Neurophysiological Correlates of Autistic Social and Emotional Dysfunction


This review highlights some of the findings from brain imaging studies about the structure and function of brain regions and their interconnectivity in autism, with evidence for a range of anomalies both in specific sites and within neural systems in the brains of individuals with autism.  The significance of brain volume and impaired neurotransmission is emphasised.

There follows a summary of a further description of biological bases of autism, with particular reference to genetic influences, notably an impairment of genes clustering on the X chromosome.


M.J.Connor                                                                                                    June 2005



Brain Imaging and Emerging Data about Neural Systems


In her introduction to the review of current evidence about neural anomalies associated with autism, Herbert (2004) highlights the range of structures, from the brain stem to the associational cortex, which contribute to the neural systems underpinning social and emotional processing.

The research to identify the specific abnormalities in these systems has produced complex and varied findings.  The most replicated finding has been a tendency towards enhanced brain size among young children, but the pattern is not consistent or systematic; while evidence from metabolic and functional imaging has highlighted substantial a-typicality in activation within the brains of individuals with autism, but limited consistency in respect of the specific sites involved.

The concept of a spectrum disorder is underlined by the wide individual differences observed in the nature and severity of dysfunctions.


Herbert suggests that it the social and emotional processing capacities that are the essence of human nature and behaviour, and the neurobiological substrates of such processing are multiple and complex.  The processing depends upon networks ranging from evolutionarily old parts of the brain to more complex associational processing parts that are evolutionarily much younger and only observable in larger-brained primates.


In autism, social and emotional functions are impaired in a distinctive way, with sets of neuro-behavioural abnormalities forming three clusters …. language impairment, limited social reciprocity, and stereotyped or repetitive behaviour.

Each of the areas that make up this autistic “triad” may be significant in its own right, and will combine with the other areas in idiosyncratic permutations of strengths and qualities; but the very clustering suggests some linkage among underlying mechanisms and some descriptive models have been formulated as the means of linking various behavioural manifestations, such as weak central coherence, or impaired theory of mind. 


Various neurological, biochemical, immunological, or genetic abnormalities have been identified among individuals with autism, and autism can be linked to a number of disease entities.  Autism can also be idiopathic.

However, there are currently no reliable biological markers; and the underlying biology is noted for the heterogeneity.

One common belief has it that the variability among findings is a function of the differences in research methods, either in terms of participant selection or the kind of measures used.  This belief carries an implication that greater consistency among research methods and homogeneity among participants would lead to more rapid identification of underlying common links …. but it is equally plausible that some of the variability is a reflection of the true state of affairs whereby there are multiple pathways from some biological source to observable symptoms.


The component disorders, or their symptoms, show a further variability … viz, a difference in observed behaviours over time. 

These differences may be a matter of rapid switching of performance on a day to day basis, with “good” periods and “difficult” periods, or of change occurring over a longer time interval (and  permanent improvements may be associated with some therapeutic intervention).

One implication is that at least part of the underlying disabilities or abnormalities are not “hardwired” but are subject to change according to experience.


A further practical problem in seeking correlations between specific brain areas and behaviour is that, apart from language, the relevant specialist regions of the brain are not adequately defined …. and, even with regard to language performance, there are changes in the regional specialisation over time, and it is not clear which are the brain regions relevant to the language weaknesses, such as pragmatics, which are especially impaired in autism.


Herbert’s ongoing hypothesis is that the varied set of underlying biological abnormalities can underlie similar patterns of behavioural abnormalities because autism reflects some disruption to systems, with the negative impact arising from a variety of sources and in any number of ways (albeit raising the question of the circumstances under which system disruption can lead to a less severe set of symptoms or to a different syndrome altogether).


