configural vs holistic processing
Holistic processing is the integration of facial features into a unified whole whereas configural is encoding of spatial distances between features
Inversion effects (Yin 1969)
when faces are upside down
○ More difficult to recognise when inverted, we do not use the same holistic processing
○ Upright and inverted faces are processed in qualitatively different ways
The thatcher effect (Thompson 1980)
features not flipped
○ Do not notice difference when presented inverted
○ We process holistically rather than looking at specific features
- The composite and part whole effects (Young et al 1987; Tanaka and Farah 1993)
○ Upper and lower facial halves of two different images are combined, yet brain processes it as one whole face
○ But fusion does not occur when the images are misaligned or inverted
○ Up right faces benefit from holistic processing only
§ Hole 1994; le grand 2006: originally used similar faces but newer procedures used more differing faces.
Composite effects criticisms
○ Konar et al 2010: lack of relationship with facial recognition ability
○ Wang et al 2012: Lack of relationship with other measures of holistic processing
○ Richler et al 2009; 2011: Attentional situational demands and response bias
○ Susilo et al 2013: Composite effects with inverted faces with larger sample sizes
Murphy et al 2016; design issues
Problems with the holistic hypothesis:
Question that does inversion impair discrimination of local features rather than holistic processing? (Leder et al 2001)
- Aperture viewing- looking at an image through a small, moveable window restricting view of the whole so image is recognised by piece, akin to serial analysis of local features
What happens when a face is inverted is still unclear
Specificity hypothesis:
Faces recruit dedicated perceptual and neural mechanisms:
- Yin 1969; Diamond and Carey 1987: inversion impairs facial recognition to a greater extent than object recognition
Composite effects are observed for faces but not other objects like cars (Cassia et al 2009)
Expertise hypothesis:
Perceptual and neural mechanisms recruited by faces overlap with objects of expertise. Regions dedicated to face recognition are also used for recognising objects of expertise
- Inversion effects for objects of expertise (Diamond and Carey 1986; Gauthier and Tarr 1997)
- Thatcher effect not face specific (Wong et al 2010)
- Composite effects seen for objects of expertise (Anstis 2005)
Fusiform face area:
- Reduced response in FFA found to inverted faces, aligning with holistic accounts but findings are slightly mixed (Kanwisher AND Yovel 2006)
- Neural basis of composite effects where greater release from adaptation when the top half if aligned with a different bottom half.
○ Not found for inverted or misaligned images (Schiltz et al 2010)
FFA also however responds to objects of expertise (Gauthier et al 2000)
- Neural basis of composite effects where greater release from adaptation when the top half if aligned with a different bottom half.
fMRI:
When faces/ objects/ scrambled faces, there are some areas which specifically respond when shown faces, withing the VOTC, particularly the fusiform face area. (Kanwisher et al 1997)
- We also have higher levels of cognition which allows us to recognise differences between faces, the lateral fusiform gyrus. The superior temporal sulcus processes perception of eye gaze, expression and lip movement.
EEG evidence:
The presentation of an upright face associated with a characteristic electrophysical response the N170 which is delayed and enhanced by inversion (Bentin et al 1999), it is not observed for objects or scrambled faces (Eimer 2000)
- Similar effects seen with composite effects
- But there is some evidence that the N170 is evoked by the eye region alone (McCarthy et al 1999) So cannot assume N170 represents holistic coding
Also found selective engagement in the FFA for objects of expertise as well as faces (Rossion et al 2002)
TMS- transcranial magnetic stimulation
TMS to the DOFA (occipital face area) disrupts face perception abilities (Pitcher et al 2007) as well as inverted face (2011), so is it disrupting facial features?
Whereas tdcs of the OFA enhances perception of both faces and objects (Barbierie et al 2016)
Prosopagnosia:
Acquired prosopagnosia:
A lifelong condition with apparent brain lesions resulting from trauma, asphyxiation, epilepsy, stroke etc.
Other visual cognition problems like object recognition also observed alongside and issues perceiving entire visual field.
Developmental prosopagnosia:
Lifelong but not associated with general intelligence, affecting people with normal vision and in the absence of apparent brain lesions.
- Affects ability to recognise familiar faces and distinguish between two unfamiliar identities (Susilo and Duchaine 2013)
- Develop compensatory strategies to cope such as voice, gair, style, distinctive features and context)
No formal diagnosis defined by DSM-V so diagnosed using self-reported tests, memory and recognition tests
Is DP domain-specific?
- There are cases of DP who show good recognition of other complex objects like cars, guns and tools (Duchaine et al 2006)
However many present object recognition difficulties and appear slower at discrimination
Is DP a deficit in holistic processing?
- DeGutis et al 2004 suggested training on holistic processing improves face recognition.
- Possible as the process measured by the composite face illusion is closely related to face recognition ability.
- However, comparing peoples sensitivity to CFI and their face recognition does not always produce intuitive results and the relationship is not so clear.
- Mixed results but majority found comparable effects in DPs and controls and some less sensitivity to the illusion by DPs
Electrophysiological findings:
Direct measure of brain’s electrical response to an event andN170 us the component associated to the neural processing of faces.
- Only some DPs show reduced N170 to upright faces, showing heterogeneity of the condition
- DP’s fail to show the typical enhancement of N170 to inverted faces
- Show normal repetition suppression effects in FFA and STS
Some studies found reduced face-selectivity in fusiform gyrus (Hadjikhani and de gelder 2002) but others failed to find significant dysfunctions (Hasson et al 2003)
Structural findings:
- Reduced grey matter in DP’s inferior temporal lobe, FFA and STS associated with face recognition performance.
- The Inferior longitudinal fasciculus connects occipital with temporal regions and associated with many behaviours including face recognition.
○ Global approach: reduced structural integrity of ILF correlates with errors in face recognition (Thomas et al 2009)
Local approach: local white matter properties of ILF, rather than the whole tract account for category specific behaviours in DP (Gomez et al 2015) The ILF connects areas that mediate different functions and so different profiles of disruption may account for different individual differences in D
- The Inferior longitudinal fasciculus connects occipital with temporal regions and associated with many behaviours including face recognition.
Genetic findings
DP tends to run in families (Duchaine, Germine and Nakayama 2007) and is worse in family members of a DP than general public
- Polk et al 2007: twin study found structural and functional mri response in the face areas is more similar in MZ than DZ
