A Discussion about Autism and Possible Treatment Pathway

Diseases come in many different shapes and sizes inflicting multiple tiers of damage on the sufferer and his/her family and friends. Obviously those diseases that significantly and rapidly shorten an individual’s life span are the worst; however, some would argue that non-fatal neurological diseases like Alzheimer’s and autism are insidiously devastating, taking away the enjoyment from life in a way that cannot be alleviated by simple pain killers or anti-psychotics. Among detrimental neurological conditions it can be argued that autism is one of the worst because most neurological conditions manifest at advanced ages at least giving the individual a “normal” life for a considerable period of time whereas autism typically arises during the first few years of existence negatively impacting life from the beginning.

The symptoms of autism are passively detrimental and heterogeneous with various levels of mental delay, increased probabilities for epilepsy, various levels of attention deficit, language and learning impairments, obsessive compulsive disorders and difficulty interacting with others.1-3 Also like Alzheimer’s, the incident of autism and other similar disorders (autistic
spectrum disorders (ASDs)) among the general population continues to expand increasing from 4 per 10,000 in the 1960s to 30-60 (depending on the strict definition applied) per 10,000 in 2000s.4 While life expectancies of autistic individuals are only slightly reduced, the level of supplemental care that these individuals require demands significant resource expenditure.

A number of individuals believe that this expansion in autism is driven by new genuine cases. To support this position individuals cite that there have been no corresponding significant decreases in other similar diseases to compensate for the increase in autism.5 Basically individuals are not being diagnosed with condition x and then later having the diagnosis reevaluated as autism. However, while there are few substitutions, there is little evidence to contest the position of past under-diagnosis, which would consist of individuals with autism not being properly diagnosed due to the specific diagnostic condition(s) used at the time or due to lack of experience/knowledge of the individual performing the diagnosis. Now with better therapeutic and clinical diagnostic experience and ability, this undiagnosed segment of the population is now being properly diagnosed, thus there is actually no genuine increase. While this reasoning is logical, the rate of the increase seems too great to simply be accounted for by under-diagnosis alone. The rate of increase also seems too large to be explained by genetic mutations alone, thus it stands to reason that if this increase is legitimate then environmental causes are also influencing the change.

The cause(s) of autism, either genetically or environmentally, is currently unknown notwithstanding an incredible display of stupidity from certain individuals attempting to foolishly link vaccination and an increased probability of developing autism despite zero valid scientific evidence supporting such a position. Incidentally these people must also believe that little green men dance the samba on Mars everyday at 4:45 EST because there is an equal amount of empirical evidence supporting either position. Currently there are no specific biomarkers that have been identified to reliably diagnose autism,6-8 although a number of biomarkers have been identified to have differing concentrations between autistic and non-autistic individuals. However, these differences are marginalized because of their non-specificity towards autism versus other metabolic and neurological conditions.

Most modern study of autism has focused on searching for a genetic cause and some progress has been made for the belief is that a majority of autism cases result from probabilistic interactions between multiple common gene variants where each variant contributes some small influence on the overall condition.9,10 Another 10-20% of autism cases are believed to occur through known genetic effects including Synapsin 1, SynGAP1, SHANK3, NLGN4, CNTNAP2, retinoic acid-related orphan receptor-alpha (RORA), MET and cytochrome P450 genes, some which may have significant heritability rates.11-20

However, with all of the hundreds of genes involved in autism, focusing on genetic research and any potential genetically engineered treatment appears to be an inefficient strategy. Instead it seems reasonable that more attention should be paid, by both researchers and the media (especially the media), to the mechanisms associated with the development and maintenance of ASD for addressing the mechanism provides a higher probability of general treatment versus genetic strategies. Two common biological symptoms of autism are an increase in brain size and weight and an increased probability of suffering epileptic seizures. One mechanism to explain these symptoms is a deficiency in the synaptic elimination process during development (a.ka. synaptic pruning).

Soon after birth an overabundance of synapses are formed as neurons produce multiple synapses between a singe post-synaptic and pre-synaptic neuron. During development important synapses, those that fire more frequently and/or have greater firing duration, are strengthened and neighboring synapses are weakened and non-selectively eliminated until there is a one-to-one relationship between axons and cell body.21,22 The reason this process occurs is because unlike mature neuron-axon relationships where electrical impulses either facilitate an action potential or nothing, immature neuron-axon relationships, especially in Purkinje cells (PCs), operate on an intensity mechanism.23,24 This intensity mechanism exists because there is no initial molecular method for the brain to identify which synapses are important and which are not, thus importance is determined by firing duration/rate. Most notably this firing is driven from visual stimulus, especially before a recognized conscious understanding develops in the individual.

