A revolution for the at-risk

Around 1 per cent of UK children have autism spectrum disorder (ASD), a condition that fundamentally affects their ability to understand other people. Such children struggle to communicate with others, can have difficulty with change, and may be overwhelmed by new sights and sounds. Many adults with ASD experience a reduced quality of life. Financial costs are also high: supporting a person with ASD across their lifespan is estimated to cost more than £1 million (Buescher et al., 2014).

In this article we will argue that the time is right for a significant increase in investment in early intervention for children with ASD and other neurodevelopmental disorders. Using ASD as an example, we will illustrate how recent research identifies revolutionary new avenues for developing and targeting interventions in early development. We will also highlight how this is applicable beyond ASD by discussing the example of another common childhood-onset disorder, attention deficit hyperactivity disorder (ADHD). We suggest that new approaches may transform current debates on the ethics of early screening and early intervention. Finally, we consider how such approaches may narrow the gap between research and practice. Taken together, we believe that we are poised to make transformational changes in detection and treatment for early emerging neurodevelopmental disorders.

Early intervention and early identification
Early intervention offers the greatest potential for optimal outcomes for children with ASD. In a groundbreaking 2014 study, Pickles and colleagues studied language development in 192 children with autism followed longitudinally from age two to age 19. Between age two and six years, there was substantial variability in language trajectories. Some children made substantial gains and ended with language in the typical range, whilst others remained significantly delayed. After age six, trajectories remained remarkably stable such that children with poor language skills at the age of six still had poor language skills 13 years later. These data indicate that the effects of a supportive environment may be maximal in the first years of life, providing a powerful illustration of early neurodevelopmental plasticity.

Randomised controlled trials have indeed shown that interventions are more successful when started at a younger age (e.g. Dawson et al., 2010; Green et al., 2010). Early intervention is economically beneficial: a 2012 Dutch study led by Peters-Scheffer estimated the potential lifetime savings as 1.1 million euros per person. However, current intervention models are intensive (often 20 to 40 hours per week), placing a significant burden on individuals and families. A further major challenge is that access to existing interventions typically requires early diagnosis. Jeremy Parr and colleagues recently showed that the average age of diagnosis in the UK has remained stable at 55 months for the last decade (Brett et al., 2016). Even within children diagnosed under age three, the average age of diagnosis was 30 months. Since parents first show concerns at 10 to 16 months (Herlihy et al., 2013), this diagnostic gap is a substantial challenge to the provision of early intervention for children with emerging ASD.

We propose that there is a common solution to the twin challenges of developing better early identification and intervention approaches. Traditional approaches to mental health focus on identifying and targeting the surface symptoms that are used in diagnostic classification systems. Diagnosis is often required before treatment can commence, because the 'disorder' has to be identified in order to provide relevant treatments. We contend that we require a revolution in this approach to mental health conditions. Instead of focusing on surface features of the condition, we should be targeting the neurodevelopmental mechanisms that produce troubling symptoms in early development (E.J. Jones et al., 2014).

This approach is comparable to the prescription of statins for those at risk of heart disease, a drastic change in the management of this condition. Such a mechanistic approach would allow infants at heightened risk for particular symptom clusters to be identified prior to emergence of a recognisable clinical syndrome. Intervention could be provided based on the presence of the mechanism, and need not wait for clinical diagnosis. This would significantly reduce the troubling delays experienced by children in accessing intervention services. Early mechanistic interventions may in the long-term ameliorate or even prevent the emergence of troubling symptoms (e.g. lack of language), whilst leaving potential strengths (such as creativity or memory) untouched. Finally, mechanistic approaches are not limited to particular diagnostic categories and may more faithfully 'carve nature at its joints'. For example, in the latter part of the article, we discuss how attention difficulties may be relevant to both ASD and ADHD risk in early development (Johnson et al., 2015). These revolutionary changes will be made possible through a radical new approach to the study of neurodevelopmental disorders: prospective longitudinal studies of infants at heightened risk.

