In cognitive neuroscience, visual modularity is an organizational concept concerning how vision works. The way in which the primatevisual system operates is currently under intense scientific scrutiny. One dominant thesis is that different properties of the visual world (color, motion, form
and so forth) require different computational solutions which are
implemented in anatomically/functionally distinct regions that operate
independently – that is, in a modular fashion.
Motion processing
Akinetopsia is an intriguing condition brought about by damage to the Extrastriate cortex MT+ that renders humans and monkeys unable to perceive motion, seeing the world in a series of static "frames" instead
and indicates that there might be a "motion centre" in the brain. Of
course, such data can only indicate that this area is at least necessary
to motion perception, not that it is sufficient; however, other
evidence has shown the importance of this area to primate motion
perception. Specifically, physiological, neuroimaging, perceptual,
electrical- and transcranial magnetic stimulation
evidence (Table 1) all come together on the area V5/hMT+. Converging
evidence of this type is supportive of a module for motion processing.
However, this view is likely to be incomplete: other areas are involved
with motion perception, including V1, V2 and V3a and areas surrounding V5/hMT+ (Table 2). A recent fMRI study put the number of motion areas at twenty-one.
Clearly, this constitutes a stream of diverse anatomical areas. The
extent to which this is ‘pure’ is in question: with Akinetopsia come
severe difficulties in obtaining structure from motion. V5/hMT+ has since been implicated in this function as well as determining depth.
Thus the current evidence suggests that motion processing occurs in a
modular stream, although with a role in form and depth perception at
higher levels.
Evidence for a "motion centre" in the primate brain
Similar converging evidence suggests modularity for color. Beginning with Gowers’ finding that damage to the fusiform/lingual gyri in occipitotemporal cortex correlates with a loss in color perception (achromatopsia), the notion of a "color centre" in the primate brain has had growing support. Again, such clinical evidence only implies that this region is critical to color perception, and nothing more. Other evidence, however, including neuroimaging and physiology converges on V4 as necessary to color perception. A recent meta-analysis has also shown a specific lesion common to achromats corresponding to V4. From another direction altogether it has been found that when synaesthetes experience color by a non-visual stimulus, V4 is active.
On the basis of this evidence it would seem that color processing is
modular. However, as with motion processing it is likely that this
conclusion is inaccurate. Other evidence shown in Table 3 implies
different areas’ involvement with color. It may thus be more instructive
to consider a multistage color processing stream from the retina
through to cortical areas including at least V1, V2, V4, PITd and TEO. Consonant with motion perception, there appears to be a constellation of areas drawn upon for color perception.
In addition, V4 may have a special, but not exclusive, role. For
example, single cell recording has shown that only V4 cells respond to
the color of a stimuli rather than its waveband, whereas other areas
involved with color do not.
Evidence against a "color center" in the primate brain
Other areas involved with color/Other functions of V4
Another clinical case that would a priori suggest a module for modularity in visual processing is visual agnosia. The well studied patient DF is unable to recognize or discriminate objects
owing to damage in areas of the lateral occipital cortex although she
can see scenes without problem – she can literally see the forest but
not the trees. Neuroimaging
of intact individuals reveals strong occipito-temporal activation
during object presentation and greater activation still for object
recognition.
Of course, such activation could be due to other processes, such as
visual attention. However, other evidence that shows a tight coupling of
perceptual and physiological changes
suggests activation in this area does underpin object recognition.
Within these regions are more specialized areas for face or fine grained
analysis, place perception and human body perception. Perhaps some of the strongest evidence for the modular nature of these processing systems is the double dissociation between object- and face (prosop-) agnosia. However, as with color and motion, early areas are implicated too, lending support to the
idea of a multistage stream terminating in the inferotemporal cortex
rather than an isolated module.
Functional modularity
One of the first uses of the term "module" or "modularity" occurs in the influential book "Modularity of Mind" by philosopher Jerry Fodor.
A detailed application of this idea to the case of vision was published
by Pylyshyn (1999), who argued that there is a significant part of
vision that is not responsive to beliefs and is "cognitively
impenetrable".
Much of the confusion concerning modularity exists in
neuroscience because there is evidence for specific areas (e.g. V4 or
V5/hMT+) and the concomitant behavioral deficits following brain insult
(thus taken as evidence for modularity). In addition, evidence shows
other areas are involved and that these areas subserve processing of
multiple properties (e.g. V1)
(thus taken as evidence against modularity). That these streams have
the same implementation in early visual areas, like V1, is not
inconsistent with a modular viewpoint: to adopt the canonical analogy in
cognition, it is possible for different software to run on the same
hardware. A consideration of psychophysics
and neuropsychological data would suggest support for this. For
example, psychophysics has shown that percepts for different properties
are realized asynchronously. In addition, although achromats experience other cognitive defects they do not have motion deficits when their lesion is restricted to V4, or total loss of form perception. Relatedly, Zihl and colleagues' akinetopsia
patient shows no deficit to color or object perception (although
deriving depth and structure from motion is problematic, see above) and
object agnostics do not have damaged motion or color perception, making
the three disorders triply dissociable.
Taken together this evidence suggests that even though distinct
properties may employ the same early visual areas they are functionally
independent. Furthermore, that the intensity of subjective perceptual
experience (e.g. color) correlates with activity in these specific areas
(e.g. V4), the recent evidence that synaesthetes
show V4 activation during the perceptual experience of color, as well
as the fact that damage to these areas results in concomitant behavioral
deficits (the processing may be occurring but perceivers do not have
access to the information) are all evidence for visual modularity.