In abnormal pruning conditions the process does not create the typical one-one ratio instead it creates a neurological environment where numerous synapses from a single pre-synaptic neuron feed into single post-synaptic neurons. These additional synapses create additional brain volume offering an explanation for the autistic symptomology, where very young children with autism (ranging from 18 months to 4 years) have 5-10% more total brain volume.25-27 The additional weight can be explained by additional myelination because the myelination process is based on proximity, thus all axons in a given proximity will express the necessary oligodendritic targets to drive myelination. MRI studies have demonstrated growth abnormalities concerning gray and white matter supporting the belief that myelination accounts for weight variances.28,29

A lack of effective pruning could also explain the typical autistic symptom of early overgrowth followed by abnormal slowed growth.26,30 The excess synapses create the overgrowth, but due to conflicting excitatory and inhibitory signaling due to unnecessary action potentials, incomplete synchronization, incomplete action potentials and/or neuronal strengthening patterns, like long-term potentiation, autistic individuals have disadvantages hindering learning and other neuronal expansion. Disruption of these processes result in a developmental regression for autistic children of 20-40% between 15 to 24 months of age31 with higher probabilities corresponding to higher autism severity.32

An interesting and ironic side effect to a lack of pruning in the development of autism could be that over time if signal interaction is continuously disrupted by confliction between the almost competing synapses then all of the synapses can be negatively affected to the point where the neuron itself dies due to inactivity. This result may explain some of the outcomes in post-mortem autistic brains that are ripe with stunted neurons, especially PCs.33,34

Overall there are four major chronological phases in the process of creating a normal functional synapse: 1) functional differentiation among multiple creeping fiber inputs; 2) evolution of a superior synapse resulting in dendritic translocation; 3) early phase synapse elimination; 4) late phase synapse elimination.24 For the purpose of autism it appears that the third and fourth phases are the significant ones. Between the two elimination phases late phases appears to be more important due to its parallel fiber (PF)-PC synapse requirements and its longer period of operation in mice and assumed longer period of operation in humans (P9-P17 versus P7-P8 in mice).21

In general early elimination removes synapses that do not form properly and late elimination removes synapses that do not “win” the synapse augmentation competition. Therefore, there is a higher probability that a faulty late elimination process will create a more detrimental outcome than a faulty early elimination process because the late elimination process has a higher probability of eliminating functional synapses or not eliminating competing synapses. However, while in general short-term occurs prior to long-term elimination the exact timing of this elimination is not homogeneous throughout the brain (different areas go through the elimination process at different times during development).35 It is thought that pruning occurs first in primary sensory and motor areas followed by temporal and parietal areas and ends with the frontal cortex.35,36 Interestingly if this pruning order is accurate then sensory and motor abnormalities should be the first autistic indicator over emotional or social interaction.

A key molecule in late stage synaptic elimination is glutamate. The principle reason for its importance is that glutamate drives neuronal excitation probabilities and magnitudes in both non-mature neuronal networks and mature ones. Due to the importance of glutamate the late stage pruning process could find its roots late in stage one during functional differentiation when creeping fibers synthesize type-2 vesicular glutamate transporters during the formation of their initial synaptic butons.24 These transporters are important because they play a key role in determining the amount of glutamate that is released into the synaptic cleft after electrical stimulation of the given neuron. Synapses that express more vesicular glutamate transporters have an advantage in the synapse competition that determines which synapses are eliminated in stage four.

The strength of the particular synapse is largely dictated by the transient rise of glutamate within the cleft at any point in time. Clearly the availability of more synaptic vesicles will increase glutamate concentration and indicate a stronger connection, but that is only one element. Another important element is expression of metabotropic glutamate receptor 1 (mGluR1). As a metabotropic receptor mGluR1, when bound, activates an intracellular G-protein in order to initiate a signal cascade. mGluR1 is a C type G-protein receptor, over the more common A type, and their activation typically results in the activation of calcium channels and protein kinase C (PKC) resulting in a single calcium concentration spike.37 Additional less common downstream effectors include phospholipase D, casein kinase 1, cyclin dependent protein kinase 5, Jun kinase, mitogen-activated protein kinase/extracellular receptor kinase (MAPK/ERK) pathway and mammalian target of rapamycin (MTOR)/p70 S6 kinase pathway, which influence among other things synaptic plasticity.38-41

mGluRs have a ligand binding site on their N-terminal domain, which is formed by two hinged globular domains, commonly known as the Venus Flytrap Domain, due to its structure and action.42 These Venus domains also have allosteric binding sites for agonist and antagonist molecules. When glutamate binds these hinged domains close initiating the activation of the associated G-protein. There is some question to whether or not glutamate binding to a single site will lead to activation (no activation versus negative cooperativity),43 but agonist binding to a protomer seems to induce cis and trans-activation.44 This cis and trans-activation may provide an interesting association with autism development.

mGluR1 can be found both on post-synaptic and pre-synaptic membranes in the cerebellum, thalamus and the hippocampus.42 mGluR1 that is pre-synaptically located could drive feedback on the pre-synaptic cell resulting in a longer duration of glutamate release; thus in such a situation greater expression of pre-synaptic mGluR1 will increase the probability of strengthening the particular synapse. Other feedback activity can also be mediated by depolarization of the post-synaptic cell and the release of endocannabinoids.45-47 Post-synaptic receptor location is highly specific regionalized around the post-synaptic density, but not within it and induces post-synaptic cellular depolarization.42 This localization is thought to be achieved by interaction with Homer family regulatory proteins.48