Paths to autism: sibs studies
In 2005 Dr Lonnie Zwaigenbaum and colleagues published a seminal study of infants with older siblings with ASD. Because ASD runs in families, about 20 per cent of such infants are diagnosed with autism by their third birthday (Ozonoff et al., 2011). For the first time, researchers could study the emergence of ASD in real time. Dr Zwaigenbaum's team showed that babies diagnosed with ASD at 24 months showed subtle developmental problems by 12 months of age. These included unusual eye contact, poor imitation, poor visual tracking, lack of smiling and laughter, and being slow to shift attention between two toys. This groundbreaking study has inspired more than a decade of 'baby sib' research that in turn has revolutionised our understanding of the earliest signs and symptoms of autism.

Baby sibs research has shown that by the second year of life, clear behavioural warning signs emerge in infants with later autism. These include failure to respond to name, poor eye contact and slowed language development. Any loss of skills such as walking or talking is of substantial concern.

These 'red flags' are now widely publicised by charities and other organisations. But what mechanisms underlie these early symptoms? In infants under 12 months, there are few clear behavioural signs of autism that could be used to identify individual children at risk. However, there are subtle differences between groups of infants with later autism and those who develop typically. For example, at six months infants with later autism often struggle to hold their head steady when pulled to sit (Flanagan et al., 2012), and other early motor delays have been observed when large groups are studied (Estes et al., 2015). Brain growth may also be subtly different, with faster expansion of head circumference and brain size in the first year (Shen et al., 2013). These changes suggest that broad changes in brain development precede the emergence of specific autism symptoms.

Contrary to expectations, researchers have identified very few changes in overt social behaviour in young infants with later autism. For example, in 2010 Ozonoff's team showed that infants with later autism look at people just as much as typically developing infants in the first year of life. However, developmental trajectories may be critical in detecting changes that are not apparent at a single time-point. For example, W. Jones and Klin's 2013 paper in Nature suggested that infant boys with later autism show declining patterns of gaze to eyes between two and six months that can be detected with eye-tracking technology. There may also be differences in how the infant's brain is responding to the incoming socialinformation. Using EEG (see also 'New studies…' box), Mayada Elsabbagh and colleagues (2012) showed that six-month-old infants with later ASD show a reduced ability to detect changes in eye gaze direction. Neural responses to faces are also slower and less prolonged in six-month-old infants with later ASD (E.J. Jones et al., 2016), suggesting reduced engagement of attention to social stimuli. As a group, infants at heightened risk for ASD also show markedly reduced social brain activity in response to social videos (Lloyd-Fox et al., 2013. Taken together, these results suggest an alteration of social brain specialisation in the early development of infants with ASD.

The implications of this for early intervention are clear. The infant brain becomes socially specialised through a complex interaction between innate programming and experience of the early environment. If early social brain development is altered in ASD, interventions that support the early social environment may be powerful. In 2015 Green and colleagues reported the results of the first randomised controlled trial of a parent-mediated intervention for infants with older siblings with ASD. The 12-week intervention helped to enrich the child's social environment by teaching parents to boost their responsivity to their infant's bids for attention. Results showed that at 14 months, infants who received the treatment tended to show increased attentiveness to their parent. Promising effects were found on other potential early markers for later ASD, such as better attention shifting between two objects on a screen. An independent study using a similar intervention approach identified significant improvements in neurocognitive markers of social attention (E.J. Jones et al. 2016). The small size of both studies means that further work is needed, but these results are highly promising in suggesting that relatively low-cost interventions could be used to target the mechanisms that underpin symptom emergence in infants with ASD.