Modularity of mind is the notion that a mind may, at least in part, be composed of innate neural structures or mental modules which have distinct, established, and evolutionarily developed functions. However, different definitions of "module" have been proposed by different authors.
Early investigations
Historically, questions regarding the functional architecture
of the mind have been divided into two different theories of the nature
of the faculties. The first can be characterized as a horizontal view
because it refers to mental processes as if they are interactions
between faculties such as memory, imagination, judgement, and
perception, which are not domain specific
(e.g., a judgement remains a judgement whether it refers to a
perceptual experience or to the conceptualization/comprehension
process). The second can be characterized as a vertical view because it
claims that the mental faculties are differentiated on the basis of
domain specificity, are genetically determined, are associated with
distinct neurological structures, and are computationally autonomous.
The vertical vision goes back to the 19th century movement called phrenology and its founder Franz Joseph Gall,
who claimed that the individual mental faculties could be associated
precisely, in a sort of one-to-one correspondence, with specific
physical areas of the brain. Hence, someone's level of intelligence, for
example, could be literally "read off" from the size of a particular
bump on his posterior parietal lobe. This simplistic view of modularity
has been disproven over the course of the last century.
Fodor's Modularity of Mind
In the 1980s, however, Jerry Fodor revived the idea of the modularity of mind, although without the notion of precise physical localizability. Drawing from Noam Chomsky's idea of the language acquisition device and other work in linguistics as well as from the philosophy of mind and the implications of optical illusions, he became a major proponent of the idea with the 1983 publication of Modularity of Mind.
According to Fodor, a module falls somewhere between the behaviorist and cognitivist views of lower-level processes.
Behaviorists
tried to replace the mind with reflexes which Fodor describes as
encapsulated (cognitively impenetrable or unaffected by other cognitive
domains) and non-inferential (straight pathways with no information
added). Low level processes are unlike reflexes in that they are
inferential. This can be demonstrated by poverty of the stimulus
arguments in which the proximate stimulus, that which is initially
received by the brain (such as the 2D image received by the retina),
cannot account for the resulting output (for example, our 3D perception
of the world), thus necessitating some form of computation.
In contrast, cognitivists
saw lower level processes as continuous with higher level processes,
being inferential and cognitively penetrable (influenced by other
cognitive domains, such as beliefs). The latter has been shown to be
untrue in some cases, such as with many visual illusions (ex. Müller-Lyer illusion),
which can persist despite a person's awareness of their existence. This
is taken to indicate that other domains, including one's beliefs,
cannot influence such processes.
Fodor arrives at the conclusion that such processes are
inferential like higher order processes and encapsulated in the same
sense as reflexes.
Although he argued for the modularity of "lower level" cognitive processes in Modularity of Mind he also argued that higher level cognitive processes are not modular since they have dissimilar properties. The Mind Doesn't Work That Way, a reaction to Steven Pinker's How the Mind Works, is devoted to this subject.
Fodor (1983) states that modular systems must—at least to "some interesting extent"—fulfill certain properties:
Domain specificity: modules only operate on certain kinds of inputs—they are specialised
Informational encapsulation: modules need not refer to other psychological systems in order to operate
Obligatory firing: modules process in a mandatory manner
Fast speed: probably due to the fact that they are encapsulated
(thereby needing only to consult a restricted database) and mandatory
(time need not be wasted in determining whether or not to process
incoming input)
Shallow outputs: the output of modules is very simple
Limited accessibility
Characteristic ontogeny: there is a regularity of development
Fixed neural architecture.
Pylyshyn
(1999) has argued that while these properties tend to occur with
modules, one—information encapsulation—stands out as being the real
signature of a module; that is the encapsulation of the processes inside
the module from both cognitive influence and from cognitive access. One example is that conscious awareness of the Müller-Lyer illusion being an illusion does not correct visual processing.
Evolutionary psychology and massive modularity
Other perspectives on modularity come from evolutionary psychology, particularly from the work of Leda Cosmides and John Tooby.
This perspective suggests that modules are units of mental processing
that evolved in response to selection pressures. On this view, much
modern human psychological activity is rooted in adaptations that
occurred earlier in human evolution, when natural selection was forming the modern human species.
Evolutionary psychologists propose that the mind is made up of genetically influenced and domain-specific mental algorithms or computational modules, designed to solve specific evolutionary problems of the past. Cosmides and Tooby also state in a brief "primer" on their website,
that "…the brain is a physical system. It functions like a computer,"
"…the brain’s function is to process information," "different neural
circuits are specialized for solving different adaptive problems," and
"our modern skulls house a stone age mind."
The definition of module has caused confusion and dispute.
J. A. Fodor initially defined module as "functionally specialized
cognitive systems" that have nine features but not necessarily all at
the same time. In his views modules can be found in peripheral
processing such as low-level visual processing but not in central
processing. Later he narrowed the two essential features to domain-specificity and information encapsulation. Frankenhuis and Ploeger
write that domain-specificity means that "a given cognitive mechanism
accepts, or is specialized to operate on, only a specific class of
information". Information encapsulation means that information
processing in the module cannot be affected by information in the rest
of the brain. One example is that being aware that a certain optical
illusion, caused by low level processing, is false does not prevent the
illusion from persisting.
Evolutionary psychologists instead usually define modules as
functionally specialized cognitive systems that are domain-specific and
may also contain innate knowledge about the class of information
processed. Modules can be found also for central processing. This
theory is sometimes referred to as massive modularity.
