24 Behavioral genetics
Ms. C, a 32-year-old woman with no history of mental disorder, visits her family physician exhibiting mild symptoms of depression. She has been listless and riddled with guilt since her long-time boyfriend moved out the previous month, along with their golden retriever. They had been having money problems since her shifts at work had been cut back, and she missed her boyfriend and their dog. She had been having trouble sleeping and had spent her sleepless nights scouring the Internet for hints about her state of mind and how to “snap out of it.” Ms. C tells her physician that she read a website about the serotonin transporter gene, 5-HTT or something, and that mutant versions of this gene make people more susceptible to depression. She asks for the physician to administer a test for this genetic mutation and warns that if he won’t order it she will just find another doctor who will.
Mr. and Mrs. D have recently moved to the area. They have an 11-year-old son, E, who has been acting like “a child they don’t even know anymore” since making friends with older kids in the new neighborhood. He won’t listen to them around the house, refuses to clean his room, always breaks his curfew, acts up in class, and sometimes smells like cigarette smoke. The same thing happened with Mrs. D’s brother at that age, and he wound up in a juvenile facility by age 14. She and her husband are concerned that maybe this kind of behavior runs in the family, and so they make an appointment with an internist to run a battery of genetic tests on E.
What is behavioral genetics?
Behavioral genetics is the statistical and, more recently, the molecular study of normal and pathological behaviors, their heredity, and their development. Behavioral genetics is usually understood to comprise psychiatric genetics plus the genetics of other behaviors and behavioral traits (Schaffner, 2006a). Behavioral traits include aggressiveness, criminality, fearfulness, homosexuality, intelligence, novelty seeking, political affiliation, and xenophobia, whereas psychiatric traits include depression, mood disorders, and schizophrenia. Behavioral genetics has a long history; some of the earliest family and twin studies began over a century ago, and the field has been under critical scrutiny for approximately 40 years (since Jensen’s [1969] controversial claims about race and intelligence). Until recently, most of the findings of behavioral genetics, while suggestive of genetic etiologies, have not borne fruit in terms of clinical or social application. The result is a well-justified and widespread skepticism about this field of inquiry. However, the recent history of behavioral genetics, characterized by advances in molecular genetics and genomics, suggests a more hopeful future for the field. In particular, the hopes of behavioral geneticists are buoyed by recent advances in the genetics of psychiatric disorders, which may translate into clinical applications for a range of conditions, including schizophrenia and depression. As a whole, though, behavioral genetics is not yet ready for the clinic. It is poorly understood, faces significant methodological challenges, and, by focusing on variation in the normal range, may strain the proper limits of the domain of medicine.
Behavioral geneticists employ a variety of methods, including so-called classical, quantitative, or epidemiological methods (family, twin, and adoption studies) as well as molecular techniques (such as linkage and association studies), to explore individual differences in behaviors. The individual differences approach helps to distinguish behavioral genetics from other enterprises, such as evolutionary psychology or developmental psychobiology, where the focus is not on differences in traits but rather on the traits themselves. Evolutionary psychologists study putatively universal traits, while developmental psychobiologists study developmental pathways within individuals. In contrast, behavioral geneticists attempt to understand why some individuals exhibit a trait while others (especially closely related others) do not. Why do some people have schizophrenia while others do not? Why are two diagnoses of schizophrenia more likely in twin pairs than in other sibling pairs, and more likely still in identical twins than in fraternal twins? These are the kinds of question that motivate behavioral genetics.
The logic of the individual differences approach helps to explain why behavioral genetics has not (yet) had an impact in the clinic: answers to questions about individual differences tell us nothing about why this particular individual has schizophrenia or another behavioral trait or disorder. This means that behavioral geneticists may be able to explain what makes a trait (say, aggressiveness) vary in a population without being able to explain why Johnny is aggressive and Jill is not. Put more technically, using the individual differences approach, behavioral geneticists try to explain how much of the total variance in a trait in a population can be explained by genetic variance, by environmental variance, and by the “interaction” between these two sources of variance. Yet, to explain variance is not to explain causation – the phenotypic variance may correlate with (be explained by) genetic variance even where genes are only one of many important causal factors. That is, it may be that the differences in a population with regard to political affiliation correlate with genetic differences, even if no genes are involved in actually causing one political affiliation or another.
