Autism and Attachment: A Need for Conceptual Clarity

Avatar photo
Barry is an NIHR School of Primary Care doctoral student at Cambridge University. His research focuses on how practitioners make sense of developmental differences in children. Prior to moving to Cambridge, Barry worked as an Assistant Psychologist in the Health Service Executive in Ireland.

Posted on

It is not uncommon to hear clinicians and practice-focused researchers speak of an overlap in critical features between autism spectrum conditions (hereafter autism) and the various attachment disorders and patterns [1, 2]. The argument follows that there are times when it can be difficult to differentiate what is ‘autism-related’ from what is ‘attachment-related’. Much of the empirical work on this topic has focused on the differentials between autism and the ‘attachment disorders’ and in particular reactive attachment disorder. Yet some researchers and clinicians have suggested that the symptomatic similarities extend into the broader category of ‘attachment difficulties’. So what do we mean by ‘attachment disorders’? Is there a meaningful distinction between these disorders and ‘attachment difficulties’? And where is the overlap with autism?

What is autism?

Autism is defined as a lifelong developmental condition which is manifest through atypical social communication (e.g. atypical eye contact and difficulties with some aspects of social understanding/interaction) and behaviours deemed restricted and repetitive (e.g. hand-flapping, intense or circumscribed interests)[3] Autism features in the neurodevelopmental constellation and is considered, by most, to be present from birth and multifactorial. Establishing the biomedical causes has proved particularly complex, nevertheless, there are a robust suite of standardised assessments which have been shown to reliably measure autism. Although autism almost certainly has multiple causes, there is no known association with maltreatment or early adversity.

What are ‘attachment disorders’?

‘Attachment disorders’ are diagnoses which are sometimes applied to children who have experienced significant patterns of insufficient care. According to classification systems, there are two types of these conditions: Reactive Attachment Disorder (hereafter RAD) and Disinhibited Social Engagement Disorder (hereafter DSED) [3] Underpinning both of these conceptualisations is the idea that there has been a complete collapse of the attachment system as a result of extremely limited opportunities to forge a relationship with an attachment figure [4, 5]. Both RAD and DSED are characterised by atypical social responsiveness [3], although, they diverge in the nature of these social behaviours. Specifically, RAD is associated with ‘inhibited’ behaviours which mirror ‘internalising’ conditions, while DSED is associated with more ‘disinhibited’ behaviour which are closer to those seen in children with ‘externalising’ conditions.

To date there has been little agreement, among researchers, regarding the prevalence of attachment disorders. Some researchers maintain that they are rare even among high risk populations (e.g. children who have experienced institutional care)[6]. Whereas others argue that relatively high rates of attachment disorders can be observed in some samples even within the general population[7]. Yet these debates are part of a wider conversation regarding which symptoms are considered part of RAD [8]. For some, RAD is characterised by the absence of behaviours targeted toward an attachment figure, atypical comfort seeking/accepting behaviours in response to distress, atypical socio-emotional responsiveness, limited positive affect, and frightfulness [9, 10]. This understanding of RAD is closely tied with the descriptions contained in current nosologies [3]. In contrast, others have interpreted the diagnostic criteria regarding ‘minimal social responsiveness to others’[3] to include symptoms such as avoiding eye and physical contact, as well as difficulties with being affectionate. [11] This description of RAD is more aligned with the assessment tool The Child and Adolescent Psychiatric Assessment (CAPA-RAD) [12].

Symptomatic similarities between RAD and Autism

The thread of atypical social responsiveness and atypical positive affect runs through both autism and RAD. Previous work has observed inflated rates of autism-behaviours in children who have experienced institutional deprivation.[13] Specifically, these studies found that children who had experienced institutional deprivation had atypical social and communication behaviours comparable to that which is typically seen in children with autism. Crucially, however, there was often a marked change in these behaviours when the child entered a more adaptive environment [14], which is not typically the case with autism.

Subsequent work comparing children with autism and children with RAD found that a substantial number of children with RAD scored in the clinical range on the social domain for autism on gold standard assessments.[15] Diagnostic guidelines [3] acknowledge these symptomatic similarities and suggest that restricted and repetitive behaviour may be differential. Yet the authors of the aforementioned study [15] found that 20% of the children with RAD also demonstrated clinically significant levels of restricted and repetitive behaviours.

Unsurprisingly, the overlap becomes more acute when symptoms such as avoiding eye contact are included in the conceptualisation and assessment of RAD. For instance, Davidson and colleagues [16]  observed that children with autism can score in the clinical range on RAD assessments measures.

This does not, of course, suggest a possible epidemic of ‘misdiagnosis’. Rather, it serves as a reminder of the limitations of assessment tools, and underscores the importance of clinical judgment in making a diagnosis. Or to echo the words of Prof Catherine Lord at the recent ACAMH Jack Tizard Memorial Lecture and Conference, “The ADOS is a measurement, not an answer”.