Social and emotional systems may be characterised as evolved modes of responding to perceived threats.  The most basic stages of operation of the autonomic nervous system involve modifications of cardiac functioning, mobilise the individual for

“ fight or flight ”, and heighten sensitivity to, and engagement with, the environment.  This last function involves several of the cranial nerves and is, thus, linked to regulation of elements of social engagement such as hearing/listening, facial expression, orientation via head movement, and coordination of breathing with vocalising.  Early mother-child interaction may also be seen as a function of vagal regulation in that biochemical changes in the brain (the inhibition or release of vasopressin or oxytocin under vagal control) are associated with the experience of security and danger as in social bonding or in isolation. 

Anomalies in this kind of biochemical neural system have been associated with the autistic spectrum.


A higher function concerns the evaluation of environmental stimuli and their emotional content or significance, including social encounters; and converging evidence implicates the amygdala in this level of social and emotional functioning, including the recognition of emotions conveyed by facial expressions.


(Incidentally, one notes the ongoing research at St. Andrew’s University reported in the June 2005 edition of Cognitive Psychology that, as a general rule, men are less competent than women at identifying the meaning of facial expressions.  It is hypothesised that the different sensitivities have evolved to match the different social needs of men and women.  Women are traditionally the care-givers and it is more salient for them to be very sensitive to the messages sent out via non-verbal channels.  Evidence also indicates that the density of grey matter in parts of the limbic system – including the amygdala – in typical female brains is greater than the density observed in the male brains; and it is the limbic system which is known to be critical for emotion processing. 

The orbital frontal lobe concerned with inhibition of aggression is also larger in female brains.  

This kind of finding has stimulated the view of autism as a kind of highly exaggerated form of male behavioural style, while also highlighting the likely significance of genetic factors in the aetiology of autism.)


Meanwhile, it is noted that participation in social interactions requires the capacity to recognise a given individual and to discriminate him/her from a group.  The relevant face recognition skill appears to be a function of part of the (ventral) cerebral cortex.  Faces are seen to be a special case in perceptual processing given that they are recognised as wholes rather than part by part.  However, individuals with autism do not process faces in the same way, although this deficit might be part of a wider visual processing anomaly.


Moving up the scale of social complexity ……it is noted that reciprocal social interaction requires an ability to make accurate inferences about the thoughts, and feelings, and perspectives of other people.  Imaging studies have identified a network of brain regions associated with this capacity, including the anterior paracingulate cortex. 

Further modulation of social and emotional experience involves other brain regions, notably the cerebellum which receives information from the associational cortices and is part of several neural circuits, and the cerebral hemispheres (notably the right hemisphere).

What also matters is the evaluation of new perceptual input and the drawing upon past experiences.  This appears to involve the operation of the associational cortex which has high interconnectivity with the limbic system, and subcortical relays such as the thalamic nuclei.


In sum, social and emotional processing does not involve isolated operations but requires the interaction of multiple and distributed components in various neural systems.  Disturbance of processing, of the kind reflected in anomalous behaviours characteristic of autism, can result from impairment of a specific part of the brain at some point in a network, or from altered or inhibited connectivity among areas which contribute to neural systems.


Herbert then discusses specific neural features and the possible significance for autism, beginning with the measurement of the volumes of brain structures held to be relevant because of the common view that volume of a given brain structure is proportional to processing capacity (but with a further possibility that total brain volume may be negatively correlated with inter-connectivity ??).


The cerebellum features commonly in existing studies although findings are inconsistent, with hypoplasia noted in some research but not replicated in others.

Nevertheless, Herbert cites one study, unique in studying the volume of cerebellar white matter, which found that this matter was significantly greater in autistic than in non-autistic pre-school children.


In the limbic system, the amygdala has been frequently but not consistently found to be smaller than among control participants.


Total brain volume has been found consistently to be enhanced among young children with autism.  (This is not universal, but other ongoing studies suggest that it is sufficiently frequent to be a kind of marker of risk, warranting ongoing monitoring.)

The brain volume increase appears to occur post-natally, during the first two years.  Among older children, smaller brain volume has been found to co-exist with larger head circumference suggesting the likelihood of a loss of brain volume over time.

The increased brain volume does not appear to have clear cognitive correlates, although one study cited by Herbert suggested a trend towards a higher non-verbal IQ.


