1. ¹Department of Psychiatry and Psychotherapy, Johannes Gutenberg University, Mainz, Germany
2. ²Institute of Neuroradiology, Johannes Gutenberg University, Mainz, Germany
3. ³Department of Psychiatry, Heinrich-Heine University, Dusseldorf, Germany

Correspondence: Dr A Konrad, Department of Psychiatry and Psychotherapy, Johannes Gutenberg University, Untere Zahlbacher Str. 8, 55131 Mainz, Germany. Tel: +49 6131 173701, Fax: +49 6131 176690, E-mail: andreas_konrad@gmx.de

Received 21 April 2008; Revised 14 June 2008; Accepted 2 July 2008; Published online 30 July 2008.

Abstract

Diminished left frontotemporal connectivity is among the most frequently reported findings in schizophrenia and there is evidence that altered neuronal myelination may in part account for this deficit. Several investigations have suggested that variations of the genes that encode the Neuregulin 1 (NRG1)–ErbB4 receptor complex are associated with schizophrenia illness. As NRG1-–ErbB4 has been implicated in neuronal myelination, we investigated with diffusion tensor imaging (DTI) whether fractional anisotropy (FA)—a putative measure of neuronal myelination—is predicted by a risk haplotype of the ErbB4 gene. The effects of the ErbB4 genotype were investigated in healthy subjects (N=59; mean age: 22.61.8 years). We also measured reaction time (RT) during a selective attention/working memory paradigm (visual oddball). In the schizophrenia risk genotype group, we found lower FA in the temporal lobe white matter (WM) including frontotemporal fiber tracts, predominantly in the left hemisphere. RT was increased in the risk genotype group and correlated with FA in the affected brain region. As FA is considered to index structural integrity of WM, to which neuronal fiber myelination is contributing, our results suggest that variations of the ErbB4 genotype may confer risk for schizophrenia illness via its impact on left frontotemporal connectivity in human brain. Reliability and validity of the result is suggested by our observation that (1) the FA–genotype association was not only obtained in the entire sample but also in both the split halves and (2) a statistical relationship was found among RT, genotype and FA.