Recent exchanges of correspondence involving the Educational Psychology section of Children’s Services referred to that branch of research into autism concerned with brain system functioning, particularly the amygdala and other limbic system components.

These short notes set out to provide a brief re-introduction to this issue, highlighting the significance of such brain areas for the regulation of social awareness and reactivity, and of damage therein for the possible aetiology of characteristic social signs and symptoms of autism.


M.J. Connor                                                                                                                  December 2003



As an introduction to this set of notes about the possible neurological foundation of the triad of symptoms associated with autism, one might usefully quote Schultz and Klin (2002) in arguing that the early onset of the condition, the nature of the symptoms, and the time scales involved, would all point towards a biological basis.

They further note the heavy input of genetics into the aetiology of autism; and a biological basis is also inferred by the frequency with which individuals with autism experience seizures or display abnormal EEGs.

However, they also acknowledge the absence of clear biological markers across all cases, and the variability in the number or severity of symptoms, which have presented difficulties for research into underlying physical deficits or anomalies.


These authors then describe the rapid advances made possible by the development of new techniques such as magnetic resonance imaging (MRI) and positron emission tomography (PET) by which to make direct observations of the structure and functioning of neural systems.  One can now map those brain areas which are activated in social and emotional functioning, including those which appear specific to recognising faces and interpreting expressions into underlying feelings, or those involved in understanding the intentions of other people.

A common feature in many working models of the “pathophysiology” of autism is the amygdala operating in the middle of various cortical and subcortical systems.


It is argued by Schultz and Klin that the social, communicative, and behavioural manifestations of autism are such as to involve a diverse set of neural areas/systems; but the deficits must occur within some discrete system enabling other systems to operate normally because of the observation that a range of perceptual or cognitive functions continue unimpaired  in many cases of ASD.

It is their view, based upon a review of studies, that the key areas in systems affected by autism are parts of the temporal and frontal lobes, and amygdala.


They are intrigued by the finding that overall brain size is typically increased among individuals with autism (by as much as 10%) even if it is not yet known whether all brain regions and systems are equally affected. 

It is also noted that the enlargement is particularly pronounced during childhood, and the authors speculate whether the increased brain size may be associated with reduced interconnections between specialised neural areas producing the fragmented processing characteristic of weak “central coherence”.  Reference is made to some evidence for a reduction in the size of the corpus callosum – the main pathway between the two hemispheres – and for a reduction in coordinated brain activity, as measured by one PET study.


Meanwhile, postmortem studies are quoted as demonstrating a consistent pattern of fewer cells in the cerebellum among individuals with autism, although anomalies of this kind are not unique to autism but have also been observed among individuals with learning and developmental disabilities.


However, what does seem rather more autism-specific is the finding of deficits or anomalies in the limbic system, largely the temporal and frontal cortices and the amygdala, with the latter known to play a major role in emotional arousal and in attributing significance to environmental events.

Postmortem studies have shown that, among individuals who had autism, there are consistent abnormalities in the size or density or branching of neurons in the limbic system, including the amygdala, hippocampus, septum, mammillary bodies and anterior cingulate, with an increased density of neurons suggestive of a limitation in normal development.


Animal studies have also implicated the amygdala in the aetiology of autistic-like behaviours in that lesioning of the amygdala among monkeys in the very early days of life is associated with the mal-development of social and communicative functions, illustrated by social isolation, lack of eye contact, stereotyped motor behaviours, and expressionless faces.

Similar lesioning in adulthood does not bring about these effects, with the implication that autistic symptoms and behaviours are the outcome of inadequate or absent early emotional learning as a result of defects in the limbic system.  The authors suggest that the hypoactivation of the amygdala and other areas, such as the fusiform gyrus, is linked to disturbance in social and emotional orientation in early infancy, and the outcome involves a range of neurodevelopmental failures, including the non-development of the ability to discriminate faces or to recognise the expression of emotions.


