By far the most compelling evidence to date for the imprinted brain theory has come from a study of 1.7 million Danish medical records which revealed that birth size reliably predicts risk of an autistic as opposed to a psychotic spectrum illness in later life. According to the theory, psychosis results from relative over-expression of maternal genes and autism from relative over-expression of paternal ones. And because paternal genes tend to be growth-enhancing and maternal ones growth-restricting, birth size can be taken as a proxy for the relative levels of expression of parental genes.

But now two of the researchers who carried out that original study have returned to the same huge data set to test the theory against more of its predictions: namely those relating to parental age. As they point out, no previous study has simultaneously tested and compared parental age effects on autism- and schizophrenia-spectrum conditions in the same large population. Thanks to access to 30 years of national public health data in Denmark, the researchers designed their study “to rigorously assess the effects of both parents’ age, their age difference, and more than 20 covariates broadly across all autistic and schizophrenic disorders.”

As I explained in a previous post, parents are not the same where age effects related to risk of mental illness in their offspring are concerned. This is because sperm are constantly produced throughout a man’s lifetime, with the opportunity for copying errors in his DNA to accumulate, explaining the well attested finding that older fathers are more likely to have children with autism. By contrast, a woman’s DNA is sequestered undisturbed and un-copied in her ovaries from before her birth until an egg cell is finally released for fertilization in her adult life.

So what about maternal age and mental health? As the diagram below illustrates, the new study found that above-average paternal and maternal ages were consistently associated with increased risk of most autistic disorders in offspring (Fig. 1A and B). 

Evolution, Medicine, and Public Health [2016] pp. 286–298
Figure 1. Risk of offspring psychiatric disorders by parental age. Plots are divided by autistic (A–C) versus schizophrenic (D–F) disorders and by maternal age (A, D), paternal age (B, E) and parental age difference (C, F) at birth. Dashed horizontal lines (RR = 1.0) indicate zero risk. For parental age difference plots, groups left of centre represent mothers 1–3, 4–7 or 8–14 years older than their reproductive partners at childbirth. Groups to the right are fathers 7–10, 11–15 or 16–27 years older than their reproductive partners. Dark-grey dots mark risk P-values <0.05. All P-values were Bonferroni-corrected before further interpretation. Key provides full autistic and schizophrenic disorder group names for abbreviations in plots
Source: Evolution, Medicine, and Public Health [2016] pp. 286–298

Consistent with this was the finding that children of younger fathers were the opposite (Fig. 1B). This protective effect for autism was also found in offspring born to younger mothers (Fig. 1A). While below average maternal and paternal ages were both protective for autism, younger fathers showed the effect more. The only exceptions were behavioural and emotional disorders, which showed increased relative risk in younger paternal and maternal age groups (Fig. 1A and B). As the authors speculate, this may be because these disorders are more distantly related to the other autistic ones.

Trends for the five autistic disorders were much the same for the three older maternal age groups (Fig. 1A). However, by contrast to the findings relating to autism, neither advanced maternal nor paternal ages were associated with significantly modified risk of any schizophrenic disorder (Fig. 1D and E). Instead, patterns of risk for the five schizophrenic disorders in offspring born within the three younger than average maternal age groups were consistently increased (Fig. 1D), but not in offspring of the three younger than average paternal age groups (Fig. 1E).

As well as investigating age of both parents, the parental age difference variable allowed the researchers to separate dissimilar parental ages at birth, with values on the left side of the distribution corresponding to younger fathers reproducing with older mothers and values on the right related to younger’s mothers reproducing with older fathers (Fig. 1C and F). Overall, autistic and schizophrenic risk in offspring was lowest for more similarly aged parents. Risk for autism was higher towards the distribution edges, with risk for fathers aged 7–27 years older than mothers consistently more significant than that of mothers aged 1–14 years older than fathers (Fig. 1C). Schizophrenia risk was also largely U-shaped, with risk for most disorders increasing as parental age differences increased (Fig. 1F).

Offspring risk of many psychiatric disorders increased if mothers or fathers had the same disorder, if there were diabetes or hypertension complications during pregnancy, previous abortions, and if offspring were born more recently. Risks were often decreased if babies were born closer to term, had birth weights closer to the mean and higher APGAR scores. As these authors found previously, in their current analyses the risk of autistic disorders was generally higher in sons and the risk of schizophrenic disorders generally higher in daughters—contrary to conventional psychiatric wisdom, but just as predicted by the imprinted brain theory. Because all mothers are female and all fathers male, psychosis inevitably reveals a female bias of gene expression and autism a male one.