Attachment difficulties

When thinking about the overlap with autism, some clinically-focused researchers have argued that the definitions of RAD and DAD are too restrictive [2, 17]. Instead, these  commentators suggest that the overlap extends to the wider category of ‘attachment difficulties’. The phrase ‘attachment difficulties’ is often used as a suitcase term for the attachment classifications (insecure and disorganised), as well the ‘attachment disorders’.

In accordance with this view, some practice-focused researchers have also started to develop tools which seek to differentiate between autism and attachment difficulties. Examples include The Coventry Grid [2] and The Coventry Grid Interview [17]. Both of these instruments place emphasis on the ‘emotional feel’ of symptoms and have been developed on the basis of clinical experience and clinical consensus. Many of the psychometric properties of these tools are yet to be established, however several NHS services do claim to use them. Importantly, the authors of each of these assessments all acknowledge that these instruments are not a straightforward guide to differentiation, but rather a possible complement to the information gathering process.

Whether these tools can accurately differentiate attachment difficulties from autism-related behaviours seems to be an open empirical question.


Perhaps the most important differential between autism and the attachment disorders is a history of insufficient care. In the absence of such a history it is unlikely that the child will meet the diagnostic criteria for either of the attachment disorders. Yet this is not such a clear cut indicator when thinking about attachment difficulties more generally, given that neither insecure nor disorganised attachment imply neglect or insufficient care [18].

There is also some emerging evidence that children with autism may be more likely to have an uneven cognitive profile, that is, a significant difference between performance and verbal IQ. [19] Such profiles are not typically reported in children with attachment difficulties in general or attachment disorders specifically. Obviously more research is required, but this is an intriguing line of inquiry.


The symptoms of autism and the various attachment conditions converge and depart in interesting ways. Yet the extent to which they overlap is closely tied to how the researcher or practitioner conceptualises ‘attachment disorders’ or indeed ‘attachment difficulties’. Standardised measures alone may not be a robust differentiator, thus underscoring the importance of clinical judgment and a greater focus on developmental history in the diagnostic workup. While certain tools might be helpful in the information gathering process, it is particularly important to consider how they conceptualise the conditions they are comparing.


  1. McKenzie, R. and R. Dallos, Autism and attachment difficulties: Overlap of symptoms, implications and innovative solutions. Clinical Child Psychology and Psychiatry, 2017. 22(4): p. 632-648.
  2. Moran, H., Clinical observations of the differences between children on the autism spectrum and those with attachment problems: The Coventry Grid. Good Autism Practice (GAP), 2010. 11(2): p. 46-59.
  3. Association, A.P., Diagnostic and Statistical Manual of Mental Disorders: Dsm-5. 2013: Amer Psychiatric Pub Incorporated.
  4. Woolgar, M. and S. Scott, The negative consequences of over-diagnosing attachment disorders in adopted children: the importance of comprehensive formulations. Clin Child Psychol Psychiatry, 2014. 19(3): p. 355-66.
  5. Zeanah, C.H., et al., Practice parameter for the assessment and treatment of children and adolescents with reactive attachment disorder and disinhibited social engagement disorder. Journal of the American Academy of Child & Adolescent Psychiatry, 2016. 55(11): p. 990-1003.
  6. Zeanah, C.H., et al., Reactive attachment disorder in maltreated toddlers. Child Abuse & Neglect, 2004. 28(8): p. 877-888.
  7. Minnis, H., et al., Prevalence of reactive attachment disorder in a deprived population. The British Journal of Psychiatry, 2013. 202(5): p. 342-346.
  8. Allen, B., Misperceptions of reactive attachment disorder persist: Poor methods and unsupported conclusions. Research in developmental disabilities, 2018. 77: p. 24-29.
  9. Zeanah, C.H. and M.M. Gleason, Annual Research Review: Attachment disorders in early childhood – clinical presentation, causes, correlates and treatment. Journal of child psychology and psychiatry, and allied disciplines, 2015. 56(3): p. 207-222.
  10. Rutter, M., J. Kreppner, and E. Sonuga-Barke, Emanuel miller lecture: Attachment insecurity, disinhibited attachment, and attachment disorders: where do research findings leave the concepts? Journal of Child Psychology and Psychiatry and Allied Disciplines, 2009. 50(5): p. 529-543.
  11. Pritchett, R., et al., Reactive attachment disorder in the general population: a hidden ESSENCE disorder. The Scientific World Journal, 2013. 2013.
  12. Minnis, H., et al., An exploratory study of the association between reactive attachment disorder and attachment narratives in early school‐age children. Journal of Child Psychology and Psychiatry, 2009. 50(8): p. 931-942.
  13. Rutter, M., et al., Quasi-autistic patterns following severe early global privation. The Journal of Child Psychology and Psychiatry and Allied Disciplines, 1999. 40(4): p. 537-549.
  14. Rutter, M., et al., Early adolescent outcomes of institutionally deprived and non‐deprived adoptees. III. Quasi‐autism. Journal of Child Psychology and Psychiatry, 2007. 48(12): p. 1200-1207.
  15. Sadiq, F.A., et al., Social use of language in children with reactive attachment disorder and autism spectrum disorders. European child & adolescent psychiatry, 2012. 21(5): p. 267-276.
  16. Davidson, C., et al., Social relationship difficulties in autism and reactive attachment disorder: Improving diagnostic validity through structured assessment. Research in Developmental Disabilities, 2015. 40: p. 63-72.
  17. Flackhill, C., et al., The Coventry Grid Interview (CGI): exploring autism and attachment difficulties. Good Autism Practice (GAP), 2017. 18(1): p. 62-80.
  18. Granqvist, P., et al., Disorganized attachment in infancy: a review of the phenomenon and its implications for clinicians and policy-makers. Attachment & Human Development, 2017. 19(6): p. 534-558.
  19. Melling, R. and N. Smethurst, Taking care with attachment disorders and autistic-like traits: the potential significance of cognitive markers. Educational Psychology in Practice, 2017. 33(3): p. 264-276.

