Skip Navigation
  Print Page

Pathophysiology of Autism: Brain Mechanisms

Skip sharing on social media links
Share this:

M. Denckla

Responses to NIH Questions

Question 1: What brain regions and functional pathways appear to be affected in autism? What key steps in development (timing, type, and loci) that are particularly sensitive to genetic or environmental insults are likely to be associated with autism?

Research studies in autism in the last 15 years using a wide range of technologies have provided evidence of a biological basis for autism. Information from neuropathology indicates that there may be abnormalities in the amygdala, hippocampus, septum, mamillary bodies, and the cerebellum. Autistic brains are slightly larger and heavier. (Clinical measures also indicate a larger than normal head circumference.) In the limbic system, there is an excess of cells and they are too small. The neurons themselves seem developmentally immature with a truncation in the development of their dendritic trees, which provide the basis for connections between neurons. Moreover, Purkinje cells are affected in a widespread fashion in the cerebellum. The anatomic differences found are consistent with a developmental curtailment that takes place at some point earlier than 30 weeks gestation (before birth). The neuropathologic findings are reasonably consistent and appear to dovetail with the lesion studies in primates. Exactly which findings are universal in autism and specific only to autism remain to be demonstrated.

Contemporary imaging research coupled with sophisticated neuropsychologic tools also offers exciting research possibilities for studying brain structure and function in vivo, particularly as new technology in both image acquisition and image analysis is developed. As with all research in autism, standardized diagnosis and control for age, gender, degree of mental retardation, language, and comorbid conditions are essential in interpreting these findings. The identification of reliably occurring subtypes and subgroups will be absolutely critical with all methodologies, as we can expect that a variety of brain structures and mechanisms may exist for subtypes with differing etiologies.

Across methodologies, studies reveal both higher and lower order areas that are dysfunctional. Neuropsychological studies have been uniform in finding deficits in certain aspects of higher order cognitive functions, including abstract and pragmatic language, encoding of complex information, and executive (frontal) functions. Other aspects of higher order cognitive functions, particularly those involved in verbal syntax and visuospatial organization are often spared in higher functioning individuals. Deficits in certain aspects of attentional functioning also are common and lower functioning persons with autism may also exhibit severe receptive and expressive language impairments, including mutism and a deficit in declarative memory. In contrast, rote memory often is intact. Evoked potential studies have also provided evidence of abnormalities in late information processing related to both frontal and parietal cortex. In contrast, evoked potential studies of early and midlatency potentials have demonstrated intact function in some subcortical areas. This neurophysiologic profile has been replicated with ocular motor, oculovestibular, and postural physiology methodology.

However, in terms of the timing, type, and locus of the originating abnormality in autism, the data from neuropathology suggest that other areas remote from the neocortex may be the beginning of the pathophysiological cascade. The universal impairment in social cognition found in neuropsychologic studies of autism suggests involvement of certain brain regions known to mediate social and emotional behavior, namely, regions of the limbic system, such as the amygdala and orbital frontal cortex. Animal research indicates that limbic lesions my cause secondary dysfunction in the neocortex. There is precedence in other diseases for this pathway, for example, progressive supernuclear palsy (PSP). Autopsy results in PSP show defects in the upper brain stem. However, PET scans in vivo show frontal area dysfunction. Frontal area functions are closer to the surface and have an amplified effect on scans. The upper brain stem may not properly activate the frontal area. What is most obvious in in vivo imaging may not necessarily reflect the basic defect.

Taken together, the available evidence in autism suggests that, although certain aspects of brain functioning are often spared in autism, the syndrome nevertheless involves widespread brain dysfunction at both the conical and subcortical levels. The originating site of the brain injury has not been identified. The competition of "top down" and "bottom up" hypotheses for the pathophysiological cascade in autistic development provides a fruitful area for future research.

At the subcellular level, neurochemistry research has provided consistent evidence of an elevation in a major neurotransmitter, serotonin, which affects potentiation at synapses and may play a role in the development of the nervous system. In terms of pathophysiology, it appears that there is a shared expression of a mutant gene in brain and platelet with respect to hyperserotonemia. Genetic analysis of the primary structure of the relevant neurochemicals is likely to be important for autism which has a sibling recurrence rate 4 to 10 times higher than that of insulin-dependent diabetes mellitus (IDDM) which has been found to have a genetic basis. Identified mutations could provide the first useful animal models of autism by homology, although animals will have a more limited behavioral repertoire.

Question 2: What behaviors observed in autism are consistent with the neuroanatomic findings?

