|  Autism
is a severe developmental disorder marked by significant impairments in
social, behavioral, and communicative functioning. Its early onset, symptom
profile, and chronicity strongly argue for a biological basis, and in
fact, several of lines of research implicate core biological mechanisms.
For example, autism is one of the most strongly genetic conditions, and
preliminary linkage data have already identified susceptibility regions
likely to contain genes involved in the condition. About a quarter of
individuals with autism exhibit a seizure disorder, and a larger number
of individuals have abnormal EEGs, which typically indicate bilateral
abnormalities without a consistent focus. However, the absence of consistent
biological markers presents across all cases and the pronounced heterogeneity
of the manifestations of autism have slowed research into its pathophysiology.
And
yet, major progress is being accomplished following the advent of new
tools of biobehavioral research. Novel neuroimaging techniques such as
magnetic resonance imaging (MRI) and positron emission tomography (PET)
are beginning to map out the neural systems affected by autism. These
include brain areas responsible for emotional and social functions, perceptual
systems specific to face and affect recognition, and social-cognitive
systems involved in understanding the intentions of others. Working models
of the pathophysiology of autism commonly include the amygdala at the
center of a distributed cortical-subcortical system, but other competing
models exist.
The
social, language, and behavioral problems that occur with autism suggest
that the syndrome affects a functionally diverse and widely distributed
set of neural systems. At the same time, however, the affected systems
must be discrete, because autism spares many perceptual and cognitive
systems. For example, autism is not incompatible with normative intelligence
or even superior visual-perceptual and other neuropsychological skills
and talents. Even though the full syndrome likely involves insults to
multiple systems, it remains possible that the initial insult is localized,
branching off into more pervasive impairments because of the highly interdependent
nature of early developmental processes.
Nearly
every neural system in the brain has been proposed at some point as the
cause of autism. Currently, the research data suggest that select aspects
of the temporal and frontal lobes, and portions of the amygdala, are key
nodes in systems affected by autism. Underlying the challenges to the
research effort is the fact that mental retardation is present in about
70% of individuals with autism, forcing researchers to disentangle causative
processes that are specific to autism from the nearly ubiquitous confound
of cognitive disability.
One
of the more intriguing findings to emerge in the past few years is that
overall brain size appears to be increased in autism (by about 5%–10%).
It is not yet understood whether all brain regions and systems are equally
affected by the expansion or whether this finding applies to all levels
of cognitive functioning. There is also some variability in the size of
the effect with age; some studies suggest that the enlargement is especially
pronounced in childhood, perhaps affecting white matter more than gray
matter. Whole brain enlargement could be merely a marker for a disturbance
in the fine structure of the brain that actually causes autistic symptoms.
Increased brain size might come at the expense of interconnectivity between
specialized neural systems, giving rise to a more fragmented processing
structure. In fact, some evidence suggests that the corpus callosum, the
major fiber pathway between the hemispheres, is reduced in size in autism.
Moreover, one PET study found a reduction in coordinated brain activity.
Less neural integration would be consistent with one influential theory
that attributes autistic symptoms to a lack of “central coherence,”
a cognitive processing style that makes integration of parts into wholes
problematic. One study measuring EEG gamma-band bursts associated with
the process of “perceptual binding” in individuals with autism
lends further support to this hypothesis.
Debate
continues as to whether the growth abnormality is postnatal or prenatal.
Specifying the developmental epoch with the most abnormal growth rate
would provide better clues to the underlying mechanism. An origin at particular
times of fetal brain development could suggest, for instance, disturbances
in the regulation of neuronal or glial cell proliferation, neuronal migration,
or apoptosis (programmed cell death). Prenatal origins of disturbed brain
development have been suggested by studies finding an increased frequency
of morphological abnormalities of the cerebral cortex in autistic individuals
(e.g., regional alterations in the size and number of gyri). Such abnormalities
stem from disturbances in neuronal migration during fetal brain development.
These gross neuroanatomical abnormalities are much more common among autistic
individuals with mental retardation in contrast with those with normative
IQs, and still they occur only in a minority of cases. Thus these findings
do not appear to be specific to the core social deficit in autism.
Postmortem
studies of a small number of persons with autism have revealed a range
of abnormalities, including a significant decrease in the number of Purkinje
cells and granule cells in the cerebellum. The precise nature of these
abnormalities, including a lack of gliosis indicative of scarring, suggests
a prenatal origin. A focus on the cerebellum would be consistent with
some neuroimaging evidence. A variety of posterior fossa abnormalities
seen on MRI have been reported in autism. These include abnormalities
of the pons, fourth ventricle, and cerebellar vermis, the midline portion
of the cerebellum. One influential set of findings ties autism to hypoplasia
of the neocerebellar vermis, but this abnormality has not consistently
been observed across studies. Moreover, it seems likely that posterior
fossa abnormalities are not specific to autism, but rather evident in
many persons with developmental disabilities and mental retardation. Thus
specificity of the findings for the core autistic features seems doubtful.
Of
the specific brain regions implicated in the pathobiology of autism spectrum
conditions, none has attracted as much interest as the limbic system,
especially the amygdala and its functional partners in the temporal and
frontal cortices. The limbic system lies largely within the medial and
ventral region of the temporal lobe, providing a girdle around the phylogenetically
older, deep brain structures. The amygdala, in particular, plays a critical
role in emotional arousal, assigning significance to environmental stimuli
and mediating the formation of visual-reward associations or “emotional”
learning. The amygdala has many afferent and efferent connections to the
temporal lobe, forming an important system for mediating the perception
of social stimuli.