A 2010 review by evolutionary psychologists Confer et al.
suggested that domain general theories, such as for "rationality," has
several problems: 1. Evolutionary theories using the idea of numerous
domain-specific adaptions have produced testable predictions that have
been empirically confirmed; the theory of domain-general rational
thought has produced no such predictions or confirmations. 2. The
rapidity of responses such as jealousy due to infidelity indicates a
domain-specific dedicated module rather than a general, deliberate,
rational calculation of consequences. 3. Reactions may occur
instinctively (consistent with innate knowledge) even if a person has
not learned such knowledge. One example being that in the ancestral
environment it is unlikely that males during development learn that
infidelity (usually secret) may cause paternal uncertainty (from
observing the phenotypes of children born many months later and making a
statistical conclusion from the phenotype dissimilarity to the
cuckolded fathers). With respect to general purpose problem solvers, Barkow, Cosmides, and Tooby (1992) have suggested in The Adapted Mind: Evolutionary Psychology and The Generation of Culture that a purely general problem solving mechanism is impossible to build due to the frame problem.
Clune et al. (2013) have argued that computer simulations of the
evolution of neural nets suggest that modularity evolves because,
compared to non-modular networks, connection costs are lower.
Several groups of critics, including psychologists working within evolutionary frameworks,
argue that the massively modular theory of mind does little to explain
adaptive psychological traits. Proponents of other models of the mind
argue that the computational theory of mind
is no better at explaining human behavior than a theory with mind
entirely a product of the environment. Even within evolutionary
psychology there is discussion about the degree of modularity, either as
a few generalist modules or as many highly specific modules. Other critics suggest that there is little empirical support in favor of the domain-specific theory beyond performance on the Wason selection task, a task critics state is too limited in scope to test all relevant aspects of reasoning.
Moreover, critics argue that Cosmides and Tooby's conclusions contain
several inferential errors and that the authors use untested
evolutionary assumptions to eliminate rival reasoning theories.
Wallace (2010) observes that the evolutionary psychologists' definition of "mind" has been heavily influenced by cognitivism and/or information processing definitions of the mind.
Critics point out that these assumptions underlying evolutionary
psychologists' hypotheses are controversial and have been contested by
some psychologists, philosophers, and neuroscientists. For example, Jaak Panksepp,
an affective neuroscientist, point to the "remarkable degree of
neocortical plasticity within the human brain, especially during
development" and states that "the developmental interactions among
ancient special-purpose circuits and more recent general-purpose brain
mechanisms can generate many of the "modularized" human abilities that
evolutionary psychology has entertained."
Philosopher David Buller agrees with the general argument that
the human mind has evolved over time but disagrees with the specific
claims evolutionary psychologists make. He has argued that the
contention that the mind consists of thousands of modules, including
sexually dimorphic jealousy and parental investment modules, are
unsupported by the available empirical evidence. He has suggested that the "modules" result from the brain's developmental plasticity and that they are adaptive responses to local conditions, not past evolutionary environments.
However, Buller has also stated that even if massive modularity is
false this does not necessarily have broad implications for evolutionary
psychology. Evolution may create innate motives even without innate
knowledge.
In contrast to modular mental structure, some theories posit domain-general processing,
in which mental activity is distributed across the brain and cannot be
decomposed, even abstractly, into independent units. A staunch defender
of this view is William Uttal, who argues in The New Phrenology
(2003) that there are serious philosophical, theoretical, and
methodological problems with the entire enterprise of trying to localise
cognitive processes in the brain. Part of this argument is that a successful taxonomy of mental processes has yet to be developed.
Merlin Donald argues that over evolutionary time the mind has gained adaptive advantage from being a general problem solver.
The mind, as described by Donald, includes module-like "central"
mechanisms, in addition to more recently evolved "domain-general"
mechanisms.
The
sketch of the idea of the grandmother cell: a neuron that reacts
selectively on a pattern: Jennifer Aniston cell, Dodecahedron cell, and
'Grandmother cell' cell, which reacts on the copy of this cartoon.
The grandmother cell, sometimes called the "Jennifer Aniston neuron", is a hypothetical neuron that represents a complex but specific concept or object. It activates when a person "sees, hears, or otherwise sensibly discriminates"
a specific entity, such as his or her grandmother. The term was in use
at least as early as 1966 amongst staff and students in the Department
of Experimental Psychology, University of Cambridge, England. A similar
concept, that of the gnostic neuron, was proposed two years later by Jerzy Konorski.
Support
Face selective cells
Visual neurons in the inferior temporal cortex of the monkey fire selectively to hands and faces.
These cells are selective in that they do not fire for other visual
objects important for monkeys such as fruit and genitalia. Research
finds that some of these cells can be trained to show high specificity
for arbitrary visual objects, and these would seem to fit the
requirements of gnostic/grandmother cells. In addition, evidence exists for cells in the human hippocampus that have highly selective responses to gnostic categories including highly selective responses to individual human faces.
However most of the reported face-selective cells are not
grandmother/gnostic cells since they do not represent a specific
percept, that is, they are not cells narrowly selective in their
activations for one face and only one face irrespective of
transformations of size, orientation, and color. Even the most selective
face cells usually also discharge, if more weakly, to a variety of
individual faces. Furthermore, face-selective cells often vary in their
responsiveness to different aspects of faces. This suggests that cell
responsiveness arises from the need of a monkey to differentiate among
different individual faces rather than among other categories of stimuli
such as bananas with their discrimination properties linked to the fact
that different individual faces are much more similar to each other in
their overall organization and fine detail than other kinds of stimuli.