The increase in volume seems to be largely concerned with a greater density of white matter with autistic children between 2 and 4 years of age showing a significant enlargement (around 18% more cerebral and 38% more cerebellar white matter), but adolescents between 12 and 16 years with autism tending to have less white matter than typically developing controls.  It is noted that the areas showing volume enlargement are those that complete the myelination process later, although the significance of this is not clear … or even whether there is any significance.


Where the proportions of components of neural systems are altered, there may be a change in the quality of the operation of that system.  For example, there may be a shift in the “expected” pattern of the volumetric relationship between frontal lobe and cerebellum, or hippocampus/amygdala and cerebral cortex, with possible implications for a shift in the pattern of processing. 

Meanwhile, the expanded white matter volume may coexist with unchanged corpus callosum volume, with similar implications for changes in the nature of interconnectivity and the processing of information from that which is typical.


Metabolic studies have revealed some temporal lobe abnormalities in autistic samples, with, for example, abnormal patterns of cerebral blood flow in the parietal lobe.

Delayed frontal lobe maturation has also been identified; and abnormal functioning in multiple regions has been reported, such as a reversal of the normal hemispheric dominance in the processing of auditory stimulation.

Currently, however, there is no hypothesis by which coherently to integrate these  findings which remain disparate.


Similarly, among functional imaging studies investigating brain activation linked to social and emotional processing, no clear common themes are emerging, and there are problems in determining whether an area of dysfunction represents some “primary” deficit or is a secondary effect of a deficit somewhere upstream in the circuit.

However, one specific finding from a mentalizing task (judging the emotions of others from their eyes)was the activation in normal participants of a network of regions including the orbitofrontal cortex, superior temporal gyrus, and amygdala.

On the other hand, participants with autism showed activation in those frontotemporal regions but not in the amygdala.

(The problem is that one cannot conclude that this identifies a specific deficit, and its locus, in inferring mental states since it might reflect a more general deficit in emotional processing).


Other studies have shown a common autistic trend towards deviations from the typical pattern of fusiform area activation when dealing with a face recognition task, but the precise patterns of activations of various brain regions were not consistent across participants in the various studies.  One linking hypothesis has it that autistic individuals do succeed in activating the fusiform area in face processing but show complex anomalies in the distributed network of brain areas that are also involved in social perception.


The above hypothesis might be seen as an example of a widespread problem in autism in terms of some impairment in holistic processing.  This has been described as weak central coherence or a deficit in the processing of complex information.

Local rather than holistic processing is certainly a common element within autistic style, and participants with autism can be differentiated from controls in their way of tackling certain tasks such as identifying embedded figures or block design, and in the pattern of neural activation associated with these spatial tasks ….. the autistic sample do not activate areas linked to the working memory component of the overall processing strategy used by controls.


The question is raised whether local rather than global processing is a consequence of abnormal or limited connections within neural networks.

Herbert cites studies of emotional judgement which have found that participants with autism can be differentiated from controls in terms of less activation in the fusiform area but greater activation in the temporal or lingual gyri, along with minimal activation in the amygdale.  

Such data are interpreted as consistent with a view of some dysfunction within neural circuits that has a particular impact upon higher-order cognitive processes.  A further hypothesis refers to some impairment of selectivity of relevant stimuli such that higher-order processing is overloaded with the outcome of some anomalous distribution of activation and a focus upon local processing with excessive attention to detail.


Herbert attempts to summarise the complexities within this whole area by referring to the probability that abnormalities of neural functioning may reflect both a specific and local impairment and a “distributed” abnormality to which various regions or systems have different degrees of vulnerability.  Once some impairment is present at one level, a major sequence of events can be initiated which will alter the functioning of a whole system.

The example is given of some behavioural or perceptual abnormality of face perception early in development which can inhibit attention to face or to signals of emotion at a critical later period, and the resulting inexperience in this domain leads to a kind of negative and self-amplifying loop and shifts the developmental course of the circuitry linking experience to expectation.


