Crosstalk Debates and conversations about timely topics in autism.
Opinion / Cross Talk

How early should autism treatment begin?

by  /  15 October 2014

Editor note: On Wednesday, 22 October, at 1 p.m. Eastern, we hosted a live Twitter Q & A on early intervention in autism, inspired in part by the perspectives below. The chat featured leading scientists who are designing and testing early intervention programs. The open dialogue also included families and journalists. You can find a ‘Storify’ summary of the conversation here.

In September, we described results from a small pilot study that assessed a behavioral interaction delivered by parents to infants with early signs of autism. Five of the seven high-risk children who received the intervention did not go on to develop autism by age 3.

The study, by Sally Rogers and her colleagues at the University of California, Davis MIND Institute, suggests that this low-intensity therapy may help the infants improve their behavior and language skills.

The preliminary success of these sorts of therapies raises a host of broader questions about the many caveats and challenges of treating autism early.

How early can an intervention for autism reasonably begin? Is there a critical period? Why do the interventions work well with some children and fail in so many others? What are the ‘active ingredients’ of a successful therapy?

For this installment of Cross Talk, we asked several leading scientists who are designing and testing early intervention programs for children with autism.

What do you think? Share your reactions and follow-up questions in the comments section below.

The Experts:

Connie Kasari

Professor of Human Development and Psychology, University of California, Los Angeles

Stephen Camarata

Professor, Vanderbilt Kennedy Center at Vanderbilt University

David Mandell

Associate Professor, Perelman School of Medicine at University of Pennsylvania

Aubyn Stahmer

Associate Professor, Department of Psychiatry at UC San Diego

Lynn Koegel

Clinical Director, University of California, Santa Barbara

Jennifer Stapel-Wax

Associate Professor, Department of Pediatrics at Emory University School of Medicine

Study supports early parent-mediated intervention

Intervene on all positive screens: “The preliminary data provided in the new paper explores the potential of intervening for early signs suggestive of autism risk in young infants. This uncontrolled study of a small sample of infants precludes definitive interpretation. As the authors note, and I concur, we must exercise caution in interpreting these results because we do not know the level of autism risk pre-intervention and we cannot confidently ascribe the positive outcomes to the intervention alone.

“Given the large variability in test scores at age 3, continued follow-up of these infants may provide further insight into their developmental trajectories. Nevertheless, the study provides support for the idea that early parent-mediated interventions may be beneficial to infants showing developmental signs of autism risk.”

Consolidated evidence: “Last year, Jonathan Green and his colleagues at the University of Manchester tested a video-aided, parent-mediated intervention for seven 8- to 10-month-old infant siblings of an older child with autism. They too found positive effects in infants whose parents used a number of strategies aimed at improving parent responsiveness and infant attention and engagement. Thus, early interventions incorporating the ‘active ingredients’ of those described by Green and Rogers may well emerge as best practices for any child who screens with developmental concerns.”

Critical periods misunderstood

Professor of Hearing and Speech Sciences, Vanderbilt University
Parent power: “The key active ingredient in this and most other early intervention approaches is teaching parents to better identify their child’s early attempts at communication and social engagement. Parents can then reinforce and increase social and communication behavior in their infant or toddler. In a nutshell, the new study shows that enrichment of this nature from parents is associated with growth in social and communication skills in the children receiving this intervention. However, the study does not conclusively show that early intervention for autism is effective. In that regard, the limitations in the design and the lack of diagnostic specificity for autism in infants are problematic.”
Identification the key: “With the proper control and comparison groups, early intervention research can be completed at any age, including newborn infants. The problem lies in accurately identifying autism in toddlers and infants. There are some promising markers, but accurate and stable long-term diagnostic methods are still being developed. This is important because evidence-based early intervention requires accurate and stable diagnoses in order to ensure that the intervention is actually effective over and above spontaneous recovery.”
Critical confusion? “There is probably nothing more misunderstood and perhaps misrepresented than the notion of ‘critical periods’ in early intervention. Animal and human studies indicate that there are indeed critical periods for neural development. However, the sensory input needed to activate neural development is actually rather global and nonspecific. For example, it has long been known that sensory input from the eyes is necessary for activating neural organization of the visual cortex. Studies of animals and of clinical populations have shown that even relatively poor visual input is sufficient to provide the basic neural architecture needed for processing visual information. Pawan Sinha’s research at the Massachusetts Institute of Technology is particularly noteworthy in this regard.”
Earlier still better: “On the other hand, because learning is incremental and developmental, one can conceptualize ‘critical periods’ as when infants and toddlers systematically build their language and social skills, which lay the foundation for further complex social and linguistic development. A cessation or plateau in this learning likely has a long-term impact on the infant or toddler’s ability to acquire ever more sophisticated language and social skills.
“In essence, the earlier an ‘at-risk’ infant or toddler can be taught to attend to and learn from parents and others in the ambient environment, the more opportunities there will be to learn from future ongoing social interactions.”

