The complex, hierarchical nature of the human visual processing system has contributed to a
great interest in visual perception and recognition. From the perception of motion to the basis of object
recognition to the effects of attention, there have been countless studies examining the neurological
underpinnings of our visual processing system. Despite this incredible wealth of research and study, or
perhaps because of it, there are some disagreements over the role and impact of particular brain areas on
visual processing. One of these debates centers on the role of the Fusiform Face Area (FFA). One side
claims that the FFA is dedicated specifically to the recognition of faces (and facial expressions; Blatt,
2013), while the other side claims that the FFA area is more generally dedicated to recognizing objects of
expertise, with faces being one of those objects.

The FFA is a part of the fusiform gyrus, located in the inferior temporal cortex (IT). The fusiform
gyrus is central to both object and face recognition (Blatt, 2013; Chang, 2017). Studies have demonstrated
that one area in the IT is more strongly activated by facial stimuli while another area is more strongly
activated by object stimuli (ex. Kobatake & Tanaka, 1994; McCarthy et al., 1997). The FFA is the name
given to the area more strongly activated by facial stimuli in humans. Aside from the clustering of
neurons that respond differentially to object and facial stimuli, there are also behavioral differences that
separate object processing from face processing. These include the face-inversion effect, which states that
inverting an image impairs face recognition more than object recognition; the part-whole effect, which
states that individuals are better able to remember individual facial features when they are tested in the
context of a whole face as opposed to isolated features; and the composite effect, which states that the
identification of one half of a face is more impaired when it is aligned with a half-face from another
individual than when the half-faces are misaligned (Kanwisher & Yovel, 2006).

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

The debate over whether or not the FFA is face specific was brought about due to one of these
behavioral differences. Diamond and Carey (1986) found dog experts show the inversion effect when

Rianda 1

attempting to “individuate dogs of the same breed” at a level that is comparable to that seen in face
recognition. This lent credence to the idea that features are more holistically processed by experts in a
way that is similar to facial processing. This has led some to claim that the FFA is a more general area
that deals with this holistic processing seen mainly in face processing and in areas of expertise, while
others maintain that the FFA is an area of face specific processing.

Proponents of the FFA as an area of face specific processing point first to evidence that object
and face processing engage different systems. Research on prosopagnosia and object agnosia seems to
support this claim. Prosopagnosia, or “face blindness”, refers to a disorder in which individuals have
difficulty recognizing faces while displaying normal sensory and cognitive abilities (Chang, 2017). Object
agnosia refers to a disorder in which individuals have difficulty recognizing objects while again
displaying normal sensory and cognitive abilities (Chang, 2017). Individuals with object agnosia can also
show good object memory – with the ability to draw the object based on memory – while still not being
able to recognize what the object is that they remember (Behrmann, Winocur, & Moscovitch, 1992).
Studies have demonstrated that individuals with prosopagnosia do not show similar deficits in
recognizing objects (Wada & Yamamoto, 2001) and that individuals with object agnosia do not show
similar deficits in recognizing faces (Moscovitch, Winocur, & Behrmann, 1997). There is also evidence
that there are different developmental mechanisms that cause each disorder (Duchaine et al., 2006). The
fact that face recognition or object recognition can be impaired while leaving the other intact supports the
presence of an area that processes face recognition and object recognition differentially, which allows for
an area like the FFA to be that face specific area.

Further supporting the idea that there is a face specific processing area is the presence of “face
patches” in primates. These face patches, located mainly in the primate superior temporal sulcus (STS),
contain neurons that respond more strongly to face stimuli as opposed to object stimuli (Tsao et al., 2003;
Fo?ldia?k et al., 2004; Desimone et al., 1984). Some have suggested that because of the similarity in
location and function of the STS in primates and the fusiform gyrus in humans the areas are homologous,

Rianda 2

with the FFA being homologous to those face patches, but there has not been enough research to prove
that the areas are indeed homologous (Yovel & Freiwald, 2013). Still, the presence of face patches in
primates indicates that it is possible that similar face specific processing areas could exist in the human
brain as well.

Proponents of the FFA as a face specific processing area then turn to evidence that the FFA
shows greater activation to facial stimuli than object stimuli. Studies using functional magnetic resonance
imaging (fMRI) have shown that the FFA in particular shows greater activation when viewing faces as
compared to objects or scrambled images (McCarthy et al., 1997; Kanwisher, McDermott, & Chun,
1997). Similar findings have been shown using positron emission tomography (PET) comparing
responses to gratings, faces, and objects (Sergent, Ohta, & Macdonald, 1992). Taken with the evidence
from examining individuals with prosopagnosia or object agnosia and from primate research, these
imaging techniques support the idea that the FFA is the area in the human brain that is specifically tuned
to face recognition.

