Blinding Them with Science: The Epigenetics Issue

The idea of epigenetic factors—events that shape the ways genetic material is expressed in the developing individual—goes back a long way. For example, the geneticist Waddington suggested what he called an “epigenetic landscape”, in which factors like exposure to infection could push development into a different pathway from what might otherwise have been. These different pathways, once entered on, would further direct the course of development and might take the individual farther and farther from the developmental outcome that would have occurred without the infection.

Within the last half-century, it has become apparent that a range of factors can change the ways in which genes are expressed in individual development. One of these factors, a biochemical change called methylation, alters matters like the number of repeats of a genetic pattern in the individual genome. The very interesting thing about methylation is that this genetic change sometimes is, and sometimes is not, apparent in an individual’s developmental outcomes. For example, the gene responsible for Fragile X syndrome (a problem characterized by mental retardation and behavior problems) has already begun to change several generations before any developmental problems are apparent. A Fragile X patient’s mother and grandmother already had genetic changes at work, but these had not advanced to the point where development became atypical. (Note, by the way, that when genetic problems of this kind occur, some, like Fragile X, develop only in the maternal line, while others develop only in the father’s and  grandfather’s genomes.)

Fast forward now to 2004, when researchers reported that rat pups whose mothers licked them more (as well as showing more of other maternal behavior) had fewer genetic methylation events than those with less attentive mothers, that this situation persisted into adulthood, and that the less-mothered pups had worse reactions to stress in adulthood. The researchers, Szyf and Meaney, also that by cross-fostering pups (temporarily giving the less-mothered pups to more attentive mothers, and trading those mothers’ original pups to the inattentive mothers) they could actually reverse the differences between the groups, thus showing that the pups’ original genetic make-ups had not caused the differences between groups. As far as I can see, these researchers assumed that all effects had to do with the attentive and inattentive mothers’ behavior; they do not seem to have analyzed the milk of each mother for differences that might influence genetic material and development of the pups.

Very quickly, a rush to apply the rat studies to human beings began. Look, people said, early bad experiences change individuals, and what’s more we can fix them! For example, at www.center4familydevelop.com/Epigenetics.pdf, where attentive rat mothers are anthropomorphized as “conscientious”,  Arthur Becker-Weidman uses epigenetic studies to make the following claim:

“One clear implication of this research is that the Attachment-Facilitating Parenting with Attachment-Focused Psychotherapy can be instrumental in demethylating important genes and, therefore, ‘resetting’ the stress response system to be within a more normal range. It is clear that harsh parenting methods, methods that are shaming, blaming, and critical only serve to reinforce negative expectations and the unresponsive stress-response system’s reset mechanism [N.B. I do not claim to understand this last bit; I’m just reporting the news here. J.M.]. Parenting methods that are grounded in a focus on relationships and connection of an emotionally meaningful and joyful nature may reset the stress response system by its [sic] effects.”   

Let’s examine this claim under a couple of strong lights. The first has to do with extrapolation from rats to human beings (or even from a particular strain of rats to other strains of rats). As I pointed out on this blog some years ago, if Harry Harlow had used a different type of monkey, the results of his work, and their capacity to support Bowlby’s attachment theory, might have been quite different (http://childmyths.blogspot.com/2011/09/you-can-pick-your-friends-but-you-cant.html). Generalization from one species to another takes careful work to substantiate it, especially when the species are so different, with much different life spans and times to reach reproductive maturity. Rats take about 30 days from birth to reproductive age, as opposed to 12-14 years for human beings, so how do we decide which part of a rat’s early development is parallel to any part of human development? In addition, the work of rat mothers involves building a nest, licking the young, positioning herself so they can nurse, and retrieving them in response to their squeaks if they manage to get out of the nest. If disturbed by noises or other animals, the rat mother may well eat the babies (and although I suppose this may have happened among humans, even Greek myths usually concentrate on mothers who serve the kids up to their fathers for dinner). I leave it to readers to think about parallels to human infant care, but want to point out that humans typically have one baby at a time and concentrate on that baby’s care, with at least a year ordinarily intervening before a new baby arrives—humans are not caring for a litter of young ones as rats do.

To go on, let’s look at the statement that a kind of intervention can be instrumental in methylating certain genes. This is a rather staggering leap of logic. For one thing, it suggests that  behavioral change through psychotherapy results from change at the genetic level, caused by the therapy. As we have yet to understand any specific relationships between human genes and human behavior, the suggestion that changed genes mean changed behavior, and vice versa, is far from warranted. Unless we are to assume that all learning results from genetic change (which would be absurd), this cannot be a correct connection to make. Even if we were to have evidence that some intervention changed some genetic material, which we do not, it would be necessary to show systematic evidence supporting the claim that any other intervention had the same effect.

In addition, in this particular case of generalization, there seems to be a questionable assumption that human beings can at any point in their development be influenced in their genes and behavior by treatment, in ways parallel to the effects of cross-fostering on rat pups very early in their development. It’s curious, too, that on an Internet site so devoted to the ill effects of adoption on development, the comparison is made to cross-fostering and reversal—these rat pups were reported to do better with “good” mothers rather than to suffer from separation from the birth mother.

 Unfortunately, it is becoming all too common to see claims that because established research evidence supports part of an idea or a practice, it can then be assumed that the whole idea or practice is equally evidence-based.  When this approach is applied to epigenetics, some wild conclusions can be drawn in the course of attempts to translate lab findings into social recommendations (as Juengst et al noted in Trends in Genetics, https://www.researchgate.net/publication/265853557_Serving_Epigenetics_before-Its-Time).    

(My thanks to Yulia Massino for suggesting this source.)

The lessons of “blinding with science” are applicable not only to epigenetics, but to a variety of other attempts to bring scientific concepts to bear on what are essentially pseudoscientific ideas and practices. One example is the effort to bring quantum mechanics into explanations of various “energy therapies”. Another is the enormous proliferation of supposedly brain-based educational interventions that emerged from the “split-brain” work of Roger Sperry (for an account of this, see Michael Staub, “The other side of the brain: The politics of split-brain research in the 1970s-1980s”. History of Psychology, 19(4), 259-273).

Moral: Before hurrying to conclusions about application of research findings to the more complicated world outside the lab, we need to ask ourselves—“exactly what does the research have to do with the practical problem we’re worried about?” Often the answer will be “Nothing—yet.” Realizing that can help keep us from treatment scams of which there are so many looking for suckers.