But scientists have much to learn about the condition, and their understanding of visual processing in autism had been obscured by the use of tests that couldn't accurately tell them the limits of people's visual performance.
'Local processing' and 'global processing'
Professor Dakin said the accepted view was that people with autism could perform "local processing" - or the fine detail of the world - but could not do "global processing", or the overall context.
But for the first time, he and fellow researchers at University College in London have found a way of separating the two functions in the lab.
The study tested 33 children with autism and 33 ability matched typical children on a more robust task - reporting mean direction, both with and without directional variability.
"When we apply it to children with autism, we find they don't have a global processing problem at all," he said.
"If anything, they are better at global processing."
Where the previous assumption had been that someone was bad at tasks that required pooling information because they simply couldn't pool it, he and his team now believed people with autism couldn't ignore irrelevant information - something quite different.
"That's about your ability to resist noise, and so we think it's related to sensory overload. Their sensory overload is an inability to resist the irrelevant," Professor Dakin said.
"If anything, you could almost think of that as more integration - the exact opposite of what people have thought.
"Our work shows that integration is fine in people with autism, they just can't ignore noise."
For example, when someone wanted to judge the overall direction of a shoal of fish or a crowd of people, they had to integrate motion signals across space and time.
This was usually tested using a standard "motion coherence paradigm" in which people reported the direction of coherently moving dots amid randomly moving noise dots.
"Poor performance by autistic individuals on this task has been widely interpreted as evidence of disrupted integrative processes, but critically, motion coherence is not limited only by the ability to pool information," Professor Dakin said.
"It can also be limited by a lack of precision estimating the direction of elements or by difficulties separating motion signals from noise."
Professor Dakin said the inability to deal with noise may apply in other aspects of Autism Spectrum Disorders such as other sensory systems and cognitive functions.
"If we have a better understanding of how these things work, then we can develop better approaches for dealing with them, such as in the way we present information to people with ASD, and the way we train people to improve their performance in these areas."
Study implications
In a blog post on the study, the University of Auckland's Head of Psychology, Professor Will Hawyard, pondered what being able to effectively integrate visual elements together might mean.
"One is tempted to think of a story of how it might provide an opportunity for savant skills to develop, such as an enhanced musical ability," he wrote.
"We must be careful, however, not to equate better performance on a psychophysical task with superior abilities in complex domains like music and mathematics.
"Rather, this may be an indicator of maladaptive processing. The authors speculate that such powerful integration might lead to the sensory overload that so many kids with autism subjectively experience."
In any event, Professor Hayward said, the researchers had shown the advantage to testing behaviour of developmentally delayed children using sophisticated techniques that have been carefully implemented.
"More work like this will give us a better insight into the world of autism: a world that is difficult for its inhabitants to describe to the rest of us."