Moreover, it has been suggested that these cells might in fact be
responding as specialized feature detector neurons that only function in
the holistic context of a face construct.
One idea has been that such cells form ensembles for the coarse
or distributed coding of faces rather than detectors for specific faces.
Thus, a specific grandmother may be represented by a specialized
ensemble of grandmother or near grandmother cells.
Individual specific recognition cells
In 2005, a UCLA and Caltech study found evidence of different cells that fire in response to particular people, such as Bill Clinton or Jennifer Aniston. A neuron for Halle Berry, for example, might respond "to the concept, the abstract entity, of Halle Berry", and would fire not only for images of Halle Berry, but also to the actual name "Halle Berry".
However, there is no suggestion in that study that only the cell being
monitored responded to that concept, nor was it suggested that no other
actress would cause that cell to respond (although several other
presented images of actresses did not cause it to respond). The researchers believe that they have found evidence for sparseness, rather than for grandmother cells.
Further evidence for the theory that a small neural network
provides facial recognition was found from analysis of cell recording
studies of macaque monkeys. By formatting faces as points in a
high-dimensional linear space, the scientists discovered that each face
cell’s firing rate is proportional to the projection of an incoming face
stimulus onto a single axis in this space, allowing a face cell
ensemble of about 200 cells to encode the location of any face in the
space.
Sparseness vs distributed representations
The grandmother cell hypothesis, is an extreme version of the idea of sparseness,
and is not without critics. The opposite of the grandmother cell theory
is the distributed representation theory, that states that a specific
stimulus is coded by its unique pattern of activity over a large group
of neurons widely distributed in the brain.
The arguments against the sparseness include:
According to some theories, one would need thousands of cells
for each face, as any given face must be recognised from many different
angles – profile, 3/4 view, full frontal, from above, etc.
Rather than becoming more and more specific as visual processing
proceeds from retina through the different visual centres of the brain,
the image is partially dissected into basic features such as vertical
lines, colour, speed, etc., distributed in various modules
separated by relatively large distances. How all these disparate
features are re-integrated to form a seamless whole is known as the binding problem.
Pontifical cells
William James in 1890 proposed a related idea of a pontifical cell.
The pontifical cell is defined as a putative, and implausible cell
which had all our experiences. It is in this different from a concept
specific cell in that it is the site of experience of sense data.
James's 1890 pontifical cell was instead a cell "to which the rest of
the brain provided a representation" of a grandmother. The experience of
grandmother occurred in this cell.
Human brain, bottom view. Fusiform face area shown in bright blue.
Computer-enhanced fMRI
scan of a person who has been asked to look at faces. The image shows
increased blood flow in cerebral cortex that recognizes faces (FFA).
The FFA was discovered and continues to be investigated in humans using positron emission tomography (PET) and functional magnetic resonance imaging
(fMRI) studies. Usually, a participant views images of faces, objects,
places, bodies, scrambled faces, scrambled objects, scrambled places,
and scrambled bodies. This is called a functional localizer.
Comparing the neural response between faces and scrambled faces will
reveal areas that are face-responsive, while comparing cortical
activation between faces and objects will reveal areas that are
face-selective.
Function
The human FFA was first described by Justine Sergent in 1992 and later named by Nancy Kanwisher in 1997 who proposed that the existence of the FFA is evidence for domain specificity
in the visual system. Studies have recently shown that the FFA is
composed of functional clusters that are at a finer spatial scale than
prior investigations have measured.
Electrical stimulation of these functional clusters selectively
distorts face perception, which is causal support for the role of these
functional clusters in perceiving the facial image.
While it is generally agreed that the FFA responds more to faces than
to most other categories, there is debate about whether the FFA is
uniquely dedicated to face processing, as proposed by Nancy Kanwisher
and others, or whether it participates in the processing of other
objects. The expertise hypothesis, as championed by Isabel Gauthier
and others, offers an explanation for how the FFA becomes selective for
faces in most people. The expertise hypothesis suggests that the FFA is
a critical part of a network that is important for individuating
objects that are visually similar because they share a common
configuration of parts. Gauthier et al., in an adversarial collaboration
with Kanwisher,
tested both car and bird experts, and found some activation in the FFA
when car experts were identifying cars and when bird experts were
identifying birds. This finding has been replicated, and expertise effects in the FFA have been found for other categories such as chess displays and x-rays.
Recently, it was found that the thickness of the cortex in the FFA
predicts the ability to recognize faces as well as vehicles.
A 2009 magnetoencephalography study found that objects incidentally perceived as faces, an example of pareidolia,
evoke an early (165-millisecond) activation in the FFA, at a time and
location similar to that evoked by faces, whereas other common objects
do not evoke such activation. This activation is similar to a face-specific ERP component N170.
The authors suggest that face perception evoked by face-like objects is
a relatively early process, and not a late cognitive reinterpretation
phenomenon.
One case study of agnosia provided evidence that faces are processed in a special way. A patient known as C. K., who suffered brain damage as a result of a car accident, later developed object agnosia.
He experienced great difficulty with basic-level object recognition,
also extending to body parts, but performed very well at recognizing
faces.
A later study showed that C. K. was unable to recognize faces that were
inverted or otherwise distorted, even in cases where they could easily
be identified by normal subjects. This is taken as evidence that the fusiform face area is specialized for processing faces in a normal orientation.
Studies using functional magnetic resonance imaging and electrocorticography have demonstrated that activity in the FFA codes for individual faces and the FFA is tuned for behaviorally relevant facial features. An electrocorticography
study found that the FFA is involved in multiple stages of face
processing, continuously from when people see a face until they respond
to it, demonstrating the dynamic and important role the FFA plays as
part of the face perception network.
