Wednesday, July 29, 2009

Twin Studies & Behavioral Genetics

The field of behavioral genetics is of great interest to me.

A lot of very good research has been done in this area for over 50 years.

One of the strongest methods of research in behavioral genetics is the "twin study", in which pairs of identical twins are compared with pairs of non-identical twins, looking at symptoms, traits, behaviors, disease frequencies, etc.

I would like to explore this subject in much greater detail in the future, but my very brief summary of the data is this:
1) most human traits, behaviors, and disease frequencies are strongly affected by hereditary (genetic) factors. Typically, about 50% of the variability in these measures is caused by variability of inherited genes. That is, the "heritability" is typically 50%, sometimes much higher.
2) The remaining variability is mostly due to so-called "non-shared environmental factors". This fact is jarring to those of us who have believed that the character of one's family home (a "shared environmental variable") is a major determinant of future character traits, etc.
3) Hereditary factors tend to become more prominent, rather than less prominent, with advancing age. One might have thought that, as one grows older, environmental events would play an ever-increasing role in "sculpting" our personalities or other traits. This is not the case.
4) Some of the "environmental variation" may in fact be random. Basically, good or bad luck. Getting struck by lightning, or winning the lottery, or not. Such "luck-based" events are mostly (though not entirely) outside our control.
5) All of these facts may lead to a kind of fatalism, a resignation about our traits being determined by factors outside our control. (mind, you, being "lucky" or "unlucky" may be more determined by attitudinal factors such as openness than just by random events: see the following article--

Here is some of my critical response to the above statements:

1) Statements about heritability are in fact dependent upon the average environmental conditions experienced by the population being studied. For example, if we were to measure the heritability of becoming the leader of a large country, we would find heritabilities of nearly 100% in times or places where there are hereditary monarchies, and much lower heritabilities for democracies (mind you, the case of the Bush family shows that the heritability has been non-zero in the U.S.).
2) Non-shared environmental factors are extremely important. This does not mean that the family environment is unimportant. Part of an individual's non-shared environmental experience is that person's unique experience of the family environment. The lesson in this is that families need to pay close attention to how each individual family member is adapting to the family situation, and to also pay close attention to a child's peer and school environment.
3) The influence of shared environmental factors is small, but rarely zero. Usually there is some small percentage of variability accounted for by shared factors. Often this percentage is larger in childhood, and declines towards zero during adult maturation. But it is not zero. Just because an influence is small does not mean that it is unimportant. We have limited control over our genetics, after all, but we do have more substantial control over shared and non-shared environmental variables.
4) Most studies look at the general effect of genetic & environmental factors in populations. Compelling examples are frequently cited of individual twins, separated at birth: perhaps one twin is adopted into a wealthy, privileged home with access to multiple educational resources, while the other grows up in a more impoverished setting. The story typically is that the twins both end up with similar funds of knowledge or intelligence: the first twin reads books available at home, while the other twin develops her inherited interest in knowledge by going out of her way to acquire a library card, and spending all day reading at the local library. Such case examples illustrate how inherited factors can prevail despite environmental differences.

But I'm interested to see counterexamples: examples in which differences in environment between twins did lead to substantial differences in traits later on. It is this type of example that has the most practical value, in my opinion.

5) I have considered the following idea:
For any trait or characteristic having any heritability, there may be environmental variables that can change the outcome of the trait for a given individual. Even for highly, obviously heritable traits. Consider eye color, for example. This seems obviously purely genetic. But suppose there was a medication that could change eye color. This would be a purely environmental factor (though, of course, perhaps the tendency to use a drug to change eye color would be partially inherited). Most people would not use such a drug. Measures of heritability for eye color would remain very high. But, despite this high heritability, there may well be simple, direct environmental changes which, for a given individual, could completely change the trait. Such environmental changes would have to be very different from average environmental conditions. The higher the heritability, the farther would the environmental change have to be from average, in order to effect a change in the trait.