Explanations of total variance in terms of genetic variance may, however, be suggestive of causation. For instance, the three classical study designs in behavioral genetics – family, twin, and adoption studies – produce increasingly compelling data about the heritability of traits. (“Heritability” is a term of art in genetics, referring to the percentage of the total variance in a trait that can be accounted for by genetic variance.) Individuals within families have more genes in common than between families, and so a trait that runs in families may correlate well with genetic inheritance patterns. Of course, families share much more in common than genes, and so family study designs are subject to significant confounding. Consequently, adoption studies have been employed to eliminate some of these confounding factors. In an adoption study, siblings reared apart, and their biological and social families, are compared for the presence or absence of some trait. If the trait is shared by biological parents and offspring, but not by adoptive parents or adoptive siblings, then a genetic explanation may be inferred. By contrast, if a trait is shared by adopted children and their adoptive parents, but not by the biological parents, then an environmental explanation may be inferred. Again, confounders may be present, including selective placement of adopted children in homes and with adoptive parents (environments) that are significantly similar to their initial environments.
Twin study designs may be able to reduce confounding further. Twin study designs depend on the fact that twins have more of their genes in common than they share with other siblings. On average identical (monozygotic [MZ]) twins have 100% of their genes in common, while fraternal (dizygotic [DZ]) twins, like any other siblings, have only 50% of their genes in common. On the assumption that the family environment related to a trait is equally correlated between MZ and DZ twins (the “equal environments assumption”), behavioral geneticists explain differences in traits between MZ and DZ twins by appeal to genetics. A standard pattern, under the equal environments assumption, is that traits are shared more often between MZ twins than between DZ twins (that is, MZ twins are more concordant than DZ twins for a trait). Where that pattern does not hold, genetic variance is held to be less important than environmental variance in explaining variance in the trait. The equal environments assumption is controversial. It challenges the commonsense intuition that parents treat their identical twins more alike than they treat fraternal twins. Therefore, many behavioral geneticists, and many critics of behavioral genetics, are skeptical of this assumption that environments do not vary between MZ and DZ twins in any interesting way related to the trait in question.
Even where family, adoption, and twin studies lead to consistent results about the role of genetic variance in explaining phenotypic variance with regard to a trait in a population, these results remain, at best, suggestive of causal factors. With advances in molecular genetics, the suggestive results of quantitative studies may help to lead to the identification of specific causal factors – and thus to clinically relevant strategies for prevention and treatment. Molecular genetic techniques include genetic linkage and allelic association studies. In a family with a disproportionate number of members exhibiting a behavioral trait, a genetic linkage study may be used to identify a gene of major effect. Success in a linkage study depends on three conditions: that there really is a gene of major effect involved in the genesis of the trait, that there is only one gene of major effect segregating in a family, and that we know the mode of inheritance of the gene. For a complex trait, none of these conditions is likely to be met, and so despite successes with single-gene disorders (such as Huntington disease), linkage studies have not been particularly fruitful in behavioral genetics (Robert, 2000). Allelic association studies work in the opposite direction: rather than starting with individuals manifesting a trait and searching for candidate genes, association studies begin with a candidate genetic variant and investigators test individuals with that trait to see whether they have the same genetic variant. If so, then behavioral geneticists will infer that the candidate allele really is involved, and they may conduct biochemical or other analyses to determine the mode of action. Again, allelic association studies may reveal spurious correlations – if an allele is in linkage disequilibrium with another allele, and only one of the alleles is involved in the genesis of the trait, an allelic association study may reveal association with the wrong allele.
Having knowledge of a correlation between genetic variance and variation in traits may sometimes suffice to design an intervention. For instance, if there is good evidence that a particular genotype is often associated with aggressive behavior, and if aggressive behavior may be prevented by a behavioral intervention, such evidence may guide school psychologists or social workers to direct their efforts at high-risk subgroups. But the surest route to the general and clinical utility of behavioral genetics is to take suggestive results (about individual differences) from quantitative, linkage, and association studies and then to explore causation (within individuals), so as to enable, where appropriate, prevention and treatment of the manifestation of certain behavioral traits.
Why is behavioral genetics important?
Behavioral genetics aims to study what makes human beings behave differently from each other. Accordingly, advances in behavioral genetics may threaten or reinforce long-cherished personal and social values and stereotypes, perceptions of group differences, and even perceptions of ourselves and our human nature. Moreover, findings in behavioral genetics may have broad applicability – for instance, in medical, legal, educational, and policy contexts – and so understanding the status and limits of behavioral genetics knowledge is imperative.
