No, I had something totally different in mind -- that enhancing the ability to communicate would be a good thing regardless of how strong the treated individual's other cognitive abilities might be. I was not suggesting that there should or must be any effect at all on those other abilities.
Ah, ok. I guess what caused the confusion is my expectation that the "other cognitive abilities" are bound up with the inability to communicate in a rather more Gordian fashion than is implied by this idea. But I'm far from certain, of course, and your point seems plausible regardless.
I ran across this article that I think is interesting. It suggests that type 2 diabetes and the increase in autism may have a common link
http://www.frontiersin.org/Cellular_Endocrinology/10.3389/fendo.2011.00054/full