Moving beyond autism
Mechanistic approaches to early intervention may not be restricted to particular diagnostically defined disorders. For example, a wide range of infants can show early social communication vulnerabilities and thus may benefit from early intervention that could support their development. In a large study of typically developing infants, parents with higher levels of social anxiety had infants who showed poorer social attention on a range of neurocognitive measures similar to those used with infants at risk for ASD (E.J. Jones et al., 2016). Long-term follow-up will indicate whether these infants (who are currently developing typically) are more vulnerable to shyness or social anxiety in later development, and whether they may benefit from brief parent-mediated interventions designed to support their social engagement. Such low-cost interventions may have broad positive impacts for children across the spectrum of social difficulties.

Attention is another critical neurocognitive domain in infancy. Attention difficulties are common in children with ASD, but are also a diagnostic feature of ADHD. Research groups led by Angelica Ronald (2008)  and Nanda Rommelse (2010, 2011) have provided several strands of evidence that attention difficulties in the two conditions may share developmental roots:
I    the two conditions have substantial overlap in genetic risk factors;
I    ASD and ADHD commonly co-occur within individuals and their families;
I    patterns of performance on many neurocognitive tasks are similar in the two conditions; and
I    poor attention skills in infancy (such as difficulty sustaining attention) are apparent prior to both ASD and ADHD diagnosis.

Early alterations in attention may thus be a common treatment target for infants at risk of ASD and ADHD. To test this hypothesis, we are currently conducting the first large longitudinal study of infants with older siblings with ASD and/or ADHD (see 'New studies…'). We will examine attention and other domains in very early development to identify distinct and similar causal paths.

We are currently testing new interventions previously demonstrated to improve attention in low-risk young infants. In 2011 Wass and colleagues reported that attentional control (the ability to move attention at will) can be improved in typically developing infants by playing a series of innovative gaze-contingent games over a short period. In these games, infants watch objects on a screen and can control them by moving their gaze. This can now be achieved with relatively low-cost eye-tracking systems that use infrared light to detect where an infant is looking on the screen. We are currently using these games with infants at high familial risk for ADHD to test whether helping infants to improve their attentional control skills provides significant benefits for learning and development (www.staars.org/interstaars). In an exciting new collaboration, we are also working with Zwaigenbaum and colleagues at the University of Alberta to test whether this intervention is also beneficial for infants at risk for ASD.

Ethics of early intervention
Although early detection and intervention can be effective, concerns remain about widespread implementation of screening and treatment programmes (see 'Should we screen for autism?'). Overdiagnosis is a concern, particularly for ADHD where reports indicate far higher rates of diagnosis and prescription amongst children who are young for their school year (Zoëga et al., 2012). Such children may be judged to have difficulty with attention and concentration skills because of their relative immaturity, not because they have a neurodevelopmental disorder. Developing more objective tools that do not rely on subjective comparisons made by teachers or parents may be one way to tackle this issue.

The autism and ADHD communities also stress the need to consider whether intervention is desirable. The 'neurodiversity' movement argues that neurodevelopmental 'disorders' like ASD and ADHD should instead be seen as being on the spectrum of individual differences; applying a disease model to these conditions is inappropriate. The neurodiversity movement is sometimes misrepresented as being against any form of treatment – rather, the goal is generally to move away from 'curing' and towards options that might enable individuals with ASD or ADHD to reach their full potential. Not all individuals with ASD or ADHD will want to access treatment or intervention options, since not all individuals will feel that they have difficulties with which they need help. However, the needs of those individuals (particularly with ASD) who cannot communicate and thus cannot contribute to debates in this area must also be considered. Many individuals with ASD or ADHD have significant strengths, like artistic ability, creativity, detail-orientation or skill with computers. If mechanisms that produce challenging symptoms can be disentangled from those that produce strengths, we may be able to develop more targeted treatment options.

This question is particularly problematic when applied to intervention in the early years, because infants and young children cannot themselves choose whether to receive it. Led by Sue Fletcher-Watson at the University of Edinburgh, we recently asked 2317 stakeholders across Europe about their views on early autism research. Respondents included parents of children with autism, clinicians and autistic adults. Whilst respondents were generally very positive about early autism research, adults with autism were less likely to want to prioritise research on early diagnosis than other groups. However, adults with autism were significantly more likely to prioritise intervention designs. We need interventions targeted to the mechanisms that produce unwelcome symptoms, rather than efforts to diagnose and treat the full syndrome at younger ages. Moving forward, it is critical to engage adults with autism and their family members in the research design process to ensure that all views are represented when designing intervention studies.  