They back this line of thinking by a reference to neuro-imaging data collected over the last few years which have consistently demonstrated that areas of the prefrontal cortices with particularly strong connectivity to the limbic areas are highly significant for “social cognition” (the ability to understand the points of view and the feelings of other people) and theory of mind weakness reflecting problems in social cognition is a core characteristic of autism.

A parallel amygdala-cortex circuit involving the temporal lobes is observed to underlie social-perceptual processes; and it is hypothesised that autism may be explained to a large extent by abnormalities in both of these amygdala-cortical loops.


The challenge set out by Schultz and Klin involves the application of the imaging techniques to babies and infants if one is to understand the neural basis of autism given that autism evolves over a short period in the first two or so years of life.


Meanwhile, on the subject of seizures, one notes the work of Tuunanen et al (1996) who artificially induced status epilepticus in rat subjects causing neurons to die in certain subregions of the amygdala …. those areas which, in humans, are associated with autistic behaviours and traits, such as eye-contact avoidance. 

In other words, there is the beginning of an hypothesis concerning both  causal mechanism (seizures) and neural substrate linked to autism …. a range of features associated with autism, such as anomalous sensory processing, or social deficits, may be linked via the theme of amygdala damage which, in turn, may be attributed in many cases to febrile seizures giving rise to status epilepticus.

The precise region in which the damage is located will be linked to the nature of observable symptoms given that certain groups of neurons are responsive to faces, others are linked to gaze monitoring or to eye contact, etc..


Tuunanen et al go on to speculate that examples of traits likely to become atypical as a result of the seizure - status epilepticus – amygdala damage sequence would include the processing of (negative) emotions such as fear, social behaviour and the ability to match behaviour to situation, stimulus-reward encoding, and certain aspects of memory.


A similar picture of limbic deficits or damage in autism is produced in the review by Edelson (2003) who refers particularly to the work of Bauman and Kemper (eg 1994). 

There is an emphasis upon the frequency with which damage in the limbic system, particularly in the hippocampus and amygdala, is identified among individuals diagnosed with autism; and there is further reference to both the smaller size of the neurons and to their greater density in these brain areas.


The review describes the amygdala as a central controlling mechanism of emotions including aggression, and notes how individuals with autism may be either aggressive towards themselves or others, or highly passive with behaviour appearing flat and emotionless. 

Again, the significance of the amygdala is highlighted via animal studies where it has been removed or damaged and the animals in question display behaviours similar to those observed in autism, such as social withdrawal, compulsive behaviours, failure to learn about dangerous situations, problems in memory retrieval, and difficulty in adjusting to change or novelty.


The role of the hippocampus is within learning and memory, and damage here will be associated with an inability to store new information.  This may be significant for studies of autism in that children with autism are commonly observed to have particular problems with relating new information to existing knowledge and experience.  Further, hippocampus damage in animals has been linked to stereotypic and self-stimulatory behaviours.


A note of caution is introduced by Edelson in terms of prefacing any speculation about limbic (dys)functioning with an acknowledgement that much of what is known about the working of the amygdala and other parts of the limbic system has emerged from animal studies which may not be validly extrapolated to human and autistic behaviours.   Nevertheless, he concludes that the correspondence between behaviours seen in autism and what is known about the limbic system appears significant.


The work of McCool (2002) begins with a reference to converging evidence about the role of the amygdala in humans, with divisions therein categorisable according to their developmental origin in the embryonic brain.

The amygdala is seen as an arbitrary collection of twenty different cell groups that may be divided into at least four separate functional units.  Working together, these units organise the way in which the brain integrates sensory and cognitive information to determine the emotional significance of an event or thought.


Based upon studies of individuals with amygdala damage, it has been observed that this brain area is active when the individual make social judgements such as the interpretation of facial expressions associated with negative emotions. 

The subjective interpretation by non-autistic individuals of negative expressions appeared not to require cognitive recognition of the face.


McCool argues that this combination of findings suggests that the function of the amygdala in these cognitive processes may be quite independent of attention or awareness.  He also refers to findings which suggest that males and females may use the left and right amygdala in different ways during memory tasks associated with the retrieval of emotionally-arousing stimuli.