Deficiencies in mGluR1 operation result in cerebellar gait problems, deficits in long-term depression and long-term potentiation and abnormal levels of regression of climbing fibers from cerebellar Purkinje cells49-50 basically reducing the probability that any fiber creates a strong one-one connection to an individual PC. In addition certain GluR1 mutations generate conditions that facilitate spontaneous ataxia through changes in the ligand-binding domain.51 Other important elements in pruning are GluRdelta2 and Cbln1, which bind to each other and are essential for the formation, organization and maintenance of PF-PC synapses52,53 and the proper execution of late phase synapse elimination as well as PLCbeta4 and PKCgamma which are part of the mGluR1 signaling cascade.24,54

Once the “winning” synapse has been identified, the creeping fibers that make up the synapse finish translocation to dendrites after the early stage elimination while the “losing” synapses remain around the soma. This proximity association creates two questions: 1) does the “winning” synapse produce some molecule that drives translocation towards the dendrites or do the “losing” synapses produce some molecule that inhibits translocation; 2) are somas “cluster pruned” or do each “losing” synapse produce a signal of sorts that targets it for elimination?

For the first question studies of the neuromuscular junction developed the “punishment model” of synapse competition where the strong synapse, created through the process discussed above, “punishes” other synapses by inducing two post-synaptic signals, one that protects itself and one that increases the probability of pruning for other synapses in the area.55,56 However, while this model works at the neuromuscular junction there is no direct evidence suggesting its accuracy for the central nervous system, although it has been the suggestion that complement proteins act as the protectors and punishers and induce microglia-based pruning.57,58

Recently microglia have become a leading candidate in the pruning process, including the process during development. The chief evidence to support this candidacy is electron microscopy imaging demonstrating microglia in the general proximity of RGC pre-synaptic inputs during the period of time when pruning is most likely to occur in developing dLGN.59,60 Further support comes from experiments where disruption of microglia during early development leads to incomplete and abnormal synaptic pruning.58 Based on this information one theory regarding synaptic pruning involves microglia CR3 receptors binding to C3 complement proteins produced in the pre-synaptic compartments due to a lack of glutamate release and appropriate feedback, which leads to synaptic engulfment.

However, there is an interesting immediate question regarding this potential role microglia play in developmental pruning? For example assume the pruning system fails and as a result autism develops. If microglia were involved in this process then how could microglia be available to eliminate/prune damaged neuronal connections later on in life? The general life span of individuals with autism, without another detrimental condition like epilepsy, implies that their brains are able to neutralize detrimental effects stemming from damaged neurons otherwise one would expect their life spans to be significantly shorter. How is this neutralization accomplished if the pruning process were damaged? The most plausible explanation is that there is a different mechanism to govern pruning during development versus during normal non-developmental operation. If so then specific experiments would need to be designed to determine the development pruning and developed pruning mechanisms.

Epilepsy frequency in autistic individuals is much higher than in non-autistic individuals with an occurrence rate ranging from 5% to 39-44% versus around 0.63%.61,62 The driving feature of epilepsy is excessive and abnormally consistent excitation of various neurons leading to abnormal synchronization resulting in spontaneous seizures and movements and potential brain damage. This additional excitation is thought to occur through one (or both) of two principle pathways: 1) excess excitatory neurotransmitters, usually glutamate, in synaptic clefts, which increase synaptic residence times increasing the probability of post-synaptic depolarization; 2) a disrupted inhibitory pathway usually due to a lack of GABA neurotransmitter release, which increases the probability of post-synaptic depolarization. Autistic individuals are thought to have an increased probability for developing epilepsy because of the loss of GABAergic interneurons in the cortical mini-columns and other areas.63,64 The general though is that because these mini-columns are less compact, but of greater number there is more extensive innervation for increased activation, but diminished lateral inhibition.64

It is thought that the relationship between autism and epilepsy is one in which the emergence of one increases the probability of developing the other. There is evidence to support the idea that autism develops prior to the clinical seizures that emphasize epilepsy and there is evidence to support the development of autism in individuals with more severe cases of epilepsy due to excessive neuronal damage.61,65 The development of epilepsy in autistic individuals could explain the level of neuronal death seen in some post-mortem autistic brains. Finally there appear to be two peak times of development of epilepsy in autistic individuals: early childhood and early adolescences.3 Not surprisingly younger preadolescent autistic children only have an epilepsy probability of slightly under 10%,66-68 due to limited compounded neuron damage, where adolescences and adults have an epilepsy probability of slight over 39%.69,70 This probability result makes sense as discussed above the nature of neuronal death and under/over excitement neuronal networks takes time to develop detrimental influence outside of severe cases of autism.

One of the interesting elements of autism is that the disorder shows a bias towards males by a ratio of around four to one.71 Numerous studies have supported this ratio through association of increased concentrations of testosterone in individuals with autism, but there is no clear understanding of any molecular mechanism that governs testosterone-based autistic augmentation.71-74 A recent study identified retinoic acid-related orphan receptor-alpha (RORA) as a possible candidate gene that influences autism development.16 RORA is responsible for regulating aromatase, an enzyme that converts testosterone to estrogen, thus a lack of aromatase will result in an increased concentration of testosterone.

