The best theories in science are usually the simplest ones: those that don't fudge the issue to satisfy contemporary prejudices, appease vested interests, or flatter established authorities, but go right to the heart of the matter. As a result, they explain an awful lot with very little, and often link up several seemingly unconnected insights into one central, strategic revelation which exposes the fundamental truth behind appearances.
Whatever you say about the imprinted brain theory, you have to admit that, whether right or wrong, it shares this feature of stunning simplicity combined with surprisingly wide scope and remarkable relevance to findings which at first seem completely unconnected. Take cancer as an example. The theory holds that mental illnesses like autism and psychosis are diametric opposites of one another based on the oppositely-acting effects of conflicting parental genes. Paternal genes favour growth because they get all the benefits without the costs (gestation and breast-feeding), while maternal genes also get the benefits of growth, but have to pay the costs. The mother has to give birth to the baby, after all--not the father--and as every woman who has tried it knows, the size of a baby matters!
When crucial paternally-active genes like IGF2 are abnormally expressed from both parent's copies, the baby is more than one-and-a-half times larger than normal--and usually has to be delivered by Caesarean section. IGF2 is a growth-hormone gene and this condition is called Beckwith-Wiedemann syndrome. There are many other symptoms and signs of excess growth, among them a tendency to childhood tumors--pathological growth, if you like. Beckwith-Wiedemann children are also 7 times more likely than average to be diagnosed autistic, just as the imprinted brain theory would predict. Indeed, we now know that there is a wider link with cancer, and a recent study of children across the USA registered autistic under the Individuals with Disabilities Act found an association between autism and breast and uterine cancers.
Nor is this all. As my colleague and co-author, Bernard Crespi has pointed out, autism is also associated with a 100-fold greater risk of neurofibromatosis (benign nerve-tissue tumors) and, as the theory would lead you to expect from the Beckwith-Wiedemann precedent, higher levels of growth hormones like IGF2; an invasive, proliferating placenta during gestation; larger head and brain, with a thicker cortex in childhood; and higher birth-weight in male autistics, with faster body growth.
But perhaps most surprisingly of all, there is also evidence of decreased risk of cancer among schizophrenics and bipolar patients. As another recent study notes, bipolar patients have increased incidence of several physical disorders such as cardiovascular diseases, diabetes, hepatitis, and alcohol dependence. But they also have a lower incidence of hyperlipidemia, lymphomas, and metastatic cancer. And as an earlier study pointed out, the effects of lifestyle habits and particularly heavy smoking on patients with schizophrenia suggest that cancer risk should be elevated for this vulnerable section of the population. Yet the exact opposite is the case: schizophrenics, like bipolar patients, show lower rates of cancer. Indeed, when the effects of smoking are controlled for, the result is even more robust.
Could this be because psychosis by contrast to autism is associated with intra-uterine and placental under-growth; low birth weight and brain growth-factors; smaller brain size, thinner cortex, and smaller limbic system? And where cancer risk specifically is concerned, could the explanation be the associated decreased proliferation of stem-cells (which can be the precursors of cancers), increased expression of tumor-suppressor genes, and reduced thresholds for cell-suicide (or apoptosis: the normal, healthy response of a cell if it should become pre-cancerous)?
In short, once you see cancer in the right context--that of growth, pathological or normal as the case may be--its positive association with autism and negative one with schizophrenia becomes intelligible as part of a much bigger picture. That bigger picture is the one uniquely described by the imprinted brain theory. No other theory can make sense of so many findings--and certainly not those relating to cancer and mental illness.
(With thanks and acknowledgement to Bernard Crespi and Graham Rook.)