Research suggests that children who are exposed to high levels of testosterone in the womb may be more likely to develop autistic traits. Since testosterone also plays a role in sexual differentiation as a male sex hormone, it is though that such exposure may also result in these children having more masculine characteristics irrespective of whether they are biologically male or female. On the basis of this evidence, one might therefore expect that boys with autism would be more masculine than typically developing boys, whilst girls with autism would be less feminine: a prediction that has been referred to in the literature as the "extreme male brain" (EMB) theory of autism.

Despite the biological plausibility of EMB, the theory has found only partial support in research to date. Whilst there are studies which suggest that females with autism may be more masculine in certain respects than non-autistic females (consistent with EMB), other evidence suggests a pattern of gender identity defiance (GID), where autistic females are more masculine and autistic males more feminine than non-autistic controls (inconsistent with EMB).

Our study therefore aims to test these two competing models of gender identity in autism - EMB versus GID - by examining the gendered play behaviours of children with autism and that of their non-autistic peers on various occasions throughout childhood.