The argument for the FFA being activated more generally in response to objects of expertise
focuses largely on studies that contradict the idea that the FFA shows greater activation to facial stimuli
than object stimuli. One particularly notable example, Gauthier et al. (1999) , showed that individuals
who were trained to differentiate “greebles” – cylindrical creatures with horns protruding from their body
in a uniform way – showed increased activation in the FFA compared to individuals who were new to
differentiating greebles. They further showed that this activation was impaired when the greebles were
inverted, similar to the face-inversion effect.

In a different study, Gauthier et al. (2000) found similar expertise effect in response to cars and
birds as opposed to objects, with activation in the FFA increasing as expertise increased. Individuals
viewing the object outside of their area of expertise (birds for the car experts and cars for the bird experts)
showed decreased activation compared with their expert counterparts, indicating that it was not simply an
effect of individuals having more familiarity with birds or cars. Both of these studies support the expertise

Rianda 3

hypothesis in two important ways. First, they both indicate that the FFA can be differentially activated by
more than just faces. Secondly, they show that the activation is modulated based on level of expertise.

Some studies that support the more general expertise argument focus solely on showing that the
FFA can be activated by more than just faces. For example, Chao, Martin, and Haxby (1999) found that
pictures of human faces, animals, and animals where the face had been removed caused greater activation
in the FFA than pictures of houses. The further found that the pictures of human faces, animals, and
faceless animals showed greater activation in the FFA than the area of the fusiform gyrus that responded
to objects, like the pictures of houses. Taken in conjunction with other data indicating that face
recognition and object recognition activate different areas in the brain, this would seem to suggest that
animal recognition “relies strongly on the same neural substrates that represent faces” (Chao et al., 1999).

In order to evaluate the relative merits of each side of the debate, it is important to understand
some of the important critiques of the evidence presented by each side. Kanwisher and Yovel (2006) level
several criticisms at Gauthier et al.’s greebles (Gauthier et al., 1999). They argue that the greebles are too
face-like (or body-like) to show that the FFA is not face-specific. They also claimed that the region of
interest (ROI) used by Gauthier et al. included the fusiform body area (FBA) which is tuned to objects
that look like human bodies. They also criticized Gauthier et al.’s study on bird and car experts (Gauthier
et al., 2000). They brought up replication concerns, noting that the “result has been replicated in one
study, but produced only a marginally significant trend in another study, and no effect at all in another.”
They also stated that “the effect size is very small and the response to faces… remains at least twice as
high as to any objects of expertise.”

There are also concerns with the initial Diamond and Carey (1986) experiment. Robbins and
McKone (2007) attempted to replicate Diamond and Carey’s (1986) face-inversion effect in dog experts,
extending their study to examine the composite effect and “sensitivity to contrast reversal” as well. They
found no significant effects. This calls into question whether or not Diamond and Carey’s (1986) findings

Rianda 4

are replicable. If they are not, then the behavioral component of the expertise argument would be rendered
unreliable.

On the other side, there have been criticisms of how well prosopagnosia supports the face-
specific argument. One study showed that some individuals with prosopagnosia are still able to recognize
certain facial features – namely facial expression, gender and age – despite not being able to recognize
face identity (Tranel, Damasio, & Damasio, 1988). This seems to support the idea that face recognition is
a more holistic form of visual processing, which better supports the expertise argument. Another study
indicated that individuals with prosopagnosia do show deficits on object recognition as compared to
healthy controls, undermining the idea that prosopagnosia supports differentiation between face and
object recognition (Gauthier, Behrmann, & Tarr, 1999).

It would seem that the most parsimonious explanation for the evidence given is that the FFA is
differentially activated by more than just facial stimuli, but that it responds particularly strongly to facial
stimuli. Whether or not differences in levels of activation are the result of expertise or some other factor
remain in dispute. The evidence in favor of expertise did not seem to prove definitively that there was no
other factor that could explain the increase in activation. The reason that items other than faces might
differentially activate the FFA could be that there are certain neurons within the FFA that are tuned to
more than just facial stimuli. For example, Rhodes et al. (2004) found that Lepidoptera (butterflies and
moths) experts showed activation in 15% of FFA voxels. While this was not shown to be significantly
different from control objects in the study, it does indicate that there could be areas of the FFA that
respond to items other than faces. 

Author