Allostasis, epigenetics, and autism

This week I studied dementia in human development, and I guess one thing that is interesting is that there are many causes for dementia. Most are not Alzheimers. Maybe collapsing the diagnosis of autism wasn't such a bad idea, until we've got a bead on some causes, which I'm now thinking there will be more than one of. I'm also considering the concept of allostasis. Some are critical of the concept. It's embodied in the book "Zebras don't get ulcers". I'm not sure if allostasis is the best word for something like metabolic syndrome. Though I think the concept that insulin resistance is a defense against oxidative stress might fall into that realm. I think addiction is a similar flavor of problem (as it involves the enzymatic manipulation of intracellular messenging). If we expand this idea to behavioral addictions, it's not such a leap to autism. Some people are looking at attenuation to stress as an epigenetic problem, as with the study of cortisol in 9/11 babies. Though for it to be truly epigenetic, it would have to be children born from gametogenesis of 9/11 survivors. In utero transmission of cortisol management would point to an allostatic situation. You could think of this as the Scarlet O'Hara effect. In the novel, Scarlet is constantly disappointed in her own behavior, measuring herself against her mother. And she observes that she seems to be more like what people remember about her grandmother. The classic epigenetic study from Sweden found diabetes in people whose parents experienced famine; in utero for mothers and pre-adolescent for fathers. It takes a long view.