Meanwhile, the volume anomalies may be linked to connectivity changes with implications for changes in the way distributed neural systems function …. and the nature of change and impairment will depend upon the postnatal timing of the increase in brain volume.  However, it appears that the strongest impact is upon the frontal and prefrontal lobes which are significant for social and emotional processing.


There are various “network” hypotheses for autism, with, for example, autism seen as a matter of maladaptive system circuitry that can result from a number of impairments but where the cerebellar-limbic circuits are most implicated.


Alternatively, there has been proposed a model in which either too many or too few neuronal connections are made so that there is an emphasis upon details and local processing and a poor capacity for generalisation.  A further view still emphasises the lack of integration and cohesive inter-working of specialised local neural networks, leading to abnormal processing of a kind characterised by weak central coherence.


Various forms of disturbance might influence early brain development in such a way as to underpin the autistic characteristic of disrupted social and emotional behaviour.  One hypothesis refers to an increased excitation to inhibition ratio which could lead to widespread cortical “noise” which inhibits the normal differentiation of brain processing systems, alters memory systems, and interferes with the integration of these various systems.  The shift in the ratio could be the consequence of any of a range of genetic or environmental circumstances such as infections, immune disorders, or toxic factors.


Herbert concludes by stressing that she has only covered a small sample of the mechanisms that could be implicated in the aetiology of autistic social and emotional behaviours.  Further, there will be variability in the expression of faulty genes given the potentially multiple mutations and the varying degree of modulation by combinations of other genes.  Similarly, there will be environmental factors which, impacting early in life, can lead to abnormalities in the “hard-wiring” of circuits and their biochemical functioning. 

Nevertheless, neurotransmission anomalies and impaired connectivity among brain regions appear to be a common and key issue.


The Biology of Autism


The review by Joshi et al (2005) appears to dovetail well with the above notes on neural networks given its emphasis upon structural and functional changes in the brain of individuals with autism, neurotransmitter abnormalities, and metabolic or genetic or environmental factors.


With regard, firstly, to neurotransmission, the authors note the recent focus upon the role of serotonin in autism, and its role in this or other developmental disorders because of its significance in the process of forming new neurons in the brain.

Serotonin acts as a differentiating agent in the developing brain, and, later, as a neurotransmitter, with levels higher in childhood than in puberty/adulthood.  However, elevated levels of serotonin appear common among samples of individuals with autism, with the possible implication that autism may, in some cases, involve a defect in the gene that produces the serotonin transporter …. the protein that removes serotonin from the space between two nerve cells.


Meanwhile, and on the other hand, studies of the effects of drugs have indicated that, in other samples of individuals with autism, there are unusually low levels of serotonin in the brain.

There is also the common finding that the behaviour of some autistic children appears to be the more severe when their diet is deficient in tryptophan, a precursor of serotonin.


With respect to biochemistry, various research initiatives have indicated the possible significance of amino acids, endorphins, melatonin, etc. for brain functioning as well as for the identification of markers that may aid the early diagnosis of autism and ASD. 

A further hypothesis for the aetiology of at least some cases of autism concerns possible mitochondrial dysfunction as a result of excess production of nitric oxide. 


The authors also refer to the possibility of abnormalities in purine metabolism in a subgroup pf people with autism … identified by high levels of uric acid in their systems; and to the relative frequency with which children with autism are identified with lowered levels of magnesium (with evidence available that a lack of magnesium is associated with a predisposition to irritability or a kind of emotional withdrawal). It is also recognised that a high intake of the vitamin B6 can be associated with magnesium deficiency hence the desirability of monitoring levels in those people who are taking B6 vitamin supplements.


Yet a further source of possible anomaly in functioning is thought to emerge from “xenobiotic exposure ” (exposure to substances that have some deleterious impact upon the central nervous system).  The problems may result from ingestion of foods to which the child is allergic, such as wheat and dairy products, with the child unable to detoxify the substances concerned; and there has arisen the concept of diet-responsive autism.