The researchers conclude that their “results lend support to many previous studies that have indicated that above average paternal and maternal ages are independently linked to increased risk of autism in their offspring.” They note that the change in autistic risk across the full paternal age range (Fig. 1B) appeared to be roughly consistent with the accumulated copying error theory. However, they add that this type of explanation is harder to maintain for older mothers—something that is particularly important in view of the fact that older mothers also confer increased risk of autism on their offspring.

The study found increased risk for schizophrenia, schizophrenia spectrum disorders, major depression and schizophrenia-schizotypal-delusional disorders in younger maternal age groups 15–21 and 22–24 years, but risks did not change uniformally across the entire age range (compared with autism, Fig. 1D), suggesting that psycho-social, cultural or resource-mediated factors might also be involved. For example, effects of early maternal age at birth on offspring schizophrenia risk have been linked to social and environmental stress factors that may be more common in earlier-reproducing families.

The study shows that autistic and schizophrenic disorder risks tend to be higher in more dissimilarly aged parents relative to parents with the most similar ages. However, autism risk plateaus towards the extreme parental age differences, where fathers are 11–27 years older than mothers, or where mothers are 4–14 years older than fathers (Fig. 1C). This suggests that there may be protective effects for autism conferred from younger maternal and paternal age groups 1–3 which neutralize the negative effects of the older parent (Fig. 1A and B). At the very least, this highlights the complexity of psychiatric disorders when they are affected by both parent’s age and their difference amongst many other factors within a single generation.

The overall picture is that schizophrenia risks are highest when parents (particularly mothers) are youngest, but these risks appear to fade away towards the middle of the reproductive period and start trending towards ever-larger autistic disorder risks as mothers and their partners age. As the authors point out, this is the pattern predicted by the imprinted brain theory “which hypothesizes that autism and schizophrenia are the extremes of a single perturbation gradient between paternal and maternal reproductive interests.” They also report that

older mothers tend to give birth to larger babies […], are more attentive on average, have less conflict with their offspring […], and see them grow up with fewer hospital visits and better than average general health and development […]. Our data also revealed that maternal age is significantly positively correlated with essentially all relevant birth-size traits (…). Such differences in offspring quality (i.e. health, survival) as well as a statistically decreasing rate of partner change with women’s age […] could all contribute to resource allocation during and after pregnancy being less constrained in older mothers.

The authors add that this is a consequence of evolution and that “Too few generations have passed to expect naturally selected ancestral life-history traits to have disappeared in modern humans, so maternal age appears to be a logical ultimate predictor of offspring risk of schizophrenia and autism, independently of birth weight.” Their diagram below summarizes the situation.

Evolution, Medicine, and Public Health [2016] pp. 286–298
Diagrammatic model of how diametrically opposite risks of mental disorders in offspring can be conceptualized to pivot between schizophrenia and autism as maternal age increases. (A) A 1911 reproductive value curve for Australian women drawn after R.A. Fisher’s version in his chapter on the fundamental theorem of natural selection, specifying that first births occurred at age ∼20 and last births normally at ages just beyond 40, and assuming that maternal median age at birth was ∼30 years. (B) The shift from maximal schizophrenia risk and minimal autism risk in offspring born to young mothers on the left, via minimal risk (zero when scaled relative to risk in offspring of median-aged mothers in the population) for any psychiatric disorder in offspring born to mothers of median reproductive value, to maximal autism risk and minimal schizophrenia risk in offspring born to mothers approaching menopause on the right, based on the overall patterns plotted in Fig. 1 and previously documented diametrically opposed risks of autism and schizophrenia dependent on size at birth. Differences in risk related to offspring being daughters or sons; are likely to be minor compared to the effects of maternal age-dependent patri/matrigenically induced provisioning biases and/or maternal genes whose expression is assumed to covary with age to express initially high but gradually diminishing resistance to patrigenic coercion for higher offspring provisioning in the womb and after birth
Source: Evolution, Medicine, and Public Health [2016] pp. 286–298

Finally, a point not made by the authors but perhaps worth adding is that the trend seen since World War 2 for mothers to have their families somewhat later in life might be yet another factor in the so-called “autism epidemic” of modern times. In terms of the diagram above, such a trend to later births would represent a secular societal swing of the pendulum towards an increasing risk of autism.

But women who have had their children later, or are thinking of postponing having them, can console themselves with a realization which only the imprinted brain theory and its diametric model of mental illness could have proposed and which this study strikingly endorses: This is that, if they are increasing risk of autism in their offspring, they are also simultaneously reducing the corresponding risk of psychosis! 

You are reading

The Imprinted Brain

Voyage to Kazohinia: A Diametric Dystopia

The diametric model of mental illness was anticipated in a novel of 1941.

Scans Reveal Diametric Differences in More Detail

Autism and schizophrenia are opposite in some brain areas, similar in others.

Imprinting and the Epigenetics of the Brain and Sleep

Two reviews make genomic imprinting a key factor in brain development and sleep.