3Interference
Effects

• Primed the lesioned network with faces of 5
  familiar actors and 5 politicians. The prime
  face had the same occupation as the target name,
  a different occupation, or the name was presented
  alone.
• Task Identify the occupation of the target
  name.
• Number of cycles for the occupation unit to
  become active was higher when the face and name
  were from a different occupational category at
  all but the most extreme levels of damage.
• As in PH, the effect is largely interference.

29
Farah et al (1993)Conclusions

• Impaired visual representations can lead to
  failure in tests of overt recognition residual
  knowledge can support covert (implicit)
  recognition.
• Simulations demonstrate patterns of dissociations
  between overt and covert recognition similar to
  those seen in studies of prosopagnosia patients.
• Prosopagnosia patients who show a dissociation
  and those who do not may differ in the severity
  of visual impairment

30
Dissociations in Pure Alexia

• Clinical tests Pure alexics only understand
  words they have read letter-by-letter.
• Yet, pure alexics show evidence of implicit
  reading. They can make some types of decisions
  about words (especially concrete, high frequency
  words) presented too briefly to be read using a
  letter-by-letter strategy
• lexical decisions (is it a word?) semantic
  judgments (is it living or nonliving?l)
• But not
• morphological (appluadly) or rhyme/nonrhyme
  (does it rhyme with x?) decisions
• Understanding a presented word is dissociated
  from knowing about its lexical status and
  semantic category similar to students in the
  Marcel (1983) studies.
• Evidence is consistent with right-hemisphere
  mediation of implicit reading.
• This type of dissociation does not have direct
  implications for the relationship between word
  perception and visual awareness.
• Hence, evidence from pure alexia is not likely to
  be relevant to finding the neural correlates of
  visual awareness.

31
Absence of visual awareness is revealed
differently within and across the different
syndromes

• Blindsight patients
• Claim to be guessing, have no subjective
  experience
• May report a non-visual experience
• May report a visual experience
• Apperceptive patient
• Has no awareness of size orientation of objects
  that she can grasp
• Prosopagnosia patient
• Has no sense of familiarity when viewing a face,
  low confidence in face identifications
• Neglect Extinction patients
• Are unaware of having seen anything at all
• Alexia patients
• Have a dissociation between ability to report a
  specific word and the ability to make judgments
  about its lexical status or semantic category,
  but no dissociation between word perception and
  awareness of that perception.

32
Tallying up the evidence

• Privileged Role
• Blindsight (disconnection of cortical visual
  system)
• Apperceptive Agnosia (ventral stream
  disconnection)
• Integration
• Blindsight, Associative Agnosia, Prosopagnosia,
  Neglect Extinction (perceptual impairments
  interfere with integration)
• Quality of Representation
• Blindsight (limited input from LGN to
  extrastriate visual cortex)
• Associative Agnosia (degraded object
  representations)
• Prosopagnosia (perceptual impairments)
• Neglect Extinction (perceptual impairments)

33
General conclusions regarding the
vision-awareness dissociation

• A hybrid account of awareness.
• Quality of representation is critical.
• Poor quality representations cannot play a
  privileged role in awareness, nor can they be
  integrated into global awareness.
• There is no empirical support for a dedicated
  conscious awareness system.
• The mind and brain are complex. It is
  unrealistic to expect that there is one correct
  (simple) solution.

34
Some final comments

• Course Objective To provide a foundation for
  the third year psychology courses showing how
  knowledge of cognition and neuroscience interacts
  to help us understand mind and behavior.
• Cognition serves to delineate the mental and
  physical dimensions of the problem of interest.
• Neuroscience probes the brain in the search for
  neural correlates (mechanisms) that will guide
  future theoretical development.
• Today, virtually all substantive areas in
  psychology involve some cognitive exploration of
  the mental and physical behavior in question and
  most are using or exploring the utility of
  functional neuroimaging or other brain studies.