As with any studies of human behavior, and of human genetics, studies in behavioral genetics are prone to media hype. We have been bombarded by provocative stories in the popular press asking questions about whether babies are born gay or criminal, whether aggression or hyperactivity are in the genes, and whether propensity to get divorced is a genetic trait. These stories are not made up by the media; rather, journalists play up the more sensationalistic elements of actual research programs in behavioral genetics. As with other media reports about advances in genetics research, such stories result in patient inquiries for genetic services, from preconception, prenatal, and postnatal tests to genotyping and genetic interventions. Sometimes these requests will be entirely inappropriate, whether because of the test itself (e.g., there is no indication for ordering it) or the target of the test (e.g., there are ethical concerns about testing children for “gay genes”). But even in easier cases, behavioral genetic findings will raise a number of important ethical, legal, and policy considerations.
Consider highly publicized findings that the combination of a particular genotype with an early childhood experience of abuse can result in very bad outcomes (e.g., antisocial behaviors, including criminal charges and convictions; Caspi et al., 2002). Though these findings referred specifically to an interaction between genotype and environment (as the mutation alone does not result in antisocial behaviors), the study was widely reported as having revealed “criminal genes,” genes that lead to criminal behavior (Wilson, 2002). Thus a kind of genetic determinism was inappropriately read into the study, with significant implications for social policy – and potentially legal implications, too, if some defense attorney were to argue that a mutant gene made her client commit a crime, thereby abolishing criminal responsibility.
Clinicians must be especially attuned to interpretive challenges and the specter of genetic determinism in behavioral genetics. One helpful way to think about this issue was proposed by behavioral geneticist Eric Turkheimer (2000), who claimed that there are three laws of behavioral genetics: (i) that all human behavioral phenotypes are heritable (heritability 
0); (ii) that the effect of genes on phenotypic outcome is greater than the effect of being raised in the same family; and (iii) that, despite the second law, a significant amount of phenotypic variation cannot be accounted for by either genetic or familial effects. In his assessment of what these three laws actually mean, Turkheimer noted that, though well established, the first two laws are largely artifacts of the statistical techniques employed, and so not particularly informative of biological explanations of (differences in) traits. They do not, for instance, suggest that nature prevails over nurture, except in a methodological sense, as there is no environmental equivalent of identical twins – the fiction of identical environment twins (both MZ and DZ). The major problem with this state of affairs is the likelihood of overestimating the meaning – the significance – of behavioral genetic findings. Heritability calculations must be properly contextualized, for they may be as much artifactual as genuine estimates of the genetic influence on traits.
With significant progress in human genetics and genomics enabled by the Human Genome Project, and with the refinement of molecular and developmental methods, there is every expectation that behavioral genetic explanations of differences in traits will continue to increase. While not yet ripe for clinical integration, clinicians can expect to see an increase in patient requests for behavioral genetic tests and, possibly, interventions in the coming decade. As indicated below, such requests will often be sensitive, and understanding the limitations of current methods, and the genuine prospects of the field, will be important for clinicians. For it is clinicians who will be required to discern the propriety of patient requests for behavioral genetic tests and interventions.
Ethics
Difference and discrimination
Behavioral genetics is the study of individual differences. In many social contexts, the identification of such differences may be used as the basis for discrimination (whether justifiable or not). The most controversial studies in behavioral genetics involve criminality and general intelligence (IQ or intelligence quotient), not only because the phenotypes are so poorly understood, but also because behavioral genetics findings with respect to these traits have been used to mark hierarchical distinctions between racial groups. For instance, in The BellCurve, Herrnstein and Murray (1994) claimed that genetics explains the performance gap between whites and blacks on standardized educational exams such as IQ tests and scholastic aptitude tests. On the basis of this behavioral genetic evidence, they argued that investment in social programs to narrow this gap is unwise. While Herrnstein and Murray did not explicitly argue that whites are superior to blacks, and so deserve their elevated social standing, such a conclusion has indeed been drawn by many others – however confounded the data may be. In most societies, the politics of difference are fraught, and behavioral genetics promises to exacerbate the problem. Additionally, where behavioral genetic findings are used to justify (rather than simply to help to explain) the status quo, the potential for societal harm is increased.
Normality and medicalization
Behavioral genetics studies a full range of behaviors, from the normal to the pathological. But how to draw the line between normal variation and pathology remains an open question. This is a generic problem, not one particular to behavioral genetics, but this does not make it any less important to address in this context. For, as some critics have argued, many of the traits explored by behavioral geneticists lack construct validity; that is, the traits (constructs) under study are not particularly robust and are thus inconstant and open to wide variation in interpretation (Press, 2006). Within behavioral genetics, these traits may be pathologized, or medicalized – stipulated as pathological and so as falling within the purview of medicine. As Press (2006, p. 141) argued, “the reification of a fluid, continuous, and essentially normal part of the human behavioral repertoire as a bounded entity is a necessary precondition for a behavioral genetics investigation.” One of her examples is shyness, once a normal part of human behavior, now categorized in the American Psychiatric Association’s Diagnostic and Statistical Manual (1994) and subject to treatment with powerful pharmaceuticals. A recurring theme in bioethics is the ever-expanding scope of medicine, well beyond the historical limits of the field. Sometimes such expansions are entirely appropriate, as the trait in question clearly varies well beyond acceptable norms; other times, though, medicine overreaches, with significant social and ethical sequelae.