Neuropsychological animal and human studies have demonstrated the key roles that some of the brain areas affected in autism may play, particularly in social/emotional development. Studies of the amygdala indicate its importance in recognition of the affective (emotional) significance of stimuli, in social stimulus-reward associations that allow understanding of the connections between behaviors and their consequences, in perception of body movements and eye gaze direction, in orienting toward social stimuli, and, together with the hippocampus, its role in long-term memory. Representation of action plans, motor planning and execution, and working memory are associated with the frontal lobe and the basal ganglia. There have been reports of late-onset symptoms in the frontostriatal system in monkeys who experienced early limbic system lesions. Rapid shifts in attention and modulation of sensory input have been associated with the cerebellum. Neurochemical strategies could be used to study specific behaviors in response to specific neurochemicals that are most likely to have an impact on the development of those regions thought to be involved in autism.

In terms of etiology, much debate has occurred regarding the identification of a single primary deficit at the cognitive level. Rather than focusing on the identification of a primary brain structure that is abnormal, it is important to recognize that multiple structures at multiple levels of the neuroaxis have clearly been implicated and all these structures participate in the neural systems that influence behavior. The pathophysiology of autism, or the structural and functional abnormalities of the brain and how precisely they result in the abnormal behavior of autism is far more complex than what brain structures or neurochemicals are involved. Each level of analysis is highly complex and, at present, only pieces of this puzzle in autism have been identified.

Question 3: What are the critical influences that the process of development brings to the design of experiments and the interpretation of findings?

Development clearly changes the outward expression of the signs and symptoms of autism. In addition, the changing signs and symptoms of autism must be compared to the changing backdrop of normal development in which the outward expression of normal abilities are also changing. In addition to variability associated with aging is the variability that occurs in normal humans in relation to general intellect and, in some cases, also gender and, in autism, in relation to severity of the disorder and developmental timing of onset, that is, congenital (at birth) or regression after apparently normal development. In assessing clinical functions, this means that different tests will be needed for different age- and ability-level individuals and that comparison groups must be matched on these relevant variables. With neurobiologic measures, these same variables of age, level of function, gender, and onset have a major impact and must be carefully considered in defining normative values and deviations from the norm. Several such methods including imaging and electrophysiologic cohesion measures have demonstrated that there are important and predictable changes in the relationships between measurements in different regions over the course of normal development. These factors require as careful attention to the selection of control subjects as to the rigor of diagnosis of autism.

Primate models also illustrate the importance of the role of development in the pathophysiologic cascade. Depending on the exact timing of the lesion, early injury to one part of the brain may result in later deficits in that part or in another part of the brain remote from the site of the original lesion. With certain known animal brain lesions, there is not much difficulty as an infant but there is significant social and working memory difficulty in adulthood. How profound the autistic-like behaviors are in monkeys depends on how early in the developmental process the brain lesions were made. Only through longitudinal animal studies can one find out what was primary and what was secondary. Longitudinal as opposed to cross-sectional studies could indicate whether subcortical findings are earlier and cortical findings are secondary to those deficits or vice versa.

Question 4: Does the available evidence suggest that there are prenatal/perinatal events associated with autism? If so, are they specific to autism, are they likely to be causal, and can they be used for clinical prognosis and the development of treatment strategies?

The available evidence suggests that there may be more problems in pregnancy or at birth, or more health problems immediately after birth in children with autism than in control families. Risk factors such as maternal age, prematurity, bleeding in pregnancy, toxemia, viral infection or exposure, and poor vigor in the neonatal period have been studied. However, there is little evidence that these problems are consistent across cases of autism or that they are specific to autism since they are also found in disorders such as dyslexia or developmental language disabilities. Such problems do not predict to later autism, nor do they appear to be related to asphyxia. These factors do not appear to cause autism, but may be reflections that fetal or neonatal development was compromised in some way.