Mind the gap
Translation of new research findings to improvements for service users remains a significant issue across child psychiatry. Research on early autism and ADHD is in its infancy, and substantial progress is required before some of the newest findings can be translated into practice. For example, many 'biomarkers' for ASD actually represent group differences, and are not individually predictive. Although this is a challenge, predicting ASD as a diagnostic category with high accuracy is not the goal. Rather, identifying markers of symptoms of ASD that may be particularly problematic (such as social communication problems, or sensory sensitivities) is critical. Further, markers for screening are usually judged by their sensitivity (the percentage of children  identified with the marker) and their positive predictive value (of the children with the marker, the proportion later diagnosed with the condition). However, markers for mechanisms that may be sensitive to intervention may actually have a poor positive predictive value to later diagnosis, because the child's environment between assessment of the marker and eventual diagnosis would be expected to have a relatively greater effect. Such considerations are important and under-discussed in the field.

Reproducibility is a critical challenge. There have been very few replication studies of neurocognitive markers of later ASD to date. Such efforts are under way – with a team of investigators we are currently running a multi-site study of infants with older siblings with ASD across Europe (www.eurosibs.eu). This study will attempt to replicate several key findings from the baby sib literature. Generalisability is also very important. For example, we recently showed that some early 'markers' for later ASD may only be related to later autism symptoms in boys and not girls (Bedford et al., 2016). In addition, findings from baby sibs research will need to be replicated in other populations. We are currently running such studies with infants with known genetic conditions linked to ASD and ADHD, such as tuberous sclerosis; other work should identify infants with early behavioural signs and examine whether neurocognitive markers could enhance individual prediction.

Despite the challenges, new mechanistic interventions hold significant translational potential. Parent-mediated interventions that appear efficacious in baby sibs (Green et al., 2015) are based on existing programmes that are low-cost, manualised and have been used in other populations in the community. Once sufficient evidence of their efficacy in the short and long term accumulates, roll-out would be more straightforward. Other new interventions such as gaze-controlled eye-tracking programmes rely on equipment that is becoming significantly cheaper. In the medium term, such training programmes could be operated remotely by parents, with less need for clinician input. Such advances improve the potential accessibility of interventions, and lower the bar in terms of the cost–benefit ratio of intervention provision.

Transforming the outlook
Prospective longitudinal studies of infants at heightened risk of neurodevelopmental disorders provide the potential for developing new interventions that are targeted at the mechanisms that underlie symptom emergence. There is much work to do in improving the quality and replicability of early indicators, and testing new intervention approaches in rigorously controlled trials. However, this new mechanistic approach has significant promise to overcome some of the ethical and translational obstacles to the provision of early intervention to vulnerable children. These advances could therefore transform the outlook for infants at heightened risk for conditions like ASD and ADHD. The resources that need to be devoted to these efforts are not trivial, but the potential economic, societal and personal benefits vastly outweigh the possible costs.

Box text: New studies of early ASD and ADHD

In the UK, the BASIS study (British Autism Study of Infant Siblings) is a UK-wide network dedicated to the study of infants with older siblings with ASD. The BASIS team, led by Professor Mark Johnson at Birkbeck College London and Professor Tony Charman at King's College London, have recently launched STAARS (Study of Attention and ADHD Risk in Siblings), which will follow both infants with older siblings with ASD and infants with older siblings with ADHD in the same protocol.

Infants are studied at 5, 10, 14, 24 and 36 months. Methods used include eye-tracking, electroencephalography (EEG) and near infrared spectroscopy (NIRS), both noninvasive measures of brain activity; eye-tracking, to assess what infants attend to; and measures of behaviour, cognition and arousal.

