Aims This series of studies is based on a 10% sub sample of infants from a prospective longitudinal general population cohort, the Avon Longitudinal Study of Parents and Children (ALSPAC).

Methods We examined 180 videos of parent-infant interaction: 60 with infants aged 12 months who later proved to have a psychiatric disorder at seven years; and 120 randomly selected sex-matched controls. The case group included 27 children with any oppositional or conduct disorder (ODD/CD), 16 with attention- deficit/hyperactivity disorder (ADHD), 6 with pervasive development disorder/autism, and 28 with any anxiety or depressive disorder. Raters (blind to case or control status) made a series of observations which were later assessed for their value in predicting psychopathology at age 7: two behavioural coding systems (Mellow Parenting and CARE Index); measures of infant motor activity; clinical assessments by two groups of senior clinicians; holistic assessment of parent-infant interaction; and specific individual parental and infant behaviours.

Results Positive parenting behaviours as assessed by the Mellow Parenting system predicted reduced risk of development of oppositional/conduct disorders. The Infant Passivity subscale of the CARE Index appeared to predict autism. Clinical raters could not reliably predict later psychopathology and there was limited agreement between teams of raters. Infant motor activity did not predict psychopathology. The level of adult activity and vocalisation predicted whether infants would be diagnosed with ADHD, ODD/CD, and any anxiety disorder at seven years of age.

Conclusion There are robust early indicators of later onset of child psychopathology, identifiable in early parent-infant interaction at 12 months.