We know that for autism, the causes and changes to the brain are happening long before birth. But in a groundbreaking new study, an intervention in infants showing early signs of autism has been able to reduce clinical diagnosis by two-thirds.

Autism spectrum disorder ( ASD) describes a wide-ranging set of conditions affecting a person's social, communication, and motor skills. Diagnosis is based on criteria outlined in the American Psychiatric Association's DSM-5 – such as persistent deficits in social interactions and reciprocating emotions, an absence of interest in friends, repetitive movements or speech, and extreme or unusual reactions to stimuli.

"These findings are the first evidence that a pre-emptive intervention during infancy could lead to such a significant improvement in children's social development that they then fell below the threshold for a clinical diagnosis of autism," says one of the study authors, University of Manchester child psychiatry researcher Jonathan Green.

"Many therapies for autism have tried previously to replace developmental differences with more 'typical' behaviors. In contrast, iBASIS-VIPP works with each child's unique differences and creates a social environment around the child that helps them learn in a way that was best for them."

This iBASIS-Video Interaction to Promote Positive Parenting (or VIPP) is what the team calls a parent mediated therapy. This is not in any way intended to be a 'cure' for autism, but an approach aimed at "reducing the long-term disability" of ASD.

Normally, diagnosis can occur from around two years of age, but there are also signs that can occur much earlier, such as avoiding eye contact and using fewer words than their peers. It's these early symptoms that the researchers are interested in, as making small changes early on could lead to significantly better developmental outcomes later.

The researchers tracked 103 infants who had these early signs of ASD, aged as young as nine months all the way through to three years, in a randomized, blinded experiment.

Fifty of the infants received iBASIS-VIPP – a treatment that teaches parents to change the way they interact with their babies to stimulate their socio-communicative development, while the remaining 53 received normal care.

The results were staggering – of those who had received the iBASIS-VIPP treatment, only 3 of the 45 participants who were tested again at age three met the clinical threshold for being diagnosed with autism, versus 9 of the 45 who had received regular care. That's a difference of two-thirds.

"This is a bit of a holy grail in the area of child health," first author, Telethon Kids Institute autism researcher Andrew Whitehouse, told the Australian Broadcasting Corporation.

"What we have found is providing a new clinical model, identifying children as early as possible in life, providing supportive intervention to help them be who they want to be, we can actually reduce the clinical criteria for autism by two-thirds."

The finding is a remarkable first step, but there are important details in evaluating this approach that still need to be ironed out.

Firstly, the children in this trial were only followed until age three, and although it's a typical cut-off age for clinically diagnosing ASD in someone, it's still possible for symptoms in some of these children to shift, change, or emerge later in life.

Also, receiving a clinical diagnosis is currently an important step towards accessing treatment options, and although the children who received this intervention no longer fit a diagnosis based on the DSM-5, they still might need additional help and support as they grow.

Finally, as is the case with many diagnoses, it's vital to make sure that work in this area focuses on improving accessibility and understanding towards people living with the condition, rather than simply 'eradicating' a particular diagnosis or disability.

"Autism is not a disease and not something that should be cured or lessened, so how this study assessed the impact of intervention on 'autism behavior severity' may cause concern amongst many autistic people and their families," says Tim Nicholls, a spokesperson from the UK's National Autistic Society.

"There are some strong technical points in this research, but there will be questions about its general premise … It's important that any further study into very early intervention doesn't seek to lessen 'severity' – early intervention should be about supporting autistic people with the biggest challenges they face.

"For effective research to be done in this area in the future, autistic people must be involved in every stage."

The research has been published in JAMA Pediatrics.