Following both groups of infants in the same protocol will allow us to compare and contrast the early developmental paths to the two disorders. We will be able to ask whether there may be similar or different early markers for ASD and ADHD, and whether there may be core paths that could be targeted by prodromal interventions.

Further information can be found on our website: www.staars.org.uk

Box text: Should we screen for autism?

In 2007 the American Academy of Pediatrics recommended continuous developmental surveillance and specific autism screening at 18 months, 24 months and whenever a parent or provider expresses concern. These guidelines were based on the growing understanding of early red flags for the condition (for example, see www.cdc.gov/ncbddd/autis/signs.html).Recent reports (e.g. tinyurl.com/hmuagux) suggest that this recommendation has significantly reduced age of diagnosis in the US.

However, in February 2016 the US Preventative Services Task Force decided not to recommend universal screening in the US on the basis that there is insufficient evidence of benefit. The Task Force accepted that common screening tools (like the Modified Checklist for Autism in Toddlers questionnaire) were effective. However, they noted that work on early intervention has been conducted with clinically referred rather than screen-positive populations, and so called for more research following children from screening to diagnosis and treatment.

This judgement echoes the findings of the UK National Screening Committee, who in 2012 concluded that there was no justification for universal screening for ASD. The UK panel also argued that there was a need for greater information about the long-term benefits of early intervention before the value of screening could be determined. Whilst the need for more research is widely accepted by the field, many prominent clinicians and researchers disagree with the Task Force's approach. Autism Speaks and other prominent charities have argued that the risk to benefit ratio 'strongly favours universal screening for autism' (tinyurl.com/hulrbp8). The debate on screening for autism requires full consideration of the scientific, ethical, economic and societal dimensions, and the need for further research is clear.

- Emily J.H. Jones is a Lecturer at the Centre for Brain and Cognitive Development, Birkbeck
[email protected]

- Mark H. Johnson is a Professor of Psychology and Director of the Centre for Brain and Cognitive Development, Birkbeck
[email protected]