Abnormal compounds resulting from the incomplete breakdown of proteins (such as gluten and casein) have been identified in more than a chance percentage of individuals with autism; and it has been hypothesised that the underlying problem is a leaky gut for which the best available intervention is an exclusionary diet.


With regard to genetic factors, there is a wealth of evidence from family and twin studies to indicate the significance of these hereditary influences in autism.  One practical problem in the way of greater understanding of the precise genetic mechanisms and routes is the way in which individuals with different autistic phenotypes have all been put together into autistic samples; or, given the lack of consistency among diagnostic criteria for autism, individuals who may not be directly comparable may all be part of a “unitary” sample representing autism.


However, evidence is accumulating that some forms of autism may reflect abnormalities in a single gene, while others may reflect abnormalities in two or more genes or some interaction between particular gene combinations and environmental (triggering) factors.

It is believed that between 10 and 15 genetic sites may all be implicated in autism and ASD, with the possibility that the mal-effects of the different genetic influences may be cumulative albeit not additive.


Joshi et al (op.cit) refer to the review of studies by Gillberg and Coleman (2000) which has highlighted the wide range of genes that have been identified as significant in the aetiology of autism, including the sex chromosomes.  It appears that chromosomes 2, 3, 7, 15, and X are the subject of particular interest among researchers seeking to identify the genes which are related to autism. 


For example, it has recently been observed  that mutations in a gene on the X chromosome are associated with Rett Syndrome (a condition identified as a variant of autism).  Most cases of Rett Syndrome arise sporadically, without any family history of the disorder, and there is speculation that further as yet unspecified abnormalities of this gene may underlie autism itself or other variants on the autistic spectrum.

However, it is thought likely that most cases of autism will reflect the impact of more than one abnormally-functioning gene.


With regard to environmental factors, there is ongoing research concerning the influence of such matters as lead and mercury poisoning, maternal alcohol intake, drug usage or smoking, pre- or peri-natal anoxia , and pre-natal infections.


The authors finally refer to immune disorders including the impairment of antibody production, protein deficiency leading to proneness to infection, and inflammation of the gut, accompanied by nutritional problems which, in interaction with the initial antigens, might result in alterations in behaviour and cognitive functioning.


Returning to the theme of genetic influence, it is relevant to refer to the work of Skuse (2003) who has been focusing upon a particular cluster of genes found upon the X chromosome of which women have two copies but men only one …. with the possibility of identifying specific reasons why men are significantly more vulnerable than women  to autism.


The work identified the genes which are essential to the development of the amygdala which is concerned with emotion recognition and processing and known to be anomalous in its functioning among people with autism. 

The research found similar anomalies in women who were missing a section of one X chromosome.  These individuals were also showing difficulties in recognising emotions even as gross as threat or fear in another person’s face, and found eye contact intimidating ….. difficulties which in a more extreme form are associated with autism.


Skuse has suggested that the cluster of genes on this particular section of the X chromosome appear crucial in the positive development of the amygdala; and the presence in females of the two X chromosomes appears to provide greater access to the active proteins which control amygdala growth.

Further support for the significance of X-linked genes comes from the study of females with Turner’s Syndrome which is marked by the presence of only one X chromosome.  Autism is observed to be many times more common among these females of whom around 5% will be diagnosed with autism and around 30% identified with some autistic traits.


                                 *          *          *          *          *          *


M.J.Connor                                                                                                    June 2005




Gillberg C. and Coleman M.  2000   The Biology of the Autistic Syndromes.  Cambridge University Press  


Herbert M.  2004   Neuro-imaging in disorders of social and emotional functioning.   Journal of Child Neurology  19(10)  772-784


Joshi I., Percy M., and Brown I.  2005   Advances in understanding causes of autism and effective intervention.   Journal on Developmental Disabilities  19(2)  1-27


Skuse D.  2003   The X-factor in autism.   Presentation to the conference of the British Association for the Advancement of Science.  Salford University : September 2003


This article is reproduced by kind permission of the author.

© Mike Connor 2005.

Back to NAS Surrey Branch Welcome Page