Two scientists, drawing on their own powers of observation and a creative reading of recent genetic findings, have published a sweeping theory of brain development that would change the way mental disorders like autism and schizophrenia are understood.

The theory emerged in part from thinking about events other than mutations that can change gene behaviour. And it suggests entirely new avenues of research, which, even if they prove the theory to be flawed, are likely to provide new insights into the biology of mental disease.

At a time when the search for the genetic glitches behind brain disorders has become mired in uncertain and complex findings, the new idea provides psychiatry with perhaps its grandest working theory since Freud, and one that is grounded in work at the forefront of science. The two researchers - Bernard Crespi, a biologist at Simon Fraser University in Canada, and Christopher Badcock, a sociologist at the London School of Economics, who are both outsiders to the field of behaviour genetics - have spelled out their theory in a series of recent journal articles.

"The reality, and I think both of the authors would agree, is that many of the details of their theory are going to be wrong; and it is, at this point, just a theory," said Dr. Matthew Belmonte, a neuroscientist at Cornell University. "But the idea is plausible. And it gives researchers a great opportunity for hypothesis generation, which I think can shake up the field in good ways."

Their idea is, in broad outline, straightforward. Crespi and Badcock propose that an evolutionary tug of war between genes from the father's sperm and the mother's egg can, in effect, tip brain development in one of two ways. A strong bias toward the father pushes a developing brain along the autistic spectrum, toward a fascination with objects, patterns, mechanical systems, at the expense of social development. A bias toward the mother moves the growing brain along what the researchers call the psychotic spectrum, toward hypersensitivity to mood, their own and others'. This, according to the theory, increases a child's risk of developing schizophrenia later on, as well as mood problems like bipolar disorder and depression.

In short: Autism and schizophrenia represent opposite ends of a spectrum that includes most, if not all, psychiatric and developmental brain disorders. The theory has no use for psychiatry's many separate categories for disorders, and it would give genetic findings an entirely new dimension.

"The empirical implications are absolutely huge," Crespi said in a phone interview. "If you get a gene linked to autism, for instance, you'd want to look at that same gene for schizophrenia; if it's a social brain gene, then it would be expected to have opposite effects on these disorders, whether gene expression was turned up or turned down."

The theory leans heavily on the work of David Haig of Harvard. It was Haig who argued in the 1990s that pregnancy was in part a biological struggle for resources between the mother and unborn child. On one side, natural selection should favour mothers who limit the nutritional costs of pregnancy and have more offspring; on the other, it should also favour fathers whose offspring maximize the nutrients they receive during gestation, setting up a direct conflict.

The evidence that this struggle is being waged at the level of individual genes is accumulating, if mostly circumstantial. For example, the fetus inherits from both parents a gene called IGF2, which promotes growth. But too much growth taxes the mother, and in normal development her IGF2 gene is chemically marked, or "imprinted," and biologically silenced. If her gene is active, it causes a disorder of overgrowth, in which the fetus' birth weight swells, on average, to 50 per cent above normal.

Biologists call this gene imprinting an epigenetic, or "on-genetic," effect, meaning that it changes the behaviour of the gene without altering its chemical composition. It is not a matter of turning a gene on or off, which cells do in the course of normal development. Instead it is a matter of muffling a gene, for instance, with a chemical marker that makes it hard for the cell to read the genetic code; or altering the shape of the DNA molecule, or what happens to the proteins it produces. To illustrate how such genetic reshaping can give rise to behavioural opposites - the yin and yang that their theory proposes - Crespi and Badcock point to a remarkable group of children who are just that: opposites, as different temperamentally as Snoopy and Charlie Brown, as a lively Gaugin and a brooding Goya.

Those with the genetic disorder called Angelman, or "happy puppet," syndrome practically dance through the day, have difficulty communicating and are demanding of caregivers. Those born with a genetic problem known as Prader-Willi syndrome are placid, compliant and, as youngsters, low maintenance.