In general, a trait might be selected and sustained in a sub-population in the short term by environmental conditions, but over the long term there'll be a regression to genotypic mean when selection differential disappear, unless canalization.

 

There's no reason to suppose that this doesn't apply also to at least some behavioral traits, and the epidemiology of conditions such as psychopathy and autism provide some evidence that this is the case (?).

 

So culture can be affected by phenotypic expression of deviations from genotypic mean?

I imagine the relationship, like so man in nature, is reciprocal. Certain cultural factors may exert pressure which leads to the growth of certain behavioral phenotypes like psychopathy while behavior in turn affects culture. In the case of psychopathy, this seems like a truly vicious circle: Primary psychopaths create a cut-throat enviroment which encourages people with latent psychpathic traits to become secondary psychopaths in order to survive.