We could say that the tendency to kill and devour wildebeest is heritable, among the different wild creatures of the African savanna. The genetic differences between lions and giraffes would completely determine the likelihood of such creatures devouring a wildebeest or not. We could say that lions inherit a tendency to eat wildebeest, while giraffes do not. Yet, I suppose that it is true that we could train and/or medicate lions (and also keep them well-fed with a vegetarian diet!) so that wildebeest are totally safe around them. In this way, we would be introducing a set of environmental changes which would cause a radical change in lion behavior. This does not change the fact that the heritability for lions' killing wildebeest is extremely high, it just means that the environmental change necessary to change the trait must have to be something radically different from the environmental experience of the average lion (most lions are not trained to be non-predatory!).

The clinical applications I have based on these observations are the following:

1) Many psychological phenomena are highly heritable. This does not mean that these phenomena are unchangeable though. It does mean that, in order to change the trait or behavior, an environmental change needs to occur which is substantially different from the environmental experiences of most people, or of the "average person". This may help us to use our efforts most efficiently. So, for example, it would be inefficient to merely provide everybody with a typical, average, 2-parent family living in a bungalo. The evidence shows that such "average" environmental changes have minimal impact on psychological or behavioral traits. It would be important to make sure each individual is not deprived or harmed, and has access to those basic environmental elements that are required for them to realize their potential. If there are problems, then the means of addressing those problems may require a substantial, unique, or radical environmental change.
2) The most influential environmental variables are those which are unique to the individual, not the ones which are shared in a family. This does not mean that family experiences are unimportant, but that a child's unique experience of his or her own family environment, is much more important than the overall atmosphere of the home. A chaotic household may be a pleasure, a source of boisterous social stimulation, for one child, but an injurious, disruptive, irritating source of stress for another. A calm household may allow one child to grow and develop, while it may cause another child to become bored or restless.
3) The higher the heritability, the more pronounced the environmental (or therapeutic) change is required to change the trait, compared to the average environment in the population.
4) The motivation to have a certain style of home, or parenting, etc. should logically not primarily be to "sculpt" the personality of your child, but to allow for joyous long-term memories, to be shared and recounted as stories by parent and child, and to pay attention to the unique nature of each individual child, providing for any healthy needs along the way.

Some references:

Segal, Nancy L. (2000). Entwined Lives: Twins and what they tell us about human behavior. New York: Plume.
{a 2009 review including a look at "epigenetics", the notion that one's genes are changeable, therefore identical twins are not truly "identical" in a genetic sense}
{genetics of PTSD}
{a look at how genetic factors influence environmental experience}
{a look at how choice of peers is influenced by heredity, moreso as a child grows up}
{some of the research showing different genetic influences coming "on line" during different stages of childhood and young adult development}
{a recent article by TJ Bouchard, one of the world's leading experts in twin studies}


Anonymous said...

Chapter 5..Starting on page 137.. may interest you.

I believe you can view the entire chapter for free (in a preview form)

Just a little more theory/concept definition. (Although you may already have this info)


GK said...

Thanks for the reference.

I'm just starting to look over it. The author's view appears to be quite extreme and unreasonable.

To minimize the relevance of genetic or inherited factors is, in my opinion, an eccentric and unfounded point of view akin to disputing evolutionary theory.

I do agree, though, that people tend to misunderstand the meaning and interpretative limitations of "heritability" calculations for human traits.

Twin studies such as those cited in the book I mentioned give strong evidence for a hereditary contribution to human traits. Particularly compelling are those case studies of identical twins reared apart in very different environments, who end up being substantially more similar than non-identical twins reared in very similar environments.

I am interested to develop a post describing, qualitatively, a more comprehensive concept of "heritability" which would be represented mathematically as a curve or a surface, rather than a point; this would account for heritability calculations being dependent on environmental or epigenetic factors, rather than being a "fixed" quantity which people would over-interpret.