Another study found that there is stronger activity in the FFA
when a person sees a familiar face as opposed to an unfamiliar one.
Participants were shown different pictures of faces that either had the
same identity, familiar, or faces with separate identities, or
unfamiliar. It found that participants were more accurate at matching
familiar faces than unfamiliar ones. Using an fMRI, they also found that
the participants that were more accurate in identifying familiar faces
had more activity in their right fusiform face area and participants
that were poor at matching had less activity in their right fusiform
area.
History
Function and controversy
The fusiform face area (FFA) is a part of the brain located in the fusiform gyrus with a debated purpose. Some researchers believe that the FFA is evolutionary purposed for face perception. Others believe that the FFA discriminates between any familiar stimuli.
Psychologists debate whether the FFA is activated by faces for an evolutionary or expertise
reason. The conflicting hypotheses stem from the ambiguity in FFA
activation, as the FFA is activated by both familiar objects and faces. A
study regarding novel objects called greebles determined this phenomenon.
When first exposed to greebles, a person's FFA was activated more
strongly by faces than by greebles. After familiarising themselves with
individual greebles or becoming a greeble expert, a person's FFA was
activated equally by faces and greebles. Likewise, children with autism have been shown to develop object recognition at a similarly impaired pace as face recognition. Studies of late patients of autism have discovered that autistic people have lower neuron densities in the FFA
This raises an interesting question, however: Is the poor face
perception due to a reduced number of cells or is there a reduced number
of cells because autistic people seldom perceive faces? Asked simply: Are faces simply objects with which every person has expertise?
Chinese characters similar to those used in Fu et al., which elicit a response in the FFA
There is evidence supporting the FFA's evolutionary face-perception.
Case studies into other dedicated areas of the brain may suggest that
the FFA is intrinsically designed to recognize faces. Other studies have
recognized areas of the brain essential to recognizing environments and
bodies. Without these dedicated areas, people are incapable of recognizing places and bodies. Similar research regarding prosopagnosia has determined that the FFA is essential to the recognition of unique faces.
However, these patients are capable of recognizing the same people
normally by other means, such as voice. Studies involving language
characters have also been conducted in order to ascertain the role of
the FFA in face recognition. These studies have found that objects, such
as Chinese characters, elicit a high response in different areas of the FFA than those areas that elicit a high response from faces. This data implies that certain areas of the FFA have evolutionary face-perception purposes.
Evidence from infants
The
FFA is underdeveloped in children and does not fully develop until
adolescence. This calls into question the evolutionary purpose of the
FFA, as children show the ability to differentiate faces. Two-year-old
babies have been shown to prefer the face of their mother.
Although the FFA is underdeveloped in two-year-old babies, they have
the ability to recognize their mother. Babies as early as three months
old have shown the ability to distinguish between faces. During this time, babies exhibit the ability to differentiate between genders, showing a clear preference for female faces.
It is theorized that, in terms of evolution, babies focus on women for
food, although the preference could simply reflect a bias for the
caregivers they experience. Infants do not appear to use this area for
the perception of faces. Recent fMRI work has found no face selective
area in the brain of infants 4 to 6 months old.
However, given that the adult human brain has been studied far more
extensively than the infant brain, and that infants are still undergoing
major neurodevelopmental processes, it may simply be that the FFA is
not located in anatomically familiar area. It may also be that
activation for many different percepts and cognitive tasks in infants is
diffuse in terms of neural circuitry, as infants are still undergoing
periods of neurogenesis and neural pruning;
this may make it more difficult to distinguish the signal, or what we
would imagine as visual and complex familiar objects (like faces), from
the noise, including static firing rates of neurons, and activity that
is dedicated to a different task entirely than the activity of face
processing. Infant vision involves only light and dark recognition,
recognizing only major features of the face, activating the amygdala. These findings question the evolutionary purpose of the FFA.
Evidence from emotions
Studies into what else may trigger the FFA validates arguments about its evolutionary purpose. There are countless facial expressions
humans use that disturb the structure of the face. These disruptions
and emotions are first processed in the amygdala and later transmitted
to the FFA for facial recognition. This data is then used by the FFA to
determine more static information about the face.
The fact that the FFA is so far downstream in the processing of emotion
suggests that it has little to do with emotion perception and instead
deals in face perception.
Recent evidence, however, shows that the FFA has other functions
regarding emotion. The FFA is differentially activated by faces
exhibiting different emotions. A study has determined that the FFA is
activated more strongly by fearful faces than neutral faces.
This implies that the FFA has functions in processing emotion despite
its downstream processing and questions its evolutionary purpose to
identify faces.
Functional specialization suggests that different areas in the brain are specialized for different functions.
Historical origins
1848 edition of American Phrenological Journal published by Fowlers & Wells, New York City
Phrenology, created by Franz Joseph Gall
(1758–1828) and Johann Gaspar Spurzheim (1776–1832) and best known for
the idea that one's personality could be determined by the variation of
bumps on their skull, proposed that different regions in one's brain
have different functions and may very well be associated with different
behaviours.
Gall and Spurzheim were the first to observe the crossing of pyramidal
tracts, thus explaining why lesions in one hemisphere are manifested in
the opposite side of the body. However, Gall and Spurzheim did not
attempt to justify phrenology on anatomical grounds. It has been argued
that phrenology was fundamentally a science of race. Gall considered the
most compelling argument in favor of phrenology the differences in
skull shape found in sub-Saharan Africans and the anecdotal evidence
(due to scientific travelers and colonists) of their intellectual
inferiority and emotional volatility. In Italy, Luigi Rolando carried out lesion experiments and performed electrical stimulation of the brain, including the Rolandic area.
