This issue starts with an overview of what we know from MRI neuroimaging of children and adolescents with psychiatric disorders. As is often the case in child psychopathology, accomplishments in adult psychiatry serve as the model for the study of childhood conditions. Thus Eliez and Reiss, in their clear overview of neuro-imaging studies, emphasise the results pertaining to child/adolescent onset schizophrenia, with its main conclusion that childhood-onset schizophrenia is not distinct from adult-onset schizophrenia from the standpoint of neuroanatomical variation. The use of the new imaging techniques in the search for neurobiological substrates of the more common psychiatric disorders in children and adolescents is still in its infancy. This can be concluded from the fact that only a few studies are available on children with ADHD, autism, or Tourette's disorder, with no neuroimaging studies yet available on childhood conditions such as depression, conduct disorder, anxiety disorders, or pervasive developmental disorders other than autism. A rather crude finding that stands out is the relationship between total brain volume and intelligence. The authors point out that studies that did not take IQ into account erroneously interpreted differences between groups of children to be specific for a certain condition, whereas in fact the differences could have been explained by differences between the groups in IQ. Reading this annotation, one is struck by the many contradictory findings of the neuroimaging studies that are reviewed. These discrepancies may be due to variations in neuro-imaging methodology and variations in the definition and measurement of neuroanatomical regions. Also, the large etiological heterogeneity, as well as the lack of precision with which we can define and measure behavioural phenotypes, will influence the variability in findings. The authors give an example of how one MRI study looked at children with ADHD without comorbid conditions whereas another study looked at children with ADHD of whom the majority had a comorbid diagnosis of conduct disorder or oppositional defiant disorder. It is not surprising, then, that neuroimaging studies on children with very different behavioural phenotypes, despite claiming that they studied the same phenomenon, end up with contradictory results. Instead of treating childhood conditions such as ADHD as diagnostic categories that are either present or absent, it may be advantageous for MRI studies to retain more diagnostic information by studying the covariation between brain morphology and behavioural phenotypes as continuous measures. Now that some experience with neuroimaging of child/adolescent psychiatric disorders is available, it is hoped that studies using larger sample sizes, advanced techniques such as functional MRI, longitudinal designs, and more precise diagnostic assessment techniques, will shed light on the still somewhat inconsistent findings on brain morphology of psychiatric conditions in children and adolescents.