His conclusion highlights the limitation of current understanding of individual differences in amygdala activity during emotional processing …. a conclusion reinforced by McCool’s reference to animal studies where lesioning of the amygdala in adult monkeys led to interference in the processing of novel or arousing stimuli and to a lack of inhibition in social interaction with unknown other monkeys, but where similar lesioning in infant monkeys did not lead to such effects. 

The lack of understanding, then, applies equally to the long term influence of social and physical development on amygdala functioning …. ie  there is an implication in these primate studies that the amygdala is not essential for the component processes of social behaviour, then there is a large question over its role in the impairment of social behaviour in autism.


Amara et al (2003) take up this question based upon amygdala-lesioning studies, and note that damage to the amygdala does have an effect upon a monkey’s response to normally fear inducing stimuli and removes a natural reluctance to engage with unfamiliar others.  This being so, they suggest that an important role for the amygdala is in the detection of threats and the evocation of an appropriate behavioural response, including the experience of fear …. and they note that a commonly comorbid feature of autism is anxiety. 

In other words, if the amygdala is implicated in the pathogenesis of autism or autistic symptoms, it may be a matter of contributions to abnormal fears and increased anxiety (and increased withdrawal? [MJC]) rather than directly to abnormal social behaviour.


A recently completed study organised by Yurgelen-Todd et al (Repligen Corporation 2002) through the collaboration of Harvard Medical School and Massachusetts General Hospital has assessed the impact of the use of injected secretin on amygdala activity in normally functioning adults, particularly with regard to possible effects upon social awareness including face recognition.

The choice of the amygdala as the focus for the study reflects the converging recognition that this brain area and other limbic components provide the centre of systems involved with affective responsiveness, and evidence is cited for the reduced activation of the amygdala in individuals with autism and their reduced capacity to respond to facial expressions, to perceive eye-gaze direction, and to establish recall memory for faces.


The outcome of the study showed that secretin is active in the CNS with a stimulant impact upon amygdala functioning, with explanatory implications for those studies which have claimed benefits from the use of secretin in individuals with autism. 


A second phase of the study involved young children (aged up to 7 years) as participants who were administered either secretin or a placebo.  The participants all showed moderate to severe symptoms of autism along with gastro-intestinal disorder.

Initial findings have shown that younger children, aged 3 and 4 years, showed benefits from the use of secretin in terms of reciprocal social interaction according to comparisons of baseline and follow-up clinical measures.


In summary, therefore, one can argue that evidence exists for the significance of the limbic system and the amygdala in particular when it comes to regulation of social interaction, and of damage to such brain areas when it comes to impaired social awareness and responsiveness.  However, more specific conclusions are inhibited by the individual differences observed, by the issue of the age of subjects and the point at which amygdala damage was experienced, and by the inappropriateness of direct extrapolation of data from animal studies to the study of human (mal)development.



                                           *          *          *          *          *


M.J.Connor                                                                                           December 2003







Amara D., Bauman M., and Schumann C.  2003   The amygdala and autism.  Genes, Brain, and Behaviour  2(5)  295-302


Bauman M. and Kemper T.  1994   The Neurobiology of Autism. 

Baltimore : John Hopkins University Press


Edelson S.  2003   Summary Notes - Autism and theLimbic System. 

Salem (Oregon) : Centre for the Study of Autism


McCool B.  2002   The Amygdala in brain function.  New York academy of Science. 

Neuroscience Conference presentation : March 24-27, 2002


Schultz and Klin  2002   Genetics of Childhood Disorders : XLIII. Autism.

Journal of the American Academy of Child and Adolescent Psychiatry  41 1259-1262


Repligen Corporation  2002   Clinical study to assess activation of amygdala by secretin in healthy adults.   Press release February 2002.


Tuunanen J., Halonen T., and Pitkanen A.  1996   Status epilepticus causes selective region damage and loss of GABAergic neurons in the rat amygdaloid complex.   European Journal of Neuroscience  8  2711-2725




This article is reproduced by kind permission of the author.

© Mike Connor 2003.

Back to NAS Surrey Branch Welcome Page