Phineas Gage's accident
Phineas Gage became one of the first lesion case studies in 1848 when an explosion drove a large iron rod completely through his head, destroying his left frontal lobe.
He recovered with no apparent sensory, motor, or gross cognitive
deficits, but with behaviour so altered that friends described him as
"no longer being Gage," suggesting that the damaged areas are involved
in "higher functions" such as personality. However, Gage's mental changes are usually grossly exaggerated in modern presentations.
Subsequent cases (such as Broca's patient Tan) gave further support to the doctrine of specialization.
Major theories of the brain
Currently,
there are two major theories of the brain's cognitive function. The
first is the theory of modularity. Stemming from phrenology, this theory
supports functional specialization, suggesting the brain has different
modules that are domain specific in function. The second theory,
distributive processing, proposes that the brain is more interactive and
its regions are functionally interconnected rather than specialized.
Each orientation plays a role within certain aims and tend to
complement each other (see below section `Collaboration´).
Modularity
The
theory of modularity suggests that there are functionally specialized
regions in the brain that are domain specific for different cognitive
processes. Jerry Fodor
expanded the initial notion of phrenology by creating his Modularity of
the Mind theory. The Modularity of the Mind theory indicates that
distinct neurological regions called modules
are defined by their functional roles in cognition. He also rooted many
of his concepts on modularity back to philosophers like Descartes, who
wrote about the mind being composed of "organs" or "psychological
faculties". An example of Fodor's concept of modules is seen in
cognitive processes such as vision, which have many separate mechanisms
for colour, shape and spatial perception.
One of the fundamental beliefs of domain specificity and the theory of modularity suggests that it is a consequence of natural selection
and is a feature of our cognitive architecture. Researchers Hirschfeld
and Gelman propose that because the human mind has evolved by natural
selection, it implies that enhanced functionality would develop if it
produced an increase in "fit" behaviour. Research on this evolutionary
perspective suggests that domain specificity is involved in the
development of cognition because it allows one to pinpoint adaptive
problems.
An issue for the modular theory of cognitive neuroscience is that
there are cortical anatomical differences from person to person.
Although many studies of modularity are undertaken from very specific
lesion case studies, the idea is to create a neurological function map
that applies to people in general. To extrapolate from lesion studies
and other case studies this requires adherence to the universality assumption,
that there is no difference, in a qualitative sense, between subjects
who are intact neurologically. For example, two subjects would
fundamentally be the same neurologically before their lesions, and after
have distinctly different cognitive deficits. Subject 1 with a lesion
in the "A" region of the brain may show impaired functioning in
cognitive ability "X" but not "Y", while subject 2 with a lesion in area
"B" demonstrates reduced "Y" ability but "X" is unaffected; results
like these allow inferences to be made about brain specialization and
localization, also known as using a double dissociation.
The difficulty with this theory is that in typical non-lesioned
subjects, locations within the brain anatomy are similar but not
completely identical. There is a strong defense for this inherent
deficit in our ability to generalize when using functional localizing
techniques (fMRI, PET etc.). To account for this problem, the
coordinate-based Talairach and Tournoux stereotaxic system
is widely used to compare subjects' results to a standard brain using
an algorithm. Another solution using coordinates involves comparing
brains using sulcal reference points. A slightly newer technique is to use functional landmarks,
which combines sulcal and gyral landmarks (the groves and folds of the
cortex) and then finding an area well known for its modularity such as
the fusiform face area. This landmark area then serves to orient the researcher to the neighboring cortex.
Future developments for modular theories of neuropsychology may
lie in "modular psychiatry". The concept is that a modular understanding
of the brain and advanced neuro-imaging techniques will allow for a
more empirical diagnosis of mental and emotional disorders. There has
been some work done towards this extension of the modularity theory with
regards to the physical neurological differences in subjects with
depression and schizophrenia, for example. Zielasek and Gaeble have set
out a list of requirements in the field of neuropsychology in order to
move towards neuropsychiatry:
To assemble a complete overview of putative modules of the human mind
To establish module-specific diagnostic tests (specificity, sensitivity, reliability)
To assess how far individual modules, sets of modules or their connections are affected in certain psychopathological situations
To probe novel module-specific therapies like the facial affect
recognition training or to retrain access to context information in the
case of delusions and hallucinations, in which "hyper-modularity" may
play a role [8]
Research in the study of brain function can also be applied to cognitive behaviour therapy.
As therapy becomes increasingly refined, it is important to
differentiate cognitive processes in order to discover their relevance
towards different patient treatments. An example comes specifically from
studies on lateral specialization between the left and right cerebral
hemispheres of the brain. The functional specialization of these
hemispheres are offering insight on different forms of cognitive
behaviour therapy methods, one focusing on verbal cognition (the main
function of the left hemisphere) and the other emphasizing imagery or
spatial cognition (the main function of the right hemisphere). Examples of therapies that involve imagery, requiring right hemisphere activity in the brain, include systematic desensitization and anxiety management training.
Both of these therapy techniques rely on the patient's ability to use
visual imagery to cope with or replace patients symptoms, such as
anxiety. Examples of cognitive behaviour therapies that involve verbal
cognition, requiring left hemisphere activity in the brain, include
self-instructional training and stress inoculation.