There may be at least two different means of testosterone action within the pruning mechanism that could explain its role in increasing the probability of autism. First, testosterone may act directly to support synaptic pruning creating a situation where instead of multiple synapses being created through a “masking” process no strong synapses are created between certain pre-synaptic and post-synaptic cells.75 This action could be driven by testosterone augmenting the expression of complementary protein C, which may be the signal marker that induces microglia-based synapse elimination. It would seem that this type of process would be competitive with mGluR1 processes, thus if a given synapses was producing enough glutamate release and mGluR1 interaction it could ward off the effect of testosterone.

Successful application of such a response could create gaps in neuronal signaling disrupting synchronization and creating the under-connectivity environment associated with autism. An explanation to why such an outcome is not fatal could be explained by the existence of two different pruning systems, as mentioned earlier one that operates during development and one that operates after development. Therefore, over time due to brain plasticity the pre-synaptic cell may be able to partially extend an axon creating a smaller synaptic cleft, thus increasing synchronization probability. Whether or not more potent hormones like 5-adihydrotestosterone (DHT) have similar influence is unknown.

Second, the change in estradiol concentration, which is produced from cholesterol and as an active metabolic product of testosterone, may act as a neuroprotectant through two different mechanisms. The first mechanism involves enhancing RORA expression, which in turn reduces testosterone concentration limiting possible testosterone synapse elimination augmentation or other mechanism in which testosterone seems to enhance autism.16 The second mechanism involves estradiol down-regulation of mGluR1 and mGluR5.76 Down-regulation of mGluR1 could reduce the probability of abnormal mGluR1 enhancement through environmental or mutation-derived agonists, which would increase the probability of multiple synapses through masking, while also reducing the probability of excessive competition from testosterone. However, it should be acknowledged that too much estradiol may increase mGluR1 down-regulation to the point where it becomes more detrimental than beneficial.

One might ask how these testosterone influences could occur without genetic mutations. Recall that even without mutations, autism, and most neurological detrimental conditions for that instance, does not develop the same way in all individuals. There are issues of random quantum proximity where testosterone concentrations may change minutely to the point that exerts an effect that otherwise may not occur at a slight smaller concentration. For example even without mutations, testosterone has a negative feedback effect on RORA, which can create a small cascade under certain conditions greatly enhancing testosterone concentrations. Basically there is an element of randomness as well as other direct environmental elements.

While a defective pruning mechanism may not explain all autism cases, obvious there some cases that are dictated by certain genetic mutations that exist outside the pruning mechanism, it appears to be a plausible explanation for most cases. One of the major lingering questions regarding the failure of the pruning mechanism is whether it prunes too many synapses or not enough. While one of the hallmarks of autism appears to be functional underconnectivity that leads to neuronal hypo-synchronization and a general lack of neuronal firing/excitation there is also a level of overconnectivity.77-80 For example autistic individuals suffer underconnectivity when having to process dynamic events like facial movements and auditory elements, but exhibit overconnectivity when having to process static images.81

This general neuronal trend in autism has been expanded through fMRI to demonstrate reduced functional long-range connectivity, which would affect synchronization, versus short-range connectivity.82-84 This reduced long-range connectivity could center around an imbalance between short-range excitatory and long-range inhibitory connections, involve the death of PCs.85 and negatively influence the interaction between cortical and limbic systems producing extreme negative emotional responses to change.86 So how does this over-connectivity/under-connectivity contrast develop from the pruning process?

The overall mechanism of pruning could be viewed in the following manner. Numerous synapses form, but none are dominant. The multiple synapses can be thought to be in competition with each other, so when receiving neuronal input from the cell body and dendrites the total potential of that input is divided between the synapses with the total weight of that division unknown. Suppose for the purpose of this example there are five synapses receiving inputs of 15%, 15%, 20%, 20% and 30%. Over time due to higher weighted percentages, more glutamate vesicles, greater feedback from post-synaptic response, etc. one synapse starts to exert dominance receiving more of the initial excitation where the weaker synapses receive less. Returning to the above breakdown after a short period the inputs change to something like 7%, 8%, 14%, 17% and 54%. Eventually the weaker synapses are eliminated and the single dominant synapse receive 100% of the dendritic input.

In scenarios where pruning is not functioning properly one of two situations can arise: 1) pruning fails to eliminate all but one of the existing synapses leaving more than one synapse feeding into post-synaptic dendrites; 2) pruning removes all synapses leaving no synapse feeding into post-synaptic dendrites. It is understandable to assume that the first situation is more frequent than the second because in the second the post-synaptic neuron would have a higher probability of future apoptosis and potential brain damage. The first means of failure (multiple synapses surviving) can be further broken down into two failures: 1) a breakdown in the pruning system itself during development where a dominant synapse is created, but the weaker synapses are not pruned because of a deficiency in the pruning pathway; 2) weakness masking where weaker synapses are masked and interpreted by the pruning pathway as stronger and unworthy of pruning.

