Face perception

Question 1

What is paraeidolia?

Answer 1

Tendency to see a face where there isn’t one
New-borns will orient their gaze to face-like stimuli even when unable to do fine feature discrimination - attention directed even when face-like features scrambled up

Question 2

How does face recognition differ from object recognition?

Answer 2

More holistic and configural processing (integration across the whole)
Unreliability of feature information - similar features but differences

Question 3

How has holistic processing been demonstrated?

Answer 3

Using the inversion effect, the thatcher illusion, the part-whole effect and the composite effect

Question 4

What is the inversion effect?

Answer 4

Faces are harder to recognise upside down (not used to this orientation)
Simple feature ID would mean this wasn’t so much of a problem
Inverted luminance polarity has the same effect - changes shadows

Question 5

What is the Thatcher illusion?

Answer 5

Similar to inversion effect - when we can’t see a problem with inverted face which has altered features
Evidence that configural processing is specific to familiar orientations

Question 6

What is the part-whole effect?

Answer 6

Memory for a face part is more accurate when presented as part of the whole - faces are stored in memory as their whole

Question 7

What is the composite effect?

Answer 7

Presented with 2 different half faces it is easier to identify the 2 individuals when the face is misaligned rather than aligned to form a whole new face

Question 8

What is prosopagnosia?

Answer 8

Inability to recognise faces but able to recognise objects (can be from brain injury or through developmental disorder) - possible explanation is that facial recognition is harder than simple object recognition (discrimination WITHIN rather than between categories), or there could be a specialised brain region involved

Question 9

What evidence supports the concept of a brain area specialised for face recognition?

Answer 9

Cells in monkeys in the inferotemporal cortex sensitive to expression regardless of specific facial identity
Cells found in superior temporal sulcus sensitive to identity regardless of emotional expression
Evidence in humans = FFA in IT cortex frequently found to be damaged in people with acquired prosopagnosia, and when healthy responds twice as strongly to faces as other objects (even when just blurry suggestion of a face)

Question 10

What other areas are involved in face recognition?

Answer 10

Occipital face area and superior temporal sulcus - can have intact FFA but damaged OFA –> prosopagnosia
Suggestion of integrated OFA and FFA processing

Question 11

What is the current theoretical controversy regarding the function of the FFA?

Answer 11

Is it face-selective or simply expertise-selective?
We have more experience with faces than individual members of other categories
Inversion effect would not work so well if experience not involved
Greeble experiments
FURTHER - Is it specifically experience, or just that things we have expertise in we pay more attention to?

Question 12

What evidence questions the expertise view?

Answer 12

Technically prosopagnosics should struggle to identify non-face objects for which they have experience
These individuals actually able to gain greeble expertise, car experts who couldn’t recognise faces still managed to conduct finer discrimination of cars

So faces must have special characteristics simply not shared by other objects

Question 13

What effects can be explained through our experience with face recognition?

Answer 13

The hollow face illusion - we favour the convex interpretation so that influences how we perceive shadowing on the mask
Other-race bias - better at discriminating between individuals of own race
Caricatures - features always compared to a learned norm (of an “average face”) so we can easily recognise caricatures as they simply exaggerate the differences we already rely on for recognition purposes

Question 14

What does the Bruce and Young model of facial recognition suggest?

Answer 14

When we see a face, recognition occurs in stages:
Familiarity info e.g. face detection and structural encoding (general face structure)
Personal info e.g. face memory
Details e.g. emotion, gender
Name info

Question 15

Why was a face detection stage added into the model?

Answer 15

Finding that prosopagnosic individuals can actually detect something as a face as rapidly as healthy controls
They just struggle to subsequently recognise the face

Question 16

What has been suggested regarding putting a name to a face?

Answer 16

Usually name generation can never come before you have specifically recognised a face (one of the slowest stages)
However, research suggests that frequent exposure to names can result in the naming stage becoming faster than the personal info stage

Question 17

What evidence has suggested that identity and expression are processed by different routes?

Answer 17

Double dissociations where prosopagnosia individuals have good performance on facial recognition while poor on expression ID and then vice versa
Associated symptoms in the first type of individual suggests that some facial expression impairments may reflect damage to emotion rather than face-specific systems

Question 18

How can composite stimuli tests be used to demonstrate how emotional expression is processed?

Answer 18

Asked to decide what expression shown by bottom of face:
When same person but two different expressions response is slower
When different people with 2 different expressions there is no cost in reaction time
When different people with the same emotion the response is faster
i.e. emotional expression detected independent of identity

Question 19

What is the current conclusion regarding the routes for emotional expression and facial identification processing?

Answer 19

Unlikely to be completely separate and more likely to be partially connected
Plus, there are likely multiple systems involved in detection of emotional expression - involves emotional systems such as the amygdala to a greater extent than assumed by Bruce and Young’s model