Transcript SIZE INVARIANT ADAPTATION TO FACES IN THE FUSIFORM

SIZE INVARIANT ADAPTATION TO FACES IN THE
FUSIFORM GYRUS
Timothy J. Andrews, Michael P. Ewbank & Caroline Johnson
Dept of Psychology, Wolfson Research Institute, University of Durham, UK
INTRODUCTION
3
sagittal
STS
STS
PG
FG
LO
LO
FG
MO
B
Fusiform (F>O)
Superior Temporal (F>O)
Lateral Occipital (F>T)
4
% MR response
PG
4
3
4
3
3
2
2
2
1
1
1
0
0
0
5
10
15
20
Face
Object
Place
Texture
0
0
5
10
15
20
0
5
10
15
20
C
Lateral Occipital (O>F)
Parahippocampal (P>F)
% MR response
4
4
4
3
Medial Occipital (T>F)
3
3
2
2
2
1
1
1
0
0
0
0
5
10
15
20
0
5
10
Time (s)
15
20
0
5
10
15
20
Regions of the fusiform
gyrus (FG), lateral
occipital lobe (LO) and
superior temporal sulcus
(STS) were more active
when subjects viewed
photographs of faces (F)
than when they viewed
objects (O) or textures
(T).
In contrast, areas in the
parahippocampal gyrus
(PG) and other regions of
the lateral occipital lobe
(LO) were more active
when subjects viewed
scenes (P) and objects
(O) compared to faces.
Importantly, significant
responses to the nonpreferred stimuli were
found in all regions.
2
% MR response
Same
Different
*
2
1
2
1
0
1
0
0
C
5
10
15
0
20
0
5
10
15
20
0
3
3
3
2
2
2
1
1
1
0
0
0
0
5
10
15
20
0
5
10
15
5
10
15
20
Medial Occipital (T>F)
Lateral Occipital (O>F)
Parahippocampal (P>F)
Face selective voxels in both
the FG and LO showed
reduced response to the
repeated presentation of
identical faces compared with
different faces.
In contrast, activity levels in
the face-selective region of
the STS and the non-face
selective regions failed to
show adaptation to faces.
20
0
5
10
15
20
(***=P<0.01; *=p<0.1; n=8)
Time (s)
B
Fusiform (F>O)
Lateral Occipital (F>T)
3
4
Superior Temporal (F>O)
3
Same
Different
3
2
2
*
2
1
1
1
0
0
C
0
0
5
10
15
20
0
5
10
15
20
0
3
3
3
2
2
2
1
1
1
0
0
0
0
5
10
15
20
0
5
10
15
Time (s)
20
10
15
20
Medial Occipital (T>F)
Lateral Occipital (O>F)
Parahippocampal (P>F)
5
0
5
10
15
20
Finally, we compared the response
to repeated presentations of the
same face that varied in gaze and
expression compared to images of
different faces that showed
similar changes in viewpoint. We
predicted that, if a region was
invariant to viewpoint, a reduced
response should be apparent for
the same face condition . On the
other hand, if an area represented
changeable aspects of faces, we
would expect a maximal response
same face condition.
Face selective voxels in the STS
showed an increased response for
identical faces viewed at different
viewpoints.
A difference between these
conditions was not observed in any
other face or non-face selective
region.
(***=P<0.01; n=8)
EXPERIMENT 2: VARY SIZE
A
B
Fusiform (F>O)
Lateral Occipital (F>T)
3
3
Superior Temporal (F>O)
3
***
% MR response
coronal
3
***
Same
Different
**
2
2
2
1
1
1
0
0
0
0
C
5
10
15
20
0
5
10
15
20
0
3
3
3
2
2
2
1
1
1
0
0
0
0
5
10
15
20
0
5
10
15
Time (s)
20
5
10
15
20
Medial Occipital (T>F)
Lateral Occipital (O>F)
Parahippocampal (P>F)
% MR response
axial
Superior Temporal (F>O)
3
LOCALISER SCAN
A
Lateral Occipital (F>T)
A
% MR response
Fusiform (F>O)
B
The response to repeated
presentations of the same
face was compared to images
of different faces Our
prediction was that areas
involved in face recognition
would be less active during the
same face condition
EXPERIMENT 3: VARY VIEWPOINT
% MR response
A
% MR response
Recognising complex objects, such as faces, is a
simple process for most human observers. However,
as we move about, the size and shape of the retinal
image also changes. To be useful, the visual system
must must represent the unchanging features of a
face that specify identity, as well as the changeable
aspects of a face that facilitate social
communication.
Although models of face processing have
proposed ways to deal with these different tasks, it
remains unclear how these mechanisms might be
implemented in visual cortex. One possibility is that
information is processed in specialised modules; an
opposing
model
appeals
to
a
distributed
representation across many visual areas. To address
this debate, we used the technique of fMRadaptation (the reduction in activity following
repeated presentation of identical images) .
EXPERIMENT 1: FACE ADAPTATION
Next, we compared the
response to repeated
presentations of images of the
same face that varied in size
compared to images of
different faces that also
varied in size. If the
representation of faces in
these areas is size-invariant,
we would expect a relatively
lower response to the same
face.
The results show that
adaptation to faces in the FG
and LO was size invariant.
No such adaptation was
observed in other face or nonface selective regions.
0
5
10
15
20
(***=P<0.01; **=p<0.05; n=8)
CONCLUSIONS
1) Face recognition is dependent upon the
activity of face-selective regions in the
inferior temporal lobe.
2) Perception of changeable aspects of
faces (such as eye gaze and expression)
is dependent on face-selective regions in
the superior temporal lobe.
3) These results challenge the idea that
object-selective regions are also involved
in the neural processing of faces.