It found several genetic variants commonly associated with ASD, all of them pointing two specific genes found on chromosome 5 which control production of proteins which help cells stick to each other, and make nervous connections.

One variant, linked to a gene called CDH10, was so common - present in over 65% of cases of autism - that the researchers calculated that fixing it would cut the number of autism cases by 15%.

They also linked ASD rather less strongly to a group of about 30 genes which produce proteins that play a key role in enabling brain cells to migrate to correct places, and to connect to neighbouring cells.

Professor Simon Baron-Cohen, an autism expert at the University of Cambridge, said 133 genes had now been linked to the condition, and much work was needed to piece together how they interacted with each other and the environment.

The National Autistic Society said the exact causes of autism were unknown. In a statement, the society said: "There is evidence to suggest that genetic factors are responsible for some forms of autism. "However, the difficulty of establishing gene involvement is compounded by the interaction of genes and by their interaction with environmental factors. "Various studies over many years have sought to identify candidate genes but so far inconclusively."

Previous studies have suggested between 3 percent and 25 percent of autistic kids recover. Fein says her studies have shown the range is 10 percent to 20 percent.

But even after lots of therapy - often carefully designed educational and social activities with rewards - most autistic children remain autistic. Recovery is "not a realistic expectation for the majority of kids," but parents should know it can happen, Fein said.

The children in Fein's study, which is still ongoing, were diagnosed by an autism specialist before age 5 but no longer meet diagnostic criteria for autism. The initial diagnoses were verified through early medical records.

The researchers are also doing imaging tests to see if the recovered kids' brains look more like those of autistic or nonautistic children.

Imaging scans also are being done to examine brain function in formerly autistic kids.

Results from those tests are still being analyzed.

Most of the formerly autistic kids got long-term behavior treatment soon after diagnosis, in some cases for 30 or 40 hours weekly.

Many also have above-average IQs and had been diagnosed with relatively mild cases of autism. At age 2, many were within the normal range for motor development, able to walk, climb and hold a pencil.

Significant improvement suggesting recovery was evident by around age 7 in most cases, Fein said.

None of the children has shown any sign of relapse. But nearly three-fourths of the formerly autistic kids have had other disorders, including attention-deficit problems, tics and phobias; eight still are affected.

Getting back to the genetics of autism, current models are therefore consistent with what is being found when the genetics of autism is researched - researchers are finding many genes that predispose to autism in a subset of cases but no single or simple universal cause. At present, 133 different gene variants have been linked to autism.

This new research, conducted by Dr. Hakon Hakonarson of the Children’s Hospital of Philadelphia, is a genome wide analysis involving about 10,000 individuals.

The results are especially significant because the variants lie between two genes, called CDH9 and CDH10, which are known to play an important role in forming nerve connections in the brain.

The gene variants that correlated with ASD are for proteins that are involved in the process of neurons forming connections with each other. There is already other lines of evidence that suggest what is different in ASD brains is a decrease in the amount of interconnectedness and communication among neurons. It is therefore likely no coincidence that this study found genetic correlations for proteins involved with neuronal connections.

This also is compatible with the finding that many separate genes are potentially involved with ASD - for there are many separate genes and processes involved with forming and maintaining neuronal connections.

Once we learn about this difference, we can then look for medications that block the consequences of the variant and once we make sure they are safe, we can then start testing these new medications in children who are at risk of developing autism, with the objective of preventing autism (i.e., avoid breakdown in connections between nerves and abnormality in brain connectivity).

Response: Yes, all of them could be tested in utero; we have identified 10 new variations (9 rare and 1 common) and we have replicated (and confired) four other once that were previously published (neurexin 1, contactin 4, 15q11 and 22q11). However, we do not have a yes or no answer as to whether the fetus will be autistic -- but if we are testing a fetus in an autistic family the value of the test is much higher.

The upshot of these differences is that the columns in an autistic brain seem to be more connected than normal with their close neighbours, and less connected with their distant ones. Though it is an interpretative stretch, that pattern of connection might reduce a person’s ability to generalise (since disparate data are less easily integrated) and increase his ability to concentrate (by drawing together similar inputs).

Dr Snyder argues that savant skills are latent in everyone, but that access to them is inhibited in non-savants by other neurological processes. He is able to remove this inhibition using a technique called repetitive transcranial magnetic stimulation.

all the success stories related inside the pages of EPSS are accompanied by fortuitous exposure to various forms of structured material, clearly marking out the most promising path for autistic development and growth.

And of course these discussions fly in the face of the current clamor for early intervention in autism, where it would seem the goal is always to yank the autistic child away as early as possible from his or her preferred method of engaging the world, and substitute instead an intense bombardment of socially based indoctrination, hoping to turn the child around while there is still time for the “malleable” brain to be re-molded.

In my opinion, it is no mere coincidence that the very same elements that stand at the core of the Mottron team's affirmative description of autistic perception and cognition are also the very same elements that stand at the core of humanity’s sudden departure off the savannah and leap into the modern world.

Admittedly, a thorough discussion of such a topic would be much too large for inclusion in an academic research paper such as EPSS, but the fact that the Mottron team does not mention, or even hint at, the connection between the features of autistic perception and the features of human cognitive history leave it unclear whether the team has ever considered such a connection.

At this point in time, the Mottron team seems to be the only autism research team heading in a positive and enlightening direction—a direction that is constructive for autistic individuals everywhere—and I look forward to all their future contributions.

In 2006 your lab published a paper in Nature Neuroscience linking a mirror neuron dysfunction to autism. How might reduced mirror neuron activity explain the symptoms of autism?

Reduced mirror neuron activity obviously weakens the ability of these patients to experience immediately and effortlessly what other people are experiencing, thus making social interactions particularly difficult for these patients. Patients with autism have also often motor problems and language problems. It turns out that a deficit in mirror neurons can in principle explain also these other major symptoms. The motor deficits in autism can be easily explained because mirror neurons are just special types of premotor neurons, brain cells essential for planning and selecting actions. It has been also hypothesized that mirror neurons may be important in language evolution and language acquisition.

Thus, a deficit in mirror neurons can in principle account for three major symptoms of autism, the social, motor and language problems.

There is convincing behavioral evidence linking media violence with imitative violence. Mirror neurons provide a plausible neurobiological mechanism that explains why being exposed to media violence leads to imitative violence.

I think there are two key points to keep in mind. The first one is the one we started with: mirror neurons are brain cells specialized for actions. They are obviously critical cells for social interactions but they can’t explain non-social cognition. The second point to keep in mind is that every brain cell and every neural system does not operate in a vacuum. Everything in the brain is interconnected, so that the activity of each cell reflects the dynamic interactions with other brain cells and other neural systems.

Nonetheless, this incidence is about 100 times greater than the rate at which autism affects unrelated people in the population.

One study showed that the likelihood that the identical twin of an autistic child also would be autistic was 82 percent, whereas the equivalent rate for fraternal twins was only 10 percent.

With sophisticated statistical techniques and numerous twin studies, behavioral geneticists now believe that as much as 90 percent of the behavioral phenotype of autism is related to inherited genes.

Somali women at high lattitudes may be at even greater risk of vitamin D deficiency if they observe the tradition of covering their skin.