Postmortem
examination of the brains of persons with autism finds consistent evidence
for abnormalities in size, density, and dendritic arborization of neurons
in the limbic system, including the amygdala, hippocampus, septum, anterior
cingulate, and mammillary bodies. There is a stunting of neuronal processes
and increased neuronal packing density, suggesting a curtailment of normal
development. These affected regions are strongly interconnected, and together
they comprise the majority of the limbic system. The limbic system, especially
the amygdala, is part of a neural structure that supports social and emotional
functioning. These postmortem findings, therefore, are often heralded
as the first good entrance points for understanding the pathobiology of
the autism spectrum disorders.
There
is supportive evidence for an amygdala theory of autism from animal models
of autism created through the lesioning of the amygdala of monkeys shortly
after their birth. Gradually in the course of the first year of life,
these animals develop patterns of behavior reminiscent of autism, i.e.,
social isolation, lack of eye contact, expressionless faces, and motor
stereotypies. Similar lesions in adulthood fail to produce autistic-like
sequelae. These findings are consistent with the notion that autistic
symptoms are in part a function of faulty early emotional learning mediated
by limbic system pathology. Moreover, as monkeys with early lesions to
the amygdala and surrounding entorhinal cortex mature into adulthood,
additional abnormalities are found in the neurochemistry of the frontal
cortex and in the subcortical regulation of dopaminergic activity. Thus
early discrete damage can produce widespread abnormalities across development.
Persons
with autism have deficits in their ability to recognize and discriminate
faces and to understand facial expressions. Functional neuroimaging and
lesion data show that the fusiform gyrus, a region on the underside of
the temporal lobe, is normally an area for face perception, while neighboring
regions in the posterior regions of the middle and superior temporal gyri
are important for reading facial expressions and social intent through
eye-gaze direction. Several functional MRI (fMRI) studies have now shown
hypoactivation of the fusiform gyrus during face perception tasks (Fig.
1). In a short period of time, this has now become the best replicated
finding in the neuroimaging literature. Preliminary evidence also links
hypoactivation of the amygdala and lateral temporal cortices to autism.
One hypothesis is that the principal pathology in autism resides in limbic
regions and that disturbance in social-affective orientation early in
life causes a cascade of neurodevelopmental events, including failure
to develop perceptual competence for faces and for visual and auditory
displays of emotion. The corollary of this hypothesis is that deficits
in subsequent and more complex social-cognitive skills may result from
this early derailment of the socialization processes.
Aspects
of frontal lobe integrity and function have been implicated in the pathogenesis
of autism. Older studies using lower-resolution neuroimaging techniques
reported general hypoactivation of the frontal lobes. Functional neuroimaging
data collected in the past decade are converging to show that subregions
of the prefrontal cortices with especially strong connectivity to limbic
areas are critical for “social cognition,” i.e., thinking
about others’ thoughts, feelings, and intentions. Deficits in such
“theory of mind” abilities are common in autism. Theory of
mind ability has been linked to functional activity in the medial region
of the superior frontal gyrus (primarily Brodmann area 9) and to the prefrontal
cortex immediately above the orbits of the eyes (i.e., orbital frontal
cortex). Bilateral lesions to the orbital and medial prefrontal cortices
cause deficits on theory of mind tasks. Preliminary functional imaging
evidence in autism spectrum conditions suggests altered functional representation
in the prefrontal cortex regions during theory of mind tasks. Moreover,
medial prefrontal dopaminergic activity as measured by fluorine-18-labeled
fluorodopa PET has been found to be significantly reduced in autism. Reduced
glucose metabolism during memory activities has also been reported in
a subdivision of the anterior cingulate gyrus, a region that lies along
the medial surface of the frontal lobe. Moreover, nonhuman primate studies
have documented abnormal social responsivity and loss of position within
the social group following lesions to the orbital and medial prefrontal
cortices.
The
orbital and medial prefrontal cortices have dense reciprocal connections
with the amygdala, providing the architecture for a system that can regulate
social-cognitive processes. A parallel set of amygdala-cortical circuitry
in the temporal lobes focuses on social-perceptual processes. One hypothesis
is that autism is largely caused by abnormalities in both of these amygdala-cortical
loops.
The
major findings to date on the neural basis of autism involve abnormalities
in brain size, aspects of the limbic system, functionally related and
connected regions of the orbitomedial prefrontal cortex, and visual association
areas of the temporal lobe. While good progress toward understanding the
neural basis of autism has been made in recent years, much work still
needs to be done. In this context, fMRI is revolutionizing psychiatry
and systems-level neuroscience, and it ultimately should enable researchers
to define dynamic brain processes that give rise to each specific symptom
and feature of autism. The closer synergy attained in the past few years
between more refined behavioral methods developed to isolate core aspects
of social processing and co-registered functional neuromapping further
increases the promise of this effort. A challenge for research in this
area will be to adapt in vivo neuroimaging techniques so that they are
applicable to the developing infant and toddler. Studying younger children
may be a prerequisite for a comprehensive understanding of the neural
basis of autism because the disorder evolves into its full form in a rather
short period of time in the first 2 years of life. Moreover, the initial
derailment of fundamental socialization processes is likely to unleash
a wide range of anomalous experiences that, in turn, are likely to result
in lifelong neurostructural and neurofunctional abnormalities.
top of page
|