The first means of failure is rather self-explanatory where one or more of the agents responsible for pruning are not functioning properly. The second is a little more complicated because of multiple potential targets. The most obvious means is some external environmental agent acting as an agonist for mGluR1. Most likely the system that evaluates the winning synapse does not function on a relative scale (synapse 1 is releasing 70% of the glutamate versus synapse 2 releasing 30% so synapse 2 should be pruned), but instead on an absolute scale with a competition threshold (synapse 1 is releasing 70 mmols of glutamate versus synapse 2 releasing 30 mmols of glutamate so synapse 2 should be pruned because it is not releasing enough “don’t prune me” compounds stemming from that glutamate release). Thus mGluR1 agonist binding could increase overall glutamate release from all synapses “masking” synapse weakness resulting in less overall pruning. This additional agonist from an environmental agent could also help explain the increased rate of autism if the agent is a toxin of some sort.

Again it is difficult to reconcile the bigger brain seen in most autistic patients and less connectivity from excessive pruning, so one must assume for the moment that the above dynamic is correct that there are regions of the brain with multiple synapses instead of just one. How does an environment of multiple synapses create a reduced level of connectivity and activation for dynamic visual and/or audio processes? One possible explanation is that these synapses function differently than normal synapses.

In a multiple synapse situation none of the synapse may extend as far as the single dominant synapse situation, thus there is a larger than normal synaptic cleft. This scenario could make sense because translocation involves signals from both the pre and post-synaptic regions and one could expect a limited amount of signal from the post-synaptic region. If this signal is spread between multiple targets versus a single target shorter translocation is possible. This larger synaptic cleft reduces the total amount of neurotransmitter that is able to bind to receptors on the appropriate post-synaptic cleft. In such a scenario even if the multiple synapses were able to provide more total neurotransmitters due to firing similarity and rate, the additional space that the neurotransmitter has to cover should limit the total binding potential. For example if one were to standardize the one-one situation at a binding potential of 1 a larger synaptic cleft in a multiple synapse scenario would have a binding potential less than 1. This lower binding potential would reduce the activation probability of the post-synaptic cell resulting in slower and/or less neuronal synchronization, which is required for optimal processing of dynamic stimulation where multiple isolated firings of single neurons is limited.

Two possible synaptic explanations for the over-excitation when an autistic individual experiences a static stimulus could involve a lack of axonal evolution and a lack of feedback. First, as previously mentioned developing synapses function on a graduated release system over an all or nothing system, which later develops in more mature neuronal systems. Perhaps this evolution does not occur in the presence of multiple synapses and the system maintains the graduated release response. With this response it would be easier for the system to induce multiple firings during periods of intent focus (static stimulus), thus releasing more neurotransmitters than normal synaptic areas creating a state of over-excitation or over-synchronization. Second, non-pruned synaptic areas could have abnormal feedback systems, which could result in multiple firings and excess neurotransmitters in the cleft. Taking the binding potential example from above instead of being less than 1 in the situation of static stimuli autistic individuals would have a binding potential at these multiple synapses of greater than 1.

As mentioned above there is a question to how the rate of autism prevalence is increasing at the current rate. It stands to reason that natural genetic mutations are not the cause of this increase due to the required increased variance rate. Therefore, it makes sense to assume that there is some environmental catalyst that is increasing autism prevalence rates. Assume for a moment that these environmental factors exist how do they influence autism rates? There appear to be two valid rationalities: 1) These environmental factors influence the rate of genetic mutation in one or more key genetic neuronal development elements; 2) These factors influence the rate of activity, either positively or negatively, for neuronal development. For example one popular rational could be different factors acting as agonists or antagonists on mGluR1, which could influence pruning rates resulting in the increase in autism rates.

From a standpoint of treatment if abnormal pruning is the chief problem then addressing the pruning would be a little tricky because of the multiple elemnets involved in the pruning mechanism. The principle target would be mGluR1 and based on initial assumptions one would have to neutralize multiple synapse masking, thus one would pharmaceutically introduce an mGluR1 antagonist. However, the dosage and effect this antagonist would have on other parts of the brain is unclear. Autistic diagnosis perhaps could be improved by early observation of visual and motor function over emotional and communication abnormalities based on assumed abnormal pruning alterations. Overall while the pruning theory for autism has been proposed before at points in the past87 it seems to have been neglected in the mist of more gene research, which is unfortunate and is more than likely detrimental to the future treatment of autism.


Citations –

1. Gepner, B, and Tardif, C. “Autism, movement, time and thought. E-motion mis-sight and
other temporo-spatial processing disorders in autism.” in: M. Vanchevsky (ed). Frontiers
in Cognitive Psychology. New York: Nova Science Publishers. 2006. 1-30.

2. Volkmar, R and Pauls, D. “Autism”. Lancet. 2003. 362:1133-1141.

3. Lecavalier, L. “Behavioral and emotional problems in young people with pervasive
developmental disorders: relative prevalence, effects of subject characteristics, and
empirical classification.” J Autism Dev Disord. 2006. 36:1101-1114.

4. Rutter, M. “Incidence of autism spectrum disorders: changes over time and their
meaning.” Acta Paediatr. 2005. 94:2-15.