Recommendations of the Working Group on Brain Mechanisms

  1. Investigation of brain structures in vivo with imaging methods is a major priority. At present, there are few data on most brain structures in autism. Cross-sectional, whole-brain studies at various ages are an essential first step in defining the relevant neuroanatomic focus for later studies. Functional MRI is a developing method that provides an opportunity for looking at the function of neural circuits without the hazards of radiation inherent to PET scans. Longitudinal studies in this area may be premature at this time until the rapidly changing technology stabilizes to allow for consistent measurements across time.
  2. The use of the technology of neuropsychology, both human and primate, can help sort out specific aspects of clinical functioning and refine knowledge of hypothesized relationships between cognitive deficits and behavioral difficulties. Methodological developments in this research area are also needed to define the testing paradigms necessary for nuclear magnetic resonance imaging of the functional variety.
  3. To expand knowledge of neuroanatomical findings, the need for access to a user-friendly brain bank was emphasized. Use of such a brain bank would lead to a greater number of appropriately age-, gender-, and cognitive-level matched controls being made available for study. Appropriate allocation of brain material to many different disciplines would allow the fuller use of postmortem brain samples for the study of specific anatomy and contribute to the urgently needed refinement of quantitative research methods for analysis. It would also permit staining of circuits that are associated with certain neurotransmitter pathways for use in genetically driven studies about the action of protein. 
  4. Studies of primary structure of relevant neurochemicals by genetic analysis are needed, since genetic study is mainly a tool to study neurochemicals in terms of determining which, when, and where proteins are expressed in the developing nervous system. For example, proteins involved in the development of neurons shown to be abnormal from postmortem studies can be examined by DNA analysis available from blood or saliva from well-characterized patients who may be followed prospectively.
  5. In an effort to identify key mechanisms in the pathogenesis of autism, studies of nerve growth and nerve growth migratory substances important for the modeling and remodeling of basic architectonics of certain centers of the human brain particularly important for language and social skills could be carried out. For example, family histories of affective disorder have been found in autism. In affective disorder, abnormalities have been identified in cell structure immediately adjacent to the inner surface of the cell membrane. This is also the site of action of neuronal growth factors, such as Growth Associated Protein, which guide the growth of developing neurons. This suggests an overlap or shared abnormal factor at the neurobiologic level in the regulation of brain membrane development in autism and affective disorder, particularly with regard to the inner membrane associated cytoskeleton. The association between autism and tuberous sclerosis may also be a particularly fruitful one in understanding the pathogenesis of both disorders. Research is also urgently needed that distinguishes two different developmental trajectories in autism, the one congenital (from birth) and the other characterized by apparently normal development followed by regression and onset of autism.
  6. Two important considerations for future research include the need for developmental norms for many new methodologies and consideration of norms in relation to IQ, gender, and race. Much of what is known about brain function and neuropathology is based on acquired brain damage in adults. If neurobiologic strategies are to be effective in correcting structural abnormalities of the brain, then noninvasive technology for the study of higher order cognitive abilities and their neural substrate should be employed over the course of development. The majority opinion was that newer functional magnetic resonance imaging will displace PET scans for activation studies, particularly once the enlarged windows of brain visualization are perfected.
    With regard to special considerations for such research, it is particularly important that normative data across the age span be accumulated with these new and more sophisticated methodologies for studying the brain such as volumetric MRI morphometry, functional imaging, and MR spectroscopy. It is also important to define normal in consideration of subject variables likely to have a major impact on neuronal organization including age, IQ (particularly Verbal IQ), gender, and race (especially in studies of infants and toddlers where the acquisition of milestones varies by race). It is also important that controls be chosen and matched as carefully as the autistic subjects and that they too be thoroughly assessed for evidence of current and past history of neurologic and psychiatric disorders as well as for family history status. Use of structured instruments for these purposes should be routine.
  7. Reports of abnormalities in higher order motor abilities (praxis) and higher cortical sensory abilities are now emerging. These findings may provide a basis for some of the unexplained aspects of the clinical syndrome of autism such as the sensory distortions (e.g., the relative insensitivity to pain and the sensory sensitivities) and movement disorders. Apraxis could provide a neurologic explanation for the inability of very young autistic children to use sign language. Sensory and motor abnormalities may be quite disabling and intervention depends on a better understanding of the neurologic basis of these behavioral difficulties. There is a related need for research on movement and synchrony, building on some previous research in this area and on new findings in Parkinson’s disease and autism.
  8. Replicable findings and consistency across methodologies will only occur when well-standardized methods are used for diagnosis, choice of comparison groups that control for relevant demographic and developmental variables, standardized protocols for imaging and psychological testing, and well-quantified methods of analysis. Such standardization is needed for all levels of inquiry neuro/pathophysiologic/anatomic, and etiologic (genetic and environmental), but progress at one level will not automatically result in solving questions at another.

first | previous | next | last

Last Updated Date: 08/15/2006
Last Reviewed Date: 08/15/2006
Vision National Institutes of Health Home BOND National Institues of Health Home Home Storz Lab: Section on Environmental Gene Regulation Home Machner Lab: Unit on Microbial Pathogenesis Home Division of Intramural Population Health Research Home Bonifacino Lab: Section on Intracellular Protein Trafficking Home Lilly Lab: Section on Gamete Development Home Lippincott-Schwartz Lab: Section on Organelle Biology