Eugenics
Eugenic considerations are not specific to behavioral genetics, though they are certainly germane. Whether and how behavioral genetics findings may be used to achieve eugenic goals is the subject of ongoing discussion and debate (e.g., Nuffield Council on Bioethics, 2002). The eugenics movement was founded by Sir Francis Galton in England in the 1860s. Eugenic means “well-born.” Inspired by the success of plant and animal breeders, Galton wondered whether the human race might be similarly improved through a program of eugenics: we could, he thought, decrease the number of “undesirable” humans and increase the number of “desirable” ones (Galton, 1869). Eugenics is usually divided into positive and negative varieties. Negative eugenics involves discouraging or preventing those deemed unfit from reproducing. Involuntary sterilization is an instance of negative eugenics. Positive eugenics is the encouragement of those deemed fit to reproduce in abundance, and to give birth only to the most perfect offspring. Though there was considerable social and scientific support for eugenics in the late nineteenth and early twentieth centuries, the technologies for achieving positive eugenics were not yet available. It is only in the past few decades that some of these technologies (such as prenatal and preimplantation diagnostic technologies) have been developed. Combined with findings in behavioral genetics, and especially with creeping medicalization, we may witness increasing social pressure to improve humankind by eugenic means. Indeed, some have argued (controversially) that it is morally imperative to use genetic selection technologies in support of eugenic enhancement (e.g., Savulescu et al., 2006).
Law
A different kind of concern raised by behavioral genetics – and, indeed, by any inquiry into human biology and especially the human brain – is the problem of free will. While human genetics and genomics have always raised issues of free will (Weir et al., 1994) – for instance, whether a genetically influenced disease is avoidable via behavioral modification – these issues are more acute where what is at stake is the genetics of behavior itself. In investigating the relationship between human behaviors and genes, behavioral geneticists may generate results relevant to our conceptions of ourselves, our choices, and our freedom to act. Such issues are important to our understanding of moral responsibility and accountability, but also to our understanding of legal responsibility and accountability.
Claims about free will and determinism – “my genes made me do it” – are not only the stuff of courtroom dramas on television. Behavioral genetics has made its way into the real-world criminal justice system in attempts to exonerate defendants or at least to mediate the severity of their punishment. Consider another finding regarding the monoamine oxidase A isoform. A particular mutation in the allele for this isoform eliminates all enzyme activity, thus dramatically altering metabolism of monoamines (e.g., serotonin, dopamine, epinephrine [Brunner et al., 1993a, b]. The behavioral outcome associated with this mutation is aggressive behavior. Lawyers attempted a legal defense on the basis of these “bad genes” in a murder trial in Georgia, and commentators have suggested that screening for mutations affecting monoamine oxidase A might be a good strategy for detecting potential criminals (Beckwith, 2006 [citing Morell, 1993; Felsenthal, 1994]). While the defense failed and no such genetic screening programs are (yet) in place, we can expect similar escapades in relation to genes correlated with impulsive and antisocial behavior of all sorts (e.g., Wasserman and Wachbroit, 2001; Edgar, 2006). Kenneth Schaffner (2006a, b) has written a terrifically clear imagined dialogue between a judge and a behavioral geneticist to help to shed light on the science and its implications.
Policy
One excellent policy-related report is that of the UK Nuffield Council on Bioethics. Their 2002 report, Genetics and Human Behaviour: The Ethical Context, lays out the historical, ethical, legal, and policy dimensions of behavioral genetics in clear, accessible terms. Focusing on behaviors in the “normal” range (rather than behaviors that fall clearly in the domain of psychiatry), their recommendations include the need for heightened awareness (and possibly government or professional oversight) of the possibility of inappropriate medicalization; that clinicians and policy makers evaluate proposed behavioral genetic enhancement interventions mindful of the prospect of furthering social inequalities; that any direct-to-consumer marketing of behavioral genetic tests be regulated and monitored as appropriate; that the legal system should not be persuaded by defendants’ claims that information about their genes (within the normal range) absolves them of legal responsibility for their actions (although behavioral genetic evidence may be useful in determining sentencing for crimes); that prescreening programs to identify potential criminals based on behavioral geneticss are entirely premature; that employers, educators, and insurers should have no special claim on behavioral genetic information; and that progress in public understanding of behavioral genetics begins with behavioral geneticists themselves, who should take particular care to communicate clearly their research findings and not to inflate their significance.