5. Sullivan, M. “Autism increase not a result of reclassification.” Clin. Psychiat. News. 2005. May:68.

6. Posey, D, et Al. “Antipsychotics in the treatment of autism.” J. Clin. Invest. 2008. 118:6–14.

7. Ecker, C, et Al. “Describing the brain in autism in five dimensions–magnetic resonance imaging-assisted diagnosis of autism spectrum disorder using a multiparameter classification approach.” J. Neurosci. 2010. 30:10612–10623.

8. Gepner, B, and Féron, F. “Autism: a world changing too fast for a mis-wired brain?”
Neuroscience and Biobehavioral Reviews. 2009. 33:1227-1242.

9. Abrahams, B, and Geschwind, D. “Advances in autism genetics: on the threshold of a
new neurobiology.” Nat Rev Genet. 2008. 9:341-355.

10. Persico, A, and Bourgeron, T. “Searching for ways out of the autism maze: genetic,
epigenetic and environmental cues.” Trends Neurosci. 2006. 29:349-358.

11. Fassio, A, et Al. “SYN1 loss-of-function mutations in autism and partial epilepsy cause impaired synaptic function.” Human Molecular Genetics. 2011 20:2297–2307.

12. Hamdan, F, et Al. “De novo SYNGAP1 mutations in nonsyndromic intellectual disability and autism.” Biological Psychiatry. 2011. 69:898–901.

13. Bozdagi, O, et Al. “Haploinsufficiency of the autism-associated Shank3 gene leads to deficits in synaptic function, social interaction, and social communication.” Molecular Autism. 2010. 1:15.

14. Durand, C, et Al. “Mutations in the gene encoding the synaptic scaffolding protein SHANK3 are associated with autism spectrum disorders.” Nature Genetics. 2006. 39:25–27.

15. Laumonnier, F, et Al. “X-linked mental retardation and autism are associated with a mutation in the NLGN4 gene, a member of the neuroligin family.” American Journal of Human Genetics. 2004. 74:552–557.

16. Sarachana, T, et Al. “Sex Hormones in Autism: Androgens and Estrogens Differentially and Reciprocally Regulate RORA, a Novel Candidate Gene for Autism.” PLoS ONE. 2011. 6(2):e17116.

17. Campbell, D, et Al. “A genetic variant that disrupts MET transcription is associated with autism.” PNAS. 2006. 103(45):16834–9.

18. Hu, V, et Al. “Gene Expression Profiling of Lymphoblasts from Autistic and Nonaffected Sib Pairs: Altered Pathways in Neuronal Development and Steroid Biosynthesis.” PLoS ONE. 2009. 4(6):e5775.

19. Awadalla, P, et Al. “Direct measure of the de novo mutation rate in autism and schizophrenia cohorts.” American Journal of Human Genetics. 2010. 87:316–324.

20. O’Roak, B, et Al. “Exome sequencing in sporadic autism spectrum disorders identifies severe de novo mutations.” Nature Genetics. 2011. 43:585–589.

21. Crepel, F. “Maturation of climbing fiber responses in the rat.” Brain Res. 1971. 35:272-276.

22. Mariani, J, and Changeux, J. “Ontogenesis of olivocerebellar relationships. I Studies by intracellular recordings of the multiple innervation of Purkinje cells by climbing fibers in the developing rat cerebellum.” J. Neurosci. 1981. 1:696-702.

23. Crepel, F, Mariani, J, and Delhaye-Bouchaud, N. “Evidence for a multiple innervation of Purkinje cells by climbing fibers in the immature rat cerebellum.” J. Neurobiol. 1976. 7:567-578.

24. Watanabe, M, and Kano, M. “Climbing fiber synapse elimination in cerebellar Purkinje cells.” European Journal of Neuroscience. 2011. 34(10):1697-1710.

25. Hazlett, H, et Al. “Magnetic resonance imaging and head circumference study of brain size in autism: birth through age 2 years.” Arch. Gen. Psychiatry. 2005. 62:366–1376.

26. Courchesne, E, et Al. “Unusual brain growth patterns in early life in patients with autistic disorder: an MRI study.” Neurology. 2001. 57:245–254.

27. Sparks, B, et Al. “Brain structural abnormalities in young children with autism spectrum disorder.” Neurology. 2002. 59:184–192.

28. Courchesne, E, “Autism at the beginning: microstructural and growth abnormalities underlying the cognitive and behavioral phenotype of autism.” Dev Psychopathol. 2005. 17:577-597.

29. Courchesne, E, et Al. “Mapping early brain development in autism.” Neuron. 2007 56:399-413.

30. Courchesne, E, Carper, R, and Akshoomoff, N. “Evidence of brain overgrowth in the first year of life in autism.” JAMA. 2003. 290:337–344.

31. Richler, J, et Al. “Is there a “regressive phenotype” of autism spectrum disorder associated with the measles-mumps-rubella vaccine? A CPEA study.” Journal of Autism and Developmental Disorders. 2006. 36:299–316.

