Ten years ago parents would often ask whether their child's autism spectrum disorder (ASD) might have been triggered by an important event in the family, say the birth of a sibling, moving house or starting day-care. The common factor in all these suggested causes was the timing – parents typically mentioned events that occurred when the child was around 18 months old. It was straightforward to reassure parents that no such ordinary events could cause or trigger ASD, and that these events were merely coincidental with the typical timing of the first signs of the disorder. However, when, five years ago, claims were made of a connection between MMR vaccination and ASD, parents were faced with a whole new set of worries, which were less easy to allay.

Since the original reports, there has been a huge controversy, fuelled by poor journalism and irresponsible reporting. Naturally, parents look for an identifiable cause for their child's increasingly divergent development, or even apparent loss of skills. A great deal of research money has been diverted to explore the possible role of MMR in ASD. The balance of data has rested strongly on the side of no link, but the paper by Honda, Simizu and Rutter, in the present issue, brings a new level of evidence to bear on the question. Previous studies have examined time-links between the introduction of MMR and the number of cases of ASD reported. However, changes in diagnostic criteria, general awareness, and specialist services have all contributed to better identification of ASD in recent years. Honda et al. report data on incidence of ASD in Japan, where an MMR vaccination programme was introduced in 1989 and withdrawn (due to worries about the particular mumps strain used) in 1993. If MMR was responsible for significant numbers of cases of ASD, withdrawal of the triple vaccine should result in a fall in the incidence of ASD. No such fall was found in this study, indeed the cumulative incidence of ASD up to age 7 increased in the birth cohorts of years 1988 through 1996, and especially from 1993. There was also no evidence of an increase in the proportion of children with ASD who showed a regression – which had been theoretically linked to MMR in the original claims. MMR does not appear to be responsible for the rise in ASD, then, although epidemiological studies cannot rule out the possibility that small numbers of children who develop ASD are especially vulnerable to vaccine damage.

Why then have numbers of individuals with identified ASD increased in recent years? Honda et al.'s analysis of time trends in incidence of ASD at different IQ-levels may provide a clue. While incidence of both narrowly-defined autism and ASD increased, the most notable increase was in numbers of cases with high IQ. This suggests that broadening of diagnostic criteria, greater awareness of the spectrum of autism disorders, and better diagnostic services may account of the rise in identified cases of ASD. Additional biological factors may also, however, play a part.

To identify possible environmental factors in ASD, genetically sensitive designs will be vital. Several papers in the present issue use twin designs to identify genetic and environmental contributions to individual differences in developmental difficulties. Notable among these is the paper by Viding, Blair, Moffitt and Plomin on callous-unemotional traits that may predispose youngsters for psychopathy. As Gelhorn et al. in this issue report, studies of conduct disorder have provided surprisingly variable estimates of heritability, perhaps due to use of different assessments, at different ages, resulting in heterogeneous groups. While Gelhorn et al. examine symptom heritabilities, and divide symptoms into aggressive and non-aggressive sets, Viding et al. focus on core empathic impairments that are hypothesised to characterise an important subtype of conduct disorder. From teacher report at age 7 years, children were identified who were at the extreme for antisocial behaviour, and further subdivided into those who in addition showed callous and unemotional (nonempathic) traits. This additional classification proved extremely important; callous-unemotional traits showed high heritability, as did antisocial behaviour combined with such traits. Antisocial behaviour without such callous traits, however, was only moderately heritable, with considerable influence of shared environment. As the authors discuss, this supports the importance of callous-unemotional traits for both molecular genetic investigations, and in considering the heterogeneity of response to intervention. Aetiologically heterogeneous samples may explain in part why intervention programmes for conduct disorder often have mixed results. Awareness of key subtypes (with versus without empathic disorders) may clarify these results, perhaps revealing high success rates for some subtypes. For example, conduct disorder without callous-unemotional traits appears to be more strongly affected by shared environment and may provide the best target for interventions aimed at improving family, school or neighbourhood conditions. Antisocial behaviour with empathy disorders, on the other hand, may require targeted neurocognitive intervention to improve or prosthetically replace normal responses to others’ distress.

Another paper in the present issue that suggests important subgroups for studies of aetiology and intervention, is Gathercole at al.'s article reporting data from the ALSPAC study. Gathercole at al. identified children with low performance on phonological memory tests at age 5, and followed these to age 8 years. Two subgroups emerged. The first appeared to have a specific deficit in the phonological loop; they had persisting phonological memory problems at age 8, but performed as well as controls on vocabulary, language and maths tests. The second group performed better on phonological memory at age 8, but had enduring language impairments from 5 to 8 years. This group is proposed to have semantic processing deficits, rather than phonological loop problems, underlying their persistent language difficulties. General working memory capacity was related to scholastic achievement in this group.

The other important papers in this issue include a valuable review of research on velo-cardio-facial syndrome and its association with schizophrenia (Murphy); a behaviour-genetic study of genetic and environmental contributions to parental perceptions regarding their 5-month-olds (Boivin et al.); a study of the impact of remote experience of September 11th on Washington schoolchildren (Lengua et al.); an exploration of the effects of ethnic origin and acculturation on psychiatric problems in immigrant adolescents (Oppedal et al.); and experimental evidence of normal prototype effects in memory in children with autism or Asperger Syndrome (Molesworth et al.).