Empirical studies
There are very few empirical studies of the significance of behavioral genetics. One recent study, though, helpfully attempted to discern the beliefs and attitudes of healthcare providers and parents toward genetic tests for violent behaviors in children (Campbell and Ross, 2004). While such tests do not exist, Campbell and Ross based their study on plausible (though imaginary) tests and used the method of focus groups to study beliefs and attitudes toward these hypothetical tests. Particularly interesting results from the healthcare providers included that they tended to medicalize behaviors, that they would be reluctant to order tests in the absence of available treatments, and that they were concerned about the potential misuse of information from behavioral genetics. Parents, by contrast, tended not to medicalize behaviors and were not as concerned about testing in the absence of therapies. Yet parents may be ambivalent about testing their own children, and some in this study were concerned that behavioral genetic research may be detracting from political and environmental interventions that could be expected to yield socially beneficial outcomes.
How should I approach behavioral genetics in practice?
Given the methodological and ethical challenges discussed above, it is evident that behavioral genetics, as a science, is not entirely ready for clinical integration. Yet advances in the field, and especially in psychiatric genetics, will no doubt soon appear in clinical contexts, suggesting that clinicians would be well advised to anticipate potential clinical applications and prepare accordingly. Two exceptionally useful resources for clinicians are by Catherine Baker (2004) and Erik Parens, Audrey Chapman, and Nancy Press (2006). Both books were produced as part of a National Institutes of Health-funded collaboration between the Hastings Center and the American Association for the Advancement of Science to develop “tools for talking” about behavioral genetics, and they should help clinicians to be prepared for the safe, effective, and appropriate clinical integration of behavioral genetics.
As with genetics more generally, it is important for clinicians to understand the relevant science and, especially, the limits of the science of genetics for explaining development. Additionally, it is imperative that clinicians appreciate the scientific, social, and ethical complexity of genetic information; these issues of data management and counseling of patients are addressed in other chapters, particularly Ch. 22.
The cases
Ms. C is clearly exhibiting affective, somatic, and possibly cognitive symptoms of major depression. Whether she has a mutation in the serotonin transporter gene is moot from both a diagnostic and a therapeutic perspective. But just initiating treatment, whether with pharmaceuticals or talk therapy or both, will likely not satisfy Ms. C, who is adamant that she wants a finding of genetic susceptibility. A caring physician will want to explore Ms. C’s desire for genetic self-knowledge, but also to explain to her that her current symptoms are perfectly understandable given the range of negative events she has recently experienced. There is no shame in a diagnosis of depression, and there are many different therapeutic approaches that can help relieve her symptoms. Bringing behavioral genetics findings to bear in this case is entirely unnecessary and may even be more debilitating than helpful if, for instance, Ms. C learns that she does not have the mutation (she may feel shame as a result) or if a positive test result is eventually used by a third party in ways unfavorable to Ms. C’s interests. If Ms. C really does plan to find another physician to order the genetic test, her family physician should simply step aside. In this case, the physician explained some of these details to Ms. C and she was satisfied. She began antidepression therapy (drugs and talk therapy) and her condition improved within a few weeks.
As with Ms. C, Mr. and Mrs. D’s desire for genetic knowledge is understandable. Though they know full well that their son’s environment has changed, the personality changes are so dramatic that they feel a chemical imbalance must be involved – especially given their familial history. This case is an especially difficult one, inasmuch as a minor is involved, the parents want specifically genetic tests, the parents are clearly worried that they are responsible for these negative personality changes in their son because they moved to a new town, and this is the first visit with a new clinician. The clinician had to work hard to establish the trust not only of the parents but also of E himself. While it is possible that E was at genetic risk of antisocial behavior and that the new environment has functioned as a stressor to trigger the manifestation of these behavioral problems, it is not clear what additional value genetic tests would have for diagnostic or therapeutic purposes. On the basis of current knowledge, genetic tests are simply not indicated. Instead, the clinician assessed the family’s willingness to meet with a psychologist or other therapist and to work with teachers and school administrators to help to provide appropriate social supports for E. They agreed with the clinician’s advice that adolescence is not a medical condition, or a condition necessarily requiring medical treatment. But insofar as E was experiencing performance difficulties, and insofar as Mr. and Mrs. D were having trouble coping, it was appropriate for the clinician and other healthcare providers to help this family to adjust to their new circumstances. Though he still spends time with the older kids, E is now fitting in better with his peers and is doing just fine.