Biological basis of behaviour
Behavioural face recognition
● Face recognition skills are very developed, able to recognise within hundreds
of milliseconds. We can also extract much of the information about a person
from this, such as race, gender direction of gaze and emotional cues.
● There are 2 main theories of face recognition. The specificity account (facial
recognition is a unqiue part of recognition, (McKone & Kanwisher 2005)) and
the expertise account (we can recognise faces because of frequent exposure
to them (Diamond and Carey 1986).
● The face inversion effect is the difficulty to recognise
upside down faces. The initial study from Yin 1969 is
one of the most influential in the space. It found the
large disparity in ability to recognise inverted faces
and normal, and this effect was not found with planes
and only a small effect with houses. This implies our
recognition of faces is specific, not the same as
general recognition. Therefore it supports face
specificity
● However, further research disputed this specificity.
Diamond and Carey (1986) investigated dog
breeders and novices with both dog and human
upside down photos. They showed them pictures of
upright and inverted stimuli and asked them to
recognize whether which of two options matched.
They found that dog breeders had equal incorrect
percentages for both faces and dogs, whereas novices only experience the
effect in faces. This supports the expertise account.
● FIrst order configurations are the general average representation of spacing
between features. Eg we have an average distance between eyes and lips in
mind for humans.
● Second order configurations are the spacing distances between features for
the person we are observing. We compare these to the average to recognise
the persons face (according to Diamond and Carey)
● Holistic processing is the idea we process the face as a ‘gestalt’. This means
using all the features to process it as a whole. Contrastingly, featural
processing is when there is a focus on specific individual features for
recognition.
● Diamond and Carey suggest the inversion effect occurs because we struggle
to judge the second order configurations.
● The thatcher illusion refers to the inability to realise that facial features are
rotated when the whole image is rotated. This is suggested to be because we
, lack expertise on rotated faces anyway, so the rotation of facial features is still
hard to distinct
● Face recognition in animals has also been researched. They commonly use
chimps, as they have very strong short term memories, and it is assumed they
have a form of eidetic memory. Research from the facial
recognition of chimps has found that they have similar
deficiencies in recognising inverted faces yet did not
have these in houses. This is the same result found in
humans, and so it suggests facial recognition may not
be a uniquely human trait (Parr and Heintz 2006).
● A follow up study on chimps found that they were not affected by the thatcher
illusion, a result different to humans.
● Further included the breakdown of the chimp faces into different components,
disrupting either the first order information or both
first and second order information. They found that
performance was significantly worse when first
order was disrupted, and even worse when first and
second order was disrupted. This suggests that
they use first order processing similar to humans
(Parr et al 2006).
● A follow up included the investigation of chimps recognition with different
photo manipulations and found they were significantly
worse at large pixel manipulated photos. They
performed to normal levels when pictures only had
eyes covered, or when pictures were pixelated with
small pixels. The large pixels disrupted first and
second order configuration the most, so this suggest
that these configurations were to key (Parr et al 2006).
Behavioural Face recognition
● FMRI is a non invasive techniques which detected activity by changes in
oxygen levels in the brain.
● The FFA (Face Fusiform Area) is a cortical region in the fusiform gyrus that is
highly activated during facial recognition, supports specificity account
(Kanwisher et al 1997)
● Gauthier et al (1999) - used ‘Greebles’ which were unlike faces, but had
specific facial features such as a tongue or ears. They found that the FFA can
also be activated by Greebles, as long as people are familiarised with them,
supporting an expertise account. There is no activation in FFA when people
are unfamiliar with greebles.
Behavioural face recognition
● Face recognition skills are very developed, able to recognise within hundreds
of milliseconds. We can also extract much of the information about a person
from this, such as race, gender direction of gaze and emotional cues.
● There are 2 main theories of face recognition. The specificity account (facial
recognition is a unqiue part of recognition, (McKone & Kanwisher 2005)) and
the expertise account (we can recognise faces because of frequent exposure
to them (Diamond and Carey 1986).
● The face inversion effect is the difficulty to recognise
upside down faces. The initial study from Yin 1969 is
one of the most influential in the space. It found the
large disparity in ability to recognise inverted faces
and normal, and this effect was not found with planes
and only a small effect with houses. This implies our
recognition of faces is specific, not the same as
general recognition. Therefore it supports face
specificity
● However, further research disputed this specificity.
Diamond and Carey (1986) investigated dog
breeders and novices with both dog and human
upside down photos. They showed them pictures of
upright and inverted stimuli and asked them to
recognize whether which of two options matched.
They found that dog breeders had equal incorrect
percentages for both faces and dogs, whereas novices only experience the
effect in faces. This supports the expertise account.
● FIrst order configurations are the general average representation of spacing
between features. Eg we have an average distance between eyes and lips in
mind for humans.
● Second order configurations are the spacing distances between features for
the person we are observing. We compare these to the average to recognise
the persons face (according to Diamond and Carey)
● Holistic processing is the idea we process the face as a ‘gestalt’. This means
using all the features to process it as a whole. Contrastingly, featural
processing is when there is a focus on specific individual features for
recognition.
● Diamond and Carey suggest the inversion effect occurs because we struggle
to judge the second order configurations.
● The thatcher illusion refers to the inability to realise that facial features are
rotated when the whole image is rotated. This is suggested to be because we
, lack expertise on rotated faces anyway, so the rotation of facial features is still
hard to distinct
● Face recognition in animals has also been researched. They commonly use
chimps, as they have very strong short term memories, and it is assumed they
have a form of eidetic memory. Research from the facial
recognition of chimps has found that they have similar
deficiencies in recognising inverted faces yet did not
have these in houses. This is the same result found in
humans, and so it suggests facial recognition may not
be a uniquely human trait (Parr and Heintz 2006).
● A follow up study on chimps found that they were not affected by the thatcher
illusion, a result different to humans.
● Further included the breakdown of the chimp faces into different components,
disrupting either the first order information or both
first and second order information. They found that
performance was significantly worse when first
order was disrupted, and even worse when first and
second order was disrupted. This suggests that
they use first order processing similar to humans
(Parr et al 2006).
● A follow up included the investigation of chimps recognition with different
photo manipulations and found they were significantly
worse at large pixel manipulated photos. They
performed to normal levels when pictures only had
eyes covered, or when pictures were pixelated with
small pixels. The large pixels disrupted first and
second order configuration the most, so this suggest
that these configurations were to key (Parr et al 2006).
Behavioural Face recognition
● FMRI is a non invasive techniques which detected activity by changes in
oxygen levels in the brain.
● The FFA (Face Fusiform Area) is a cortical region in the fusiform gyrus that is
highly activated during facial recognition, supports specificity account
(Kanwisher et al 1997)
● Gauthier et al (1999) - used ‘Greebles’ which were unlike faces, but had
specific facial features such as a tongue or ears. They found that the FFA can
also be activated by Greebles, as long as people are familiarised with them,
supporting an expertise account. There is no activation in FFA when people
are unfamiliar with greebles.