Both of these therapy techniques focus on patients' internal
self-statements, requiring them to use vocal cognition. When deciding
which cognitive therapy to employ, it is important to consider the
primary cognitive style of the patient. Many individuals have a tendency
to prefer visual imagery over verbalization and vice versa. One way of
figuring out which hemisphere a patient favours is by observing their
lateral eye movements. Studies suggest that eye gaze reflects the
activation of cerebral hemisphere contralateral to the direction. Thus,
when asking questions that require spatial thinking, individuals tend to
move their eyes to the left, whereas when asked questions that require
verbal thinking, individuals usually move their eyes to the right.
In conclusion, this information allows one to choose the optimal
cognitive behaviour therapeutic technique, thereby enhancing the
treatment of many patients.
Areas representing modularity in the brain
Fusiform face area
One of the most well known examples of functional specialization is the fusiform face area (FFA). Justine Sergent
was one of the first researchers that brought forth evidence towards
the functional neuroanatomy of face processing. Using positron emission
tomography (PET), Sergent found that there were different patterns of
activation in response to the two different required tasks, face
processing verses object processing.
These results can be linked with her studies of brain-damaged patients
with lesions in the occipital and temporal lobes. Patients revealed that
there was an impairment of face processing but no difficulty
recognizing everyday objects, a disorder also known as prosopagnosia. Later research by Nancy Kanwisher using functional magnetic resonance imaging (fMRI), found specifically that the region of the inferior temporal cortex, known as the fusiform gyrus,
was significantly more active when subjects viewed, recognized and
categorized faces in comparison to other regions of the brain. Lesion
studies also supported this finding where patients were able to
recognize objects but unable to recognize faces. This provided evidence
towards domain specificity in the visual system, as Kanwisher acknowledges the Fusiform Face Area as a module in the brain, specifically the extrastriate cortex, that is specialized for face perception.
Visual area V4 and V5
While looking at the regional cerebral blood flow (rCBF), using PET, researcher Semir Zeki directly demonstrated functional specialization within the visual cortex known as visual modularity. He localized regions involved specifically in the perception of colour and vision motion. For colour, visual area V4 was located when subjects were shown two identical displays, one being multicoloured and the other shades of grey. This was further supported from lesion studies where individuals were unable to see colours after damage, a disorder known as achromatopsia. Combining PET and magnetic resonance imaging
(MRI), subjects viewing a moving checker board pattern verses a
stationary checker board pattern located visual area V5, which is now
considered to be specialized for vision motion. (Watson et al., 1993)
This area of functional specialization was also supported by lesion
study patients who's damage caused cerebral motion blindness.
Frontal lobes
Studies have found the frontal lobes to be involved in the executive functions of the brain, which are higher level cognitive processes.
This control process is involved in the coordination, planning and
organizing of actions towards an individual's goals. It contributes to
such things as one's behaviour, language and reasoning. More
specifically, it was found to be the function of the prefrontal cortex,
and evidence suggest that these executive functions control processes
such as planning and decision making, error correction and assisting
overcoming habitual responses. Miller and Cummings used PET and
functional magnetic imaging (fMRI) to further support functional
specialization of the frontal cortex. They found lateralization of
verbal working memory in the left frontal cortex and visuospatial
working memory in the right frontal cortex. Lesion studies support these
findings where left frontal lobe patients exhibited problems in
controlling executive functions such as creating strategies. The dorsolateral, ventrolateral and anterior cingulate
regions within the prefrontal cortex are proposed to work together in
different cognitive tasks, which is related to interaction theories.
However, there has also been evidence suggesting strong individual
specializations within this network.
For instance, Miller and Cummings found that the dorsolateral
prefrontal cortex is specifically involved in the manipulation and
monitoring of sensorimotor information within working memory.
Right and left hemispheres
During the 1960s, Roger Sperry
conducted a natural experiment on epileptic patients who had previously
had their corpora callosa cut. The corpus callosum is the area of the
brain dedicated to linking both the right and left hemisphere together.
Sperry et al.'s experiment was based on flashing images in the right and
left visual fields of his participants. Because the participant's
corpus callosum was cut, the information processed by each visual field
could not be transmitted to the other hemisphere. In one experiment,
Sperry flashed images in the right visual field (RVF), which would
subsequently be transmitted to the left hemisphere (LH) of the brain.
When asked to repeat what they had previously seen, participants were
fully capable of remembering the image flashed. However, when the
participants were then asked to draw what they had seen, they were
unable to. When Sperry et al. flashed images in the left visual field
(LVF), the information processed would be sent to the right hemisphere
(RH) of the brain. When asked to repeat what they had previously seen,
participants were unable to recall the image flashed, but were very
successful in drawing the image. Therefore, Sperry concluded that the
left hemisphere of the brain was dedicated to language as the
participants could clearly speak the image flashed. On the other hand,
Sperry concluded that the right hemisphere of the brain was involved in
more creative activities such as drawing.
Parahippocampal place area
Located in the parahippocampal gyrus,
the parahippocampal place area (PPA) was coined by Nancy Kanwisher and
Russell Epstein after an fMRI study showed that the PPA responds
optimally to scenes presented containing a spatial layout, minimally to
single objects and not at all to faces.
It was also noted in this experiment that activity remains the same in
the PPA when viewing a scene with an empty room or a room filled with
meaningful objects. Kanwisher and Epstein proposed "that the PPA
represents places by encoding the geometry of the local environment".
In addition, Soojin Park and Marvin Chun posited that activation in the
PPA is viewpoint specific, and so responds to changes in the angle of
the scene. In contrast, another special mapping area, the retrosplenial cortex (RSC), is viewpoint invariant or does not change response levels when views change. This perhaps indicates a complementary arrangement of functionally and anatomically separate visual processing brain areas.