Traditional medical model too late

Director, Center for Mental Health Policy and Services Research, University of Pennsylvania
Prevention policy needed: “Despite the limitations acknowledged by the authors, this is an exciting study on many fronts. First, it lends further evidence to the idea that autism comprises a trajectory of brain processes that deviate from typical development early in life. If we intervene early enough, we can change this trajectory. The jury is still out on how early is early enough, but it is likely to be before we can accurately diagnose autism, especially outside of expert autism clinics. If we stick to the traditional medical model of ‘screen, diagnose, refer and treat,’ we will intervene too late.
“Instead, we should examine the intervention components here and determine whether we could apply a public health model of disease prevention and health promotion. In other words, can we prevent autism? Rogers and her colleagues describe a treatment with specific responses to six target symptoms that, together, qualify children for a diagnosis of autism. We could consider screening for these symptoms during 6- and 9-month ‘well-child visits,’ and teaching parents these basic procedures.”
Parental overload: “It will be important to study the independent effects of these intervention components. Determining the minimal intervention necessary is critical. We may be asking parents to do too much even with these brief interventions. Many pilot studies like this one have shown the promise of parent-implemented interventions, but later randomized trials often are disappointing. While we have become increasingly adept at changing the behavior of children with autism, we are less adept at changing the behavior of their parents. If we wish to expand the population to which we provide intervention, developing easy-to-implement interventions, combined with effective parent training and support, will be an even more critical issue.”

Excited for larger trial

Associate Director, Child and Adolescent Services Research Center; Associate Professor of Psychiatry, University of California, San Diego

Early promise: “Coaching parents in specific techniques to increase parent-child engagement has the potential to ameliorate the withdrawal often seen in this population during the early toddler years. Although there is growing evidence that these types of interventions support early development in children with autism, the new study is one of the first to demonstrate positive results in infants this young. It seems that social engagement was affected in this group of children, which has been a core feature of autism that is traditionally quite resistant to intervention. I am excited to see what new information a larger trial of the intervention will find.”
Natural ingredients: “Several researchers are trying to extend these evidence-based behavioral interventions set in natural or home environments to younger age groups. These studies will provide an opportunity to examine the ‘active ingredients’ that are key to positive outcomes across various successful interventions. It is possible, indeed likely, that children at risk for autism are differentially responsive to certain components.”
Tailored treatment: “Understanding the characteristics that respond best to specific ingredients may help researchers and providers develop individualized interventions that will optimally support an individual child’s progress. This may facilitate progress in children with autism who respond more slowly to our current interventions. We are not at that point. Still, the studies in which complex interventions are simplified for younger populations are beginning to provide clues.”

How early can we intervene?

Clinical Director, Koegel Autism Center; Director, Eli and Edythe L. Broad Center for Asperger Research, University of California, Santa Barbara
Progress begets progress: “The question arises as to whether a diagnosis of autism is possible before the onset of verbal communication if verbal communication itself is a diagnostic criterion. The important thing to consider is that if we can intervene within the first year of life and decrease the early social deficits that are evident, we may, consequently, decrease the likelihood of a more profound disability.”
Early signs: “Parents, especially those who have other children, are usually the first to express concern about their child’s lack of social development. While there is a wide variability within and across children during the first year of life, we often find a more consistent pattern of difficulties with social engagement in at-risk babies. The research is beginning to suggest that these early social concerns may be improved with intervention. Instead of needing large numbers of weekly hours, infants often improve with just an hour or two a week of parent education. Thus, an early intervention is far more cost- and time-efficient than a ‘wait-and-see’ approach.”

Early intervention works, and changing

Director, Infant and Toddler Clinical Research Operations, Marcus Autism Center, Children’s Healthcare of Atlanta; Associate Professor of Pediatrics, Emory University
Making most of intervention: “This pilot study represents the beginning of a new generation of studies focused on treatment in infancy, while capitalizing on the period of greatest postnatal neuroplasticity.
“Undoubtedly, this study is preliminary and requires further research and exploration. Other labs are already performing similar studies that utilize parent-implemented models focused on making the most of everyday moments. Coaching the parents must involve both empowering them with knowledge of development, as well as helping them to integrate strategies throughout their day to promote engagement and development of new skills.”
Universal design: “Early intervention works, and we need to start as soon as possible in the life of a young child. Our definition of intervention is changing from yesterday’s models of clinic-based, expensive, primarily clinician-directed treatment. Instead, we favor naturalistic settings with clinically coached caregivers using models that are uniquely generalizable to community-based, home-based and population-focused efforts.
“Increasingly, we can expect to see interventions that universally produce better outcomes in young children who are at risk for autism or other neurodevelopmental disorders. This universal design will help ensure that all children get the rich and meaningful interactions and opportunities that are integral to their healthy neurodevelopment.”

About Cross Talk
Discussions among leading experts in the field. Submit your questions to

8 responses to “How early should autism treatment begin?”

