Yang et al (Reference Yang, Raine and Lencz2005) propose a neurodevelopmental theory of pathological lying, finding increased prefrontal white matter and lower prefrontal gray/white ratios in pathological liars compared with antisocial and normal controls. Spence (Reference Spence2005) asks whether these findings represent cause or effect. Since lying is a criterion symptom for childhood conduct disorder, we reexamined a structural magnetic resonance imaging study of early-onset conduct disorders (Reference Kruesi, Casanova and MannheimKruesi et al, 2004 plus unpublished data).

Youths had been classified as liars or not based upon structured interviews and collateral information when documenting criterion symptoms of conduct disorder. Liars (n=6) were compared with individuals with conduct disorder (n=4) and with healthy volunteers (n=10). The mean ages of the three groups (191.5, 195 and 190.8 months) were similar (F(2,19)=0.015). In accordance with developmental differences, ratios of prefrontal white volume to total brain volume were lower in our three groups of youths (0.039, 0.026 and 0.034 for liars, antisocial controls and healthy volunteers respectively) than in the corresponding groups of adults reported by Yang et al (0.069, 0.054 and 0.054). However, prefrontal white to whole brain ratio, prefrontal white volume, or prefrontal grey/white ratios did not differ between our youth groups (F(2,19)=1.105, 0.973 and 0.337 respectively).

We also examined the corpus callosum morphometrically using the method of Casanova et al (Reference Casanova, Sanders and Goldberg1990). Since Raine et al's (Reference Raine, Lencz and Taylor2003) strongest effect size was seen for corpus callosum volume and limited data were available, we calculated the ratio of corpus callosum area to whole brain volume as a proxy for corpus callosum volume. A trend for ratio differences between the three groups was seen (F(2,19)=2.748, P=0.092), with the smallest ratios in the liars (0.080), followed by antisocial controls (0.086) and healthy controls (0.091).

Thus, we did not find prefrontal differences in lying youths but did find suggestion of corpus callosum differences. Our results are consistent with the notion that prefrontal findings are not causal, although they may be linked to the maintenance of the symptom of lying and consistent with myelination proceeding rostrally and from the inside (longer connections) outward (short association fibres and arcuate fibres).