32. Baird, G, et Al. “Regression, developmental trajectory and association problems in disorders in the autism spectrum: The SNAP study.” Journal of Autism and Developmental Disorders. 2008. 38:1827–1836.

33. Bauman, M, and Kemper, T. “Neuroanatomic observations of the brain in autism: a
review and future directions.” Int J Dev Neurosci. 2005. 23:183-187.

34. Whitney, E, et Al. “Density of cerebellar basket and stellate cells in autism: Evidence for a late developmental loss of Purkinje cells.” J. Neurosci Res. 2009. 87(10):2245–2254.

35. Huttenlocher, P, and Dabholkar, A. “Regional differences in synaptogenesis in human cerebral cortex.” Journal of Comparative Neurology. 1997. 387:167–178.

36. Huttenlocher, P. “Neural plasticity: The effects of environment on the development of the cerebral cortex.” 2002. Cambridge, MA: Harvard University Press.

37. Kawabata, S, et Al. “Control of calcium oscillations by phosphorylation of metabotropic glutamate receptors.” 1996. 89:92.

38. Li, X, et Al. “JNK1 contributes to metabotropic glutamate receptor-dependent long-term depression and short-term synaptic plasticity in the mice area hippocampal CA1.” Eur. J. Neurosci. 2007. 25(2):391–96.

39. Bordi, F, and Ugolini, A. “Group I metabotropic glutamate receptors: implications for brain diseases.” Progress in neurobiology. 1999. 59(1):55.

40. Page, G, et Al. “Group I metabotropic glutamate receptors activate the p70S6 kinase via both mammalian target of rapamycin (mTOR) and extracellular signalregulated kinase (ERK 1/2) signaling pathways in rat striatal and hippocampal synaptoneurosomes.” Neurochem. Int. 2006. 49(4):413–21.

41. Hou, L, and Klann, E. “Activation of the phosphoinositide 3-kinase-Akt-mammalian target of rapamycin signaling pathway is required for metabotropic glutamate receptor-dependent long-term depression.” J. Neurosci. 2004. 24(28):6352–61.

42. Niswender, C, and Conn, J. “Metabotropic Glutamate Receptors: Physiology, Pharmacology, and Disease.” Annu Rev Pharmacol Toxicol. 2010. 50:295–322.

43. Suzuki, Y, et Al. “Negative cooperativity of glutamate binding in the dimeric metabotropic glutamate receptor subtype 1.” J. Biol. Chem. 2004. 279(34):35526–34.

44. Brock, C, et Al. “Activation of a dimeric metabotropic glutamate receptor by intersubunit rearrangement.” J. Biol Chem. 2007. 282:33000–33008.

45. Anwyl, R. “Metabotropic glutamate receptors: electrophysiological properties and role in plasticity.” Brain Res. Rev. 1999. 29(1):83–120.

46. Bellone, C, Luscher, C, and Mameli, M. “Mechanisms of synaptic depression triggered by metabotropic glutamate receptors.” Cell. Mol. Life Sci. 2008. 65(18):2913–23.

47. Pinheiro, P, and Mulle, C. “Presynaptic glutamate receptors: physiological functions and mechanisms of action.” Nat. Rev. Neurosci. 2008. 9(6):423–36.

48. Ronesi, J, and Huber, K. “Homer interactions are necessary for metabotropic glutamate receptor induced long-term depression and translational activation.” J. Neurosci. 2008. 28(2):543–47.

49. Aiba, A, et Al. “Deficient cerebellar long-term depression and impaired motor learning in mGluR1 mutant mice.” Cell. 1994. 79(2):377–88.

50. Levenes, C, et Al. “Incomplete regression of multiple climbing fibre innervation of cerebellar Purkinje cells in mGLuR1 mutant mice.” Neuroreport. 1997. 8(2):571–74.

51. Sachs, A, et Al. “The mouse mutants recoil wobbler and nmf373 represent a series of Grm1 mutations.” Mamm. Genome. 2007. 18(11):749–56.

52. Matsuda, K, et Al. “Cbln1 is a ligand for an orphan glutamate receptor g2, a bidirectional synapse organizer.” Science. 2010. 328:363-368.

53. Takeuchi, T, et Al. “Control of synaptic connection by glutamate receptor g2 in the adult cerebellum.” J. Neurosci. 2005. 25:2146-2156.

54. Nakamura, M, et Al. “Signaling complex formation of phospholipase Cbeta4 with mGluR1apha and IP3R1 at the perisynapse and endoplasmic reticulum in the mouse brain.” Eur. J. Neurosci. 2004. 20:2929-2944.

55. Balice-Gordon, R, and Lichtman, J. “Long-term synapse loss induced by focal blockade of postsynaptic receptors.” Nature. 1994. 372:519–24.

56. Jennings, C. “Developmental neurobiology. Death of a synapse.” Nature. 1994. 372:498–99.

57. Huberman, A, Feller, M, and Chapman, B. “Mechanisms underlying development of visual maps and receptive fields.” Annu. Rev. Neurosci. 2008. 31:479–509

58. Stephan, A, Barres, B, and Stevens, B. “The Complement System: An Unexpected Role in Synaptic Pruning During Development and Disease.” Annu. Rev. Neurosci. 2012. 35:369–89.