Extrastriate body area
Located
in the lateral occipitotemporal cortex, fMRI studies have shown the
extrastriate body area (EBA) to have selective responding when subjects
see human bodies or body parts, implying that it has functional
specialization. The EBA does not optimally respond to objects or parts
of objects but to human bodies and body parts, a hand for example. In
fMRI experiments conducted by Downing et al. participants were asked to
look at a series of pictures. These stimuli includes objects, parts of
objects (for example just the head of a hammer), figures of the human
body in all sorts of positions and types of detail (including line
drawings or stick men), and body parts (hands or feet) without any body
attached. There was significantly more blood flow (and thus activation)
to human bodies, no matter how detailed, and body parts than to objects
or object parts.
Distributive processing
The
cognitive theory of distributed processing suggests that brain areas
are highly interconnected and process information in a distributed
manner.
A remarkable precedent of this orientation is the research of Justo Gonzalo on brain dynamics
where several phenomena that he observed could not be explained by the
traditional theory of localizations. From the gradation he observed
between different syndromes in patients with different cortical lesions,
this author proposed in 1952 a functional gradients model,
which permits an ordering and an interpretation of multiple phenomena
and syndromes. The functional gradients are continuous functions through
the cortex describing a distributed specificity, so that, for a given
sensory system, the specific gradient, of contralateral character, is
maximum in the corresponding projection area and decreases in gradation
towards more "central" zone and beyond so that the final decline
reaches other primary areas.
As a consequence of the crossing and overlapping of the specific
gradients, in the central zone where the overlap is greater, there would
be an action of mutual integration, rather nonspecific (or multisensory) with bilateral character due to the corpus callosum.
This action would be maximum in the central zone and minimal towards
the projection areas. As the author stated (p. 20 of the English
translation)
"a functional continuity with regional variation is then offered, each
point of the cortex acquiring different properties but with certain
unity with the rest of the cortex. It is a dynamic conception of quantitative localizations".
A very similar gradients scheme was proposed by Elkhonon Goldberg in 1989.
Other researchers who provide evidence to support the theory of distributive processing include Anthony McIntosh and William Uttal,
who question and debate localization and modality specialization within
the brain.
McIntosh's research suggests that human cognition involves interactions
between the brain regions responsible for processes sensory information,
such as vision, audition, and other mediating areas like the prefrontal
cortex. McIntosh explains that modularity is mainly observed in sensory
and motor systems, however, beyond these very receptors, modularity
becomes "fuzzier" and you see the cross connections between systems
increase.
He also illustrates that there is an overlapping of functional
characteristics between the sensory and motor systems, where these
regions are close to one another. These different neural interactions
influence each other, where activity changes in one area influence other
connected areas. With this, McIntosh suggest that if you only focus on
activity in one area, you may miss the changes in other integrative
areas. Neural interactions can be measured using analysis of covariance in neuroimaging.
McIntosh used this analysis to convey a clear example of the
interaction theory of distributive processing. In this study, subjects
learned that an auditory stimulus signalled a visual event. McIntosh
found activation (an increase blood flow), in an area of the occipital cortex, a region of the brain involved in visual processing,
when the auditory stimulus was presented alone. Correlations between
the occipital cortex and different areas of the brain such as the prefrontal cortex, premotor cortex and superiortemporal cortex showed a pattern of co-variation and functional connectivity.
Uttal focusses on the limits of localizing cognitive processes in
the brain. One of his main arguments is that since the late 90's,
research in cognitive neuroscience has forgotten about conventional psychophysical
studies based on behavioural observation. He believes that current
research focusses on the technological advances of brain imaging
techniques such as MRI and PET scans.
Thus, he further suggest that this research is dependent on the
assumptions of localization and hypothetical cognitive modules that use
such imaging techniques to pursuit these assumptions. Uttal's major
concern incorporates many controversies with the validly,
over-assumptions and strong inferences some of these images are trying
to illustrate. For instance, there is concern over the proper
utilization of control images in an experiment. Most of the cerebrum
is active during cognitive activity, therefore the amount of increased
activity in a region must be greater when compared to a controlled area.
In general, this may produce false or exaggerated findings and may
increase potential tendency to ignore regions of diminished activity
which may be crucial to the particular cognitive process being studied.
Moreover, Uttal believes that localization researchers tend to ignore
the complexity of the nervous system. Many regions in the brain are
physically interconnected in a nonlinear system, hence, Uttal believes
that behaviour is produced by a variety of system organizations.
Collaboration
The
two theories, modularity and distributive processing, can also be
combined. By operating simultaneously, these principles may interact
with each other in a collaborative effort to characterize the
functioning of the brain. Fodor himself, one of the major contributors
to the modularity theory, appears to have this sentiment. He noted that
modularity is a matter of degrees, and that the brain is modular to the
extent that it warrants studying it in regards to its functional
specialization.
Although there are areas in the brain that are more specialized for
cognitive processes than others, the nervous system also integrates and
connects the information produced in these regions. In fact, the
proposed distributive scheme of the functional cortical gradientes by J. Gonzalo
already tries to join both concepts modular and distributive: regional
heterogeneity should be a definitive acquisition (maximum specificity in
the projection paths and primary areas), but the rigid separation
between projection and association areas would be erased through the
continuous functions of gradient.
The collaboration between the two theories not only would provide
a more unified perception and understanding of the world but also make
available the ability to learn from it.