This is an independent article and the views are not necessarily those of ACAMH.


The comparison is a category error. Attachment is a system all are ‘affected by’, where maladaptions to caregivers (or their absence) has been proscribed as ‘disorder’; autism is conceived of very differently.

Tools such as the Coventry Grid group together what people conceive of each concept: in simplistic terms, they chart what people believe – no grounding or triangulation.

Regrettably this piece does not improve clarity. The title leads us away from the critical clinical dilemma which does not concern attachment disorders and AuSpDis but attachment strategies and how some of these may be confused with AuSpDis – especially Asperger variant. Attachment strategies concern us all and are not to be conceptualised as difficulties. The higher numbered strategies, for example A5-8, and C5-8, as well as A/C have proved advantageous to maximise care in the individual’s early developmental niche. And who would like to be using a Type B strategy in Yemen. They are adaptive. With a focus specifically on the higher Type A strategies we note that those using these strategies tend to give few affective signals in social contexts. Sharing affect has not proved to facilitate soothing by a VIP, and affects have most commonly been managed alone with a variety of techniques which can at timesmimic autistic behaviours as these are self-focussed, such as self-stimulating behaviours.
In case it can be of help may I add a reference to a couple of books which do not use the Berkeley ABCD classification system associated with Mary Main, but instead use the Dynamic maturational model of attachment and adaptation (DMM) which does not make use of the Type D attachment: Wilkinson, SR (2003) Coping and Complaining: Attachment and the language of dis-ease. Brunner-Routledge, and with a greater focus on adult attachment see Crittenden, PM & Landini, A (2011) Assessing adult attachment: A dynamic maturational approach to discourse analysis. Norton.

I have questions. Isn’t a child with autism more likely to suffer attachment issues as well if the parent is unable to cope? Is it possible that some children have both?
Most things online are looking for differences and overlaps. Only one article pointed out that disabled children are more at risk of damage being caused because parents have struggled to care for them.

Is there a general presumption amongst practitioners to shy away from suggesting attachment disorders? I am asking as a member of the extended family of a child diagnosed with autism. The wider family has suspected (and at times witnessed) emotional abuse, neglect and extremely inconsistent parenting and the child in question’s behaviour modifies significantly when placed in an environment with consistent and fair expectations and timely praise. I understand that it could, of course, be possible for an autistic child to also have an attachment disorder, especially if one or both parents have traits of high functioning autism. If any family member dares to suggest there may be more to the situation than autism, to GPs, social workers and teachers, the conversation stops almost immediately, or we are given a short lecture on autism and reprimanded for daring to imply some shortcomings on the child’s parents’ part. This is an area that needs open and honest dialogue, as well as more research. The aim is not to blame the parents, but to ensure the child is given the best possible chance to lead a happy and fulfilling life. A culture where no questions can be asked and people close to the child other than the parents have no voice has lead (in our case) to a child with limited horizons who has been actively discouraged from developing any kind of emotional resilience, or offered any assistance with developing social skills and parents whose approach has never been challenged (perhaps ‘challenged’ is the wrong word to use, but nevertheless, autistic or with attachment issues, their approach has had a significant impact) As someone who is closely involved with a child with an autism diagnosis who has been let down massively by his parents, teachers and social services, I found the article fascinating and it should be the start of the conversation (it never promised to be the end!)

I’m finding the comments as useful as the article as I am the mother of autistic twins who spent months in newborn intensive care as a result of premature birth at 25 weeks gestation. While the staff at the NICU were diligent, the atmosphere and the lack of emotional content to their care, combined with constant staff turnover and reluctance to let parents so much as touch their children (this was several years ago) could only be seen as emotional neglect.
I have always wondered if there is any data or research about how lengthy hospitalization factors into either condition. Especially as the twin who left hospital sooner fared so much better in the long run.

I have a 32 month old grandson who was developing normally at 10 months. His mother committed suicide just before his first birthday. His mothers family kept us from him for over a year. He is mute and has no problem being shifted from one person to another. He is missing the markers of autism and he is starting to speak one word occassionally on a better environment. The mothers family has not had him assessed. Thank you for the article. Very interesting.

Add a comment

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