Professor Fitzgerald comments on the ‘real-life’ utility of the ADI-R and ADOS. It is his view that this may have limited the representativeness of our sample with respect to the autistic presentation. However, the ADI-R or ADOS were used only for inclusion/exclusion when we were referred participants who did not have an existing clinical diagnosis of autism spectrum disorder (ASD). This was a small number of participants (n = 19), and the number who were excluded because they did not meet cut-offs was even smaller (n = 8). It is interesting to note that our experience actually supports Fitzgerald's observations, in that participants in the research did differ significantly from a comparison sample of people with ASD in terms of their ADI-R scores, as discussed in our article. ^{Reference Larson, Wagner, Jones, Tantam, Meng-Chuan and Baron-Cohen1}

The wider question alluded to in Fitzgerald's letter is one of categorical diagnoses and the utility of boundaries. Obviously, research requires a common language in order to facilitate discovery – I need to know that what I am measuring is equivalent to something of the same name measured in another country by another researcher at another time. Thus, instruments such as the ADI-R and ADOS (considered gold standards in ASD research) are vital. They also allow for meaningful comparison of groups – it is only when the cut-offs are applied rigidly that they become less useful, and this was not the case in our research.

However, a tremendous amount of heterogeneity exists among people with ASD (those who meet current categorical definitions), let alone those who have some symptoms but do not meet full diagnostic criteria (the broader autism phenotype). The challenge, then, is what to do with individuals who lie in different places along the spectra that comprise a standard ASD conceptualisation. As Fitzgerald rightly points out, it is those individuals who have independently learned or been supported to cope with their differences in a way that allows them to function in a ‘neurotypical’ society who are missed and excluded by the diagnostic categories that the ADOS and ADI-R conform to. However, simply because they appear to be functioning well does not mean that they are not experiencing difficulties – I agree with Fitzgerald. It was my experience conducting this research that, for many participants, it was actually because of the pressure of coping, or because they were not recognised as struggling, that many individuals got into difficult circumstances that precipitated the onset of psychosis or other serious mental health problems – an observation that is unsurprising for many clinicians, I'm sure.

The challenge for research and clinical practice, then, is to find a way to bridge the gap between rigid diagnostic categories and representative samples. This is a problem for psychiatry as a whole, not just those interested in certain conditions, which makes innovations such as the research domain criteria initiative from the National Institute for Mental Health so relevant and interesting. ^{Reference Sanislow, Pine, Quinn, Kozak, Garvey and Heinssen2}