References

Bedford, R., Jones, E.J.H., Johnson, M.H. et al. (2016). Sex differences in the association between infant markers and later autistic traits. Molecular Autism, 7, 21.
Brett, D., Warnell, F., McConachie, H. & Parr, J.R. (2016). Factors affecting age at ASD diagnosis in UK. Journal of Autism and Developmental Disorders, 46(6), 1974–1984.
Buescher, A.V., Cidav, Z., Knapp, M. & Mandell, D.S. (2014). Costs of autism spectrum disorders in the United Kingdom and the United States. JAMA Pediatrics, 168(8), 721–728.
Dawson, G., Rogers, S., Munson, J. et al. (2010). Randomized, controlled trial of an intervention for toddlers with autism. Pediatrics, 125(1), e17–23.
Elsabbagh, M., Mercure, E., Hudry, K. et al. (2012). Infant neural sensitivity to dynamic eye gaze is associated with later emerging autism. Current Biology, 22(4), 338–342.
Estes, A., Zwaigenbaum, L., Gu H. et al. (2015). Behavioral, cognitive, and adaptive development in infants with autism spectrum disorder in the first 2 years of life. Journal of Neurodevelopmental Disorders, 7(1), 1–10.
Flanagan, J.E., Landa, R., Bhat, A. & Bauman, M. (2012). Head lag in infants at risk for autism. American Journal of Occupational Therapy, 66(5), 577–585.
Fletcher-Watson, S., Apicella, F., Auyeung, B. et al. (2016). Attitudes of the autism community to early autism research. Autism, March.
Green, J., Charman, T., McConachie, H. et al. (2010). Parent-mediated communication-focused treatment in children with autism (PACT): A randomised controlled trial. The Lancet, 375(9732), 2152–2160.
Green, J., Charman, T., Pickles, A. et al. (2015). Parent-mediated intervention versus no intervention for infants at high risk of autism. The Lancet Psychiatry, 2(2), 133–140.
Herlihy, L., Knoch, K., Vibert, B. & Fein, D. (2013). Parents' first concerns about toddlers with autism spectrum disorder: Effect of sibling status. Autism. November.
Johnson, M.H., Gliga, T., Jones, E. & Charman, T. (2015). Annual research review. Journal of Child Psychology and Psychiatry, 56, 228–247.
Jones, E.J. (2016). Early intervention in infants at risk for ASD. Manuscript submitted for publication.
Jones, E.J., Gliga, T., Bedford, R. et al. (2014). Developmental pathways to autism. Neuroscience & Biobehavioral Reviews, 39, 1–33.
Jones, E.J.H., Venema, K., Earl, R. et al. (2016). Reduced engagement with social stimuli in 6-month-old infants with later autism spectrum disorder. Journal of Neurodevelopmental Disorders, 8(1), 1.
Jones, W. & Klin, A. (2013). Attention to eyes is present but in decline in 2–6-month-old infants later diagnosed with autism. Nature, 504(7480), 427–431.
Lloyd-Fox, S., Blasi, A., Elwell, C.E. et al. (2013). Reduced neural sensitivity to social stimuli in infants at risk for autism. Proceedings of the Royal Society B: Biological Sciences, 280(1758).
Ozonoff, S., Iosif, A-M., Baguio, F. et al. (2010). A prospective study of the emergence of early behavioral signs of autism. Journal of the American Academy of Child and Adolescent Psychiatry, 49(3), 256–266.
Ozonoff, S., Young, G.S., Carter, A. et al. (2011). Recurrence risk for autism spectrum disorders. Pediatrics, 128(3), e488–e495.
Peters-Scheffer, N., Didden, R., Korzilius, H. & Matson, J. (2012). Cost comparison of early intensive behavioral intervention and treatment as usual for children with autism spectrum disorder in The Netherlands. Research in Developmental Disabilities, 33(6), 1763–1772.
Pickles, A., Anderson, D.K. & Lord, C. (2014). Heterogeneity and plasticity in the development of language. Journal of Child Psychology and Psychiatry, 55(12), 1354–1362.
Ronald, A., Simonoff, E., Kuntsi, J. et al. (2008). Evidence for overlapping genetic influences on autistic and ADHD behaviours in a community twin sample. Journal of Child Psychology & Psychiatry, 49(5), 535–542.
Rommelse, N.N.J., Franke, B., Geurts, H.M., et al. (2010). Shared heritability of attention-deficit/hyperactivity disorder and autism spectrum disorder. European Child & Adolescent Psychiatry, 19(3), 281–295.
Rommelse, N.N.J., Geurts, H.M., Franke, B. et al. (2011). A review on cognitive and brain endophenotypes that may be common in ASD and ADHD and facilitate the search for pleiotropic genes. Neuroscience & Biobehavioral Reviews, 35(6), 1363–1396.
Russell, G., Rodgers, L.R., Ukoumunne, O.C. & Ford, T. (2014). Prevalence of parent-reported ASD and ADHD in the UK. Journal of Autism and Developmental Disorders, 44(1), 31–40.
Shen, M.D., Nordahl, C.W., Young, G.S. et al. (2013). Early brain enlargement and elevated extra-axial fluid in infants who develop autism spectrum disorder. Brain, July.
Wass, S., Porayska-Pomsta, K. & Johnson, M.H. (2011). Training attentional control in infancy. Current Biology, 21(18), 1543–1547.
Zoëga, H., Valdimarsdóttir, U.A. & Hernández-Díaz, S. (2012). Age, academic performance, and stimulant prescribing for ADHD. Pediatrics, 130(6), 1012–1018.
Zwaigenbaum, L., Bryson, S., Rogers, T. et al. (2005). Behavioral manifestations of autism in the first year of life. International Journal of Developmental Neuroscience, 23, 143–152.