I have been re-reading Edelman's NEURAL DARWINISM in preparation for a trip I am taking to his Institute in February. In the process of doing so, I have been trying to link his ideas of evolutionary & ontogenetic selective (versus instructive) processes as they relate to the construction of behavior.
A major challenge is how one translates from ANY genetic change to a coherent change in behavior (versus just messing up the works). The problem becomes still more complex as soon as we think about some brain processes as being "local" (affecting specific or restricted functional pathways) and others being "global" (setting behavioral tone, broadly defined goals, etc.). These layers then have to be coordinated. As far as I know, no one understands how this can work.
This may all sound esoteric (it certainly is complex), but it strikes at the very heart of evolutionary models of behavior. If anyone has ideas (or would like clarification) I would love to hear.
Getting from gene changes to coherent brain/neural circuit changes is a fundamental mystery....to me at least!