59. Hughes, V. “The Constant Gardener.” Nature. 2012. 485(31): 570-572.

60. Schafer, D, et Al. “Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner.” Neuron. 2012. 74(4):691-705.

61. Tuchman, R, and Rapin, I. “Epilepsy in autism.” Lancet Neurol. 2002. 1:352–358

62. Moshe, S, and Tuchman, R. “Convulsing toward the pathophysiology of autism.” Brain Dev. 2009. 31(2):95–103.

63. Casanova, M, et Al. “Minicolumnar pathology in autism.” Neurology. 2002. 58:428–432.

64. Casanova, M, Buxhoeveden, D, and Brown, C. “Clinical and macroscopic correlates of minicolumnar pathology in autism.” J. Child Neurol. 2002. 17:692–695.

65. Giedd, J, et Al. “Brain development during childhood and adolescence: a longitudinal MRI study.” Nat. Neurosci. 1999. 2:861–863

66. Voigt, R, et Al. “Laboratory evaluation of children with autistic spectrum disorders: a guide for primary care pediatricians.” Clin Pediatr (Phila). 2000. 39:669–671.

67. Fattal-Valevski, A, et Al. “Characterization and comparison of autistic subgroups: 10 years’ experience with autistic children.” Dev Med Child Neurol. 1999. 41:21–25.

68. Hoshino, Y, et Al. “Clinical features of autistic children with setback course in their infancy.” Jpn J Psychiatry Neurol. 1987. 41:237–245.

69. Giovanardi-Rossi, P, Posar, A, and Parmeggiani, A. “Epilepsy in adolescents and young adults with autistic disorder.” Brain Dev. 2000. 22:102–106.

70. Kawasaki, Y, et Al. “Brief report: electroencephalographic paroxysmal activities in the frontal area emerged in middle childhood and during adolescence in a follow-up study of autism.” J. Autism Dev Disord. 1997. 27:605–620.

71. Baron-Cohen, S, Knickmeyer, R, and Belmonte, M. “Sex differences in the brain: Implications for explaining autism.” Science. 2005. 310(5749):819–23.

72. Auyeung, B, et Al. “Foetal testosterone and autistic traits in 18 to 24-month-old children.” Molecular Autism. 2010. 1(1):11.

73. Chura, L et Al. “Organizational effects of fetal testosterone on human corpus callosum size and asymmetry.” Psychoneuroendocrinology. 2010. 35(1):122–132.

74. Just, M, et Al. “Functional and anatomical cortical underconnectivity in autism: Evidence from an FMRI study of an executive function task and corpus callosum morphometry.” Cereb. Cortex. 2007. 17(4):951 61.

75. Zehr, J, et Al. “Adolescent Development of Neuron Structure in Dentate Gyrus Granule Cells of Male Syrian Hamsters.” Develop Neurobiol. 2008. 68:1517–1526.

76. Hilton, G, et Al. “Glutamate-mediated excitotoxicity in neonatal hippocampal neurons is mediated by mGluR-induced release of Ca++ from intracellular stores and is prevented by estradiol.” Eur J Neurosci. 2006. 24:3008–16.

77. Nishitani, N, Avikainen, S, and Hari, R. “Abnormal imitation-related cortical activation sequences in Asperger’s syndrome.” 2004. Ann Neurol. 55:558-562.

78. Wickelgren, I. “Autistic brains out of synch?” Science. 2005. 308:1856-1858.

79. Geschwind, D, and Levitt, P. “Autism spectrum disorders: developmental disconnection syndromes.” Curr Opin Neurobiol. 2007. 17:103-111.

80. Minshew, N, and Williams, D. “The new neurobiology of autism: cortex, connectivity and neuronal organization.” Arch Neurol. 2007. 64:945-950.

81. Bertone, A, et Al. “Motion perception in autism: a ‘complex’ issue.” J. Cogn Neurosci. 2003. 15:218-225.

82. Belmonte, M, et Al. “Autism and abnormal development of brain connectivity.” J. Neurosci. 2004. 24:9228-9231.

83. Williams, E, and Casanova, M. “Autism and dyslexia: A spectrum of cognitive styles as defined by minicolumnar morphometry.” Medical Hypotheses. 2010. 74:59–62.

84. Koshino, H, et Al. “fMRI investigation of working memory for faces in autism: Visual coding and underconnectivity with frontal areas.” Cerebral Cortex. 2008. 18:289–300.

85. Gustafsson, L. “Inadequate cortical feature maps: A neural circuit theory of autism.” Biological Psychiatry. 1997. 42:1138–1147.

86. Grossberg, S, and Seidman, D. “Neural dynamics of autistic behaviors: Cognitive, emotional, and timing substrates.” Psychological Review. 2006. 113:483–525.

87. Frith, C. “What do imaging studies tell us about the neural basis of autism?” In G. Bock & J. Goode (Eds.), Autism: Neural basis and treatment possibilities: Novartis Foundation Symposium. 2003. 251:149–176. Chichester, England: Wiley.