A spark of insight into the mystery of autism

It was her husband's tinkering with their music collection, Geraldine recalled, that nearly ended their marriage. "Michael rearranged all our classical music CDs in order of the birth-date of the composer, of all things, and I couldn't find anything," she told me. Michael's response at the time had been a robotically mouthed, "Well, I'm sorry you feel that way". Cue airborne crockery and threats of divorce.

Their marriage was saved when Michael was diagnosed with Asperger's syndrome, a high-functioning form of autism. Geraldine no longer saw a cold, unsympathetic husband, but a decent man struggling to navigate a relationship without the neurological compass that guides normal social interaction. He had since learned to make eye contact and to appear sincere and less aloof, a near-theatrical feat aided by his considerable intelligence.

This memorable interview, which I conducted a decade ago, sprang to mind as I read Wednesday's report about some high-IQ children appearing to shed symptoms of autism as they grow older. It is a striking finding, because autism - or, more correctly, autism spectrum disorder - has long been thought to be a life sentence. At the severe end of the spectrum, that will probably remain tragically so for the teenagers still in nappies who have never spoken a word nor initiated a cuddle with their tireless mothers. But for young children at the milder end, who perhaps show a delay in language acquisition or challenging behaviour, the research offers a spark of hope that some therapies - or even spontaneous neurological good fortune - might push a lucky few off the spectrum and into the realms of clinical normality. It is also a reminder that scientific investigation - particularly when it comes to the human brain - has the capacity to confound and surprise, and to uncover new avenues in areas dominated by cul-de-sacs.

The study, funded by the US National Institutes of Health and published in the Journal of Child Psychology and Psychiatry, looked at 34 young people aged from eight to 21, who were previously assigned an autism diagnosis but were now considered not to have it, and were being schooled with no special help. Their earlier problems ranged from an inability to read faces to communication difficulties and repetitive behaviour. Researchers double-checked that those original diagnoses were correct, and then found a comparable group (matched for age, sex and non-verbal IQ) of young people with high-functioning autism. The most notable difference between the two groups turned out to be verbal IQ: those who had "shed" their autism had verbal IQs around seven points higher than the high-functioning group.

As some observers have noted, this correlation superficially makes sense: it takes intelligence to master coping strategies like those employed by Michael. But, as with all studies, we must note the caveats. The sample size is relatively small; as is the difference in IQ. The researchers, led by Deborah Fein at the University of Connecticut, have yet to delve into the backgrounds of these "recovered" individuals, to see if something other than IQ is at work, perhaps behavioural therapy for example. They carefully describe their findings as illustrating a "loss of autism diagnosis"; the term "cure" is deliberately, and properly, avoided. The National Autistic Society notes that there is no cure for autism, although some interventions, such as counselling and pictorial learning, can help the estimated half a million people in Britain on the autism spectrum.

Another possibility is that these clever children never genuinely shake off their condition but are instead able to summon up a masquerade of social adeptness to evade diagnosis. Furthermore, measuring and describing autism is not an exact science. For example, Asperger's syndrome and high-functioning autism are often conflated - the distinction is that those with Asperger's do not show delayed language development as children. There will always be jostling about where the spectrum begins - and whether mildly autistic individuals should simply be considered "different" rather than "disordered".

Equally, we must not be too quick to condemn the results as irrelevant and giving only false hope. All medical inroads begin with a glimpse of the unexpected, with a finding that defies conventional explanation and demands investigation. Science certainly hasn't got autism completely figured out: only now are we beginning to peer inside the black box that is the human brain, the mysterious "executive hub" that governs how we engage with other people.

The biological architecture of that hub differs in each person; it is sculpted by early environmental factors, genes and daily life (which is why London cabbies have large hippocampi, reflecting the expansion of the brain region that deals with navigation). The brain is not inert: it is an ever-changing organ, continually pruning old neural connections, making new ones, and rewiring itself in response to new experiences. That is how some stroke patients regain lost abilities; and that might be what is happening here, in a minority of children with autism.

The idea of "recovery" from autism may be controversial but the imperative is clear: researchers must not be afraid to walk down new avenues.