  1. charlesnelson says:

    Despite keen interest in developing drugs that can be used to treat autism, a pharmacological approach to a disorder so complicated and heterogeneous is still not within reach. It is therefore timely that Rogers and colleagues have attempted to adapt a well-established (although still not widely used) behavioral treatment for use in infants at high risk for developing autism – specifically, those with an older sibling with the disorder. The great promise of this approach is that it can be applied *before* signs or symptoms of the disorder first become apparent, and it comes with no side effects. Since it is well known that early intervention is more efficacious than later intervention, the beauty of this approach is that in some cases it may actually prevent the disorder from developing. Of course, this approach begs the question of how early the intervention can be implemented. One approach to this is to increase the developmental sensitivity of the treatment protocol so that it can be used in infants as young as perhaps 4-6 months, when infants are becoming increasingly social, more sophisticated motorically and the first signs of language (cooing and then babbling; some rudimentary gestures) are coming on line. Whether all children will benefit from this remains a big unknown, for the very simple reason that only approximately 1 in 5 infants with an older sibling will eventually be diagnosed with an ASD. And, among those who will develop autism, we still have no way of predicting individual differences in treatment response; such individual differences could have to do with underlying genetic differences or in neural circuitry. An additional consideration is whether there is a tipping point in the trajectory to autism. If there is, then we must determine what the critical window is for applying the intervention.

  2. usethebrainsgodgiveyou says:

    “…all children get the rich and meaningful interactions and opportunities that are integral to their healthy neurodevelopment.” ___This sounds more like Son-Rise than ABA,___which gives me hope that science is losing the “othering” aspect of the poor kids, making it more about relationships. What happened to table ready and eye-contact, the FIRST tenants of behaviorism when my son was young? I ignored both of them, thought they were inappropriate based on mother’s intuition, not science. Turns out, most adult autistics who went through ABA would agree.

    Funny…ABA guy versus Son-Rise guy..the top video, if you want to know what I knew 17 years ago. I beg you to use common sense in helping children enter a social world. (Click on top video, 10 ABA versus Son-Rise “advertisements” will follow.) This is not science, only for amusement.

    As ABA moves more towards the acceptance and child-centered work that Barry Kauffman, father, tried years ago, the more parents and children seem to love it. I used it with my son, for 2 hours, not 8 hours a day, but still… Let’s hope the nightmare of ABA abuse ends. Teachers with common sense use ABA fairly restrictively, and look to the relationship first. As I was told by a teacher at a residential ABA school, “They have to know you love them, first.” It’s just common sense, and not fear based. Good luck to new parents who will see a better day.

    • Jen says:

      I’m surprised to hear this ABA school you mention is even clued into the fact to any kind of social interaction/relationship. Most I’ve seen are rote, strict ABA where the kids make little gains and NO generalization of skills.

  3. Uta Frith says:

    While it is obvious that early intervention is highly desirable, I am concerned that our enthusiasm may make us forget two things:
    1. Learning and teaching are life long activities. Early intervention does not necessarily mean that later intervention will no longer be necessary. In many, perhaps most cases, there is a need for sustained intervention.
    2. If you missed the early start, don’t despair. Intervention can happen at any age. We all continue to learn during our whole life span.

  4. Damian Milton says:

    What nonsense – “If we intervene early enough, we can change this trajectory.” People mature and change autistic or not. The amount people think that it is what they are doing which is affecting someone else is in some areas massively exaggerated (and in other areas missed – like doing harm!). Early identification, understanding and support is desirable, not trying to normalise and cure autistic people of their natural dispositions and developmental trajectories. Learning is a lifelong activity – but if learning is how to internalise ableist ideology, we can do without it thankyou.

  5. Tracy says:

    From experiences, we introvened very early, but it only caused more problems in the school system because we couldn’t change his asperger’s personality. You know –talked and communicated very well when he wanted and if he was interested. Now I looked back and wondered how he has ever made to this point, because we all worked so hard to change who he was to be “normal” instead of accepting him. I have now realized that he is normal. And I see many other kids who are normal, but different. We need to educate people on these new “normals.” We need to shop trying to change them, but help them be accepted.

  6. bessie says:

    I think that a big part of the problem is lack of education. People are becoming very confused about the research information. Not everyone understands the terms. Professionals need to use clear straight to the point statements.You need to educate parents more about autism and what it really is (using clear language). Too many people are talking about their own private experiences and assuming that it is factual, or they are speaking about children with autism. Every child is an individual, and we must keep that in mind. Even though a child has autism it doesn’t mean everything he does is related to his autism. There are other factors in his life. And every child with autism is different.

  7. Seth Bittker says:

    Why is it that early intervention is restricted to behavioral therapy?

    It seems to me there is solid evidence in the literature for biochemical dysfunction and there have been controlled trials that certain biochemical therapies can help ameliorate this dysfunction. Why does somebody not conduct a trial of select early interventions based on biochemistry?

    For example take a population of at risk children. Then test for cysteine, oxidized glutathione and reduced glutathione. Take those with lower than typical control cysteine and higher than typical control oxidized to reduced glutathione. This group will form the population for the study. Next randomize assignment to two groups. Group 1 will get ABA only. Group 2 will get ABA, very low dose methylfolate and low dose NAC. Run the trial for X months. Then do behavioral tests. If Group 2 has significantly better behavior than Group 1, then you have evidence for a biochemical therapy in at risk children. One could do variants on this that are blinded.

Leave a Reply

Your email address will not be published. Required fields are marked *