by David E Comings


For understanding the causes of human behaviour, the twentieth century could be considered the century of Freud. By contrast, the twenty-first century will undoubtedly be considered the century of the gene. In the first half of the twentieth century Freud and other theorists proposed that most behavioural disorders were the result of problems in the early childhood years. For years the concept that a major dimension of behaviour was controlled by genes was generally ridiculed. In recent decades these concepts have almost reversed themselves. Except for extreme neglect or abuse, the early childhood years have been shown to have little effect on adult behaviour. By contrast, there is much greater appreciation of the role of genetic factors. The subject of the genetics of human behaviour is so vast, and so many different genes have now been implicated for such a wide range of behaviours, that this article will be limited to reviewing a series of basic, but important concepts.

Behavioural Traits are Due to Both Genetic and Environmental Factors

Twin studies are uniquely suited to identifying the relative contribution of genes and environment to a given trait. Identical twins share all of their genes in common, whereas fraternal twins share only 50%. The frequency with which both twins share a given trait is called the concordance rate. The ratio of the concordance rate for identical versus fraternal twins provides an estimate of the relative role of genetic versus environmental factors. If a trait is strongly genetic, the concordance rate for identical twins approaches 100%, while the rate for fraternal twins approaches 50% and the ratio of the two approaches 0.5. For purely environmentally caused traits the ratio approaches 1.0.

Personality traits are considered to be characteristics of fundamental but basically independent types of behaviour in normal individuals. For example, in the NEO Five Factor Personality Inventory the traits are agreeableness, conscientiousness, neuroticism, extraversion and openness. In the seven factor Temperament Character Inventory they are cooperativeness, harm avoidance, novelty seeking, persistence, reward dependence, self-directiveness and self-transcendence. While personality traits and the ratio of concordance in identical versus fraternal twins may vary considerably from study to study, as a generalization that keeps both the geneticists and environmentalists happy approximately 50% of most personality traits are due to genetic factors and 50% to environmental factors.

A second fundamental characteristic of humans is intelligence. Many twin studies of general intelligence have been performed. After age 15 and before extreme old age, the genetic contribution to intelligence is remarkably stable, with a modal heritability of 0.81. The examination of twins reared apart support the results of twins reared together, showing a weighted average heritability of 0.75. This contrasts to an average correlation of a range of traits of 0.04 for unrelated siblings (adoptive siblings) raised in the same home.

Some Common Behavioural Disorders are Over 70% Genetic

In addition to personality traits, human behaviour is characterized by a range of disorders. The Diagnostic and Statistical Manual of the American Psychiatric Association is divided into Axis II or Personality Disorders, and Axis I or Clinical Disorders. The personality disorders consist of borderline, antisocial, narcissistic, paranoid, obsessive-compulsive and other disorders that are considered to be related to basic personality traits but severe enough to cause difficulty in a person's life. The clinical disorders are the most severe. They include major depression; bipolar or manic-depressive disorder; schizophrenia; generalized anxiety, obsessive-compulsive, phobic and panic disorders; alcohol or drug dependence; the childhood disorders such as attention deficit hyperactivity disorder (ADHD), Tourette syndrome, conduct disorder (CD), oppositional defiant disorder (ODD), autism; and others.

While twin studies show that the genetic component in most of these disorders range from 30 to 70%, for some, especially the childhood behavioural disorders, the genetic component is even higher. This is not surprising as, in general, the higher the degree of genetic loading the earlier the age of onset. In recent years there have been a number of twin studies of childhood and adolescent ADHD. These have all been in agreement that genetic factors account for 70–95% of the variance. Between 30 and 60% of children with ADHD have comorbid or coexisting CD or ODD. Conduct disorder is characterized by the presence in children or adults of behaviours which disregard the rights of others and include stealing, fighting and physical or sexual assault. While most theorists and therapists in the twentieth century considered CD to be an environmental disorder, largely due to bad or neglectful parenting and broken homes, a large Australian study of over 2600 twins showed that, in common with ADHD, CD was over 70% genetic. The implications of these findings for the treatment of these childhood disorders are profound, as they indicate they are primarily biological and biochemical rather than psychological in origin.

Early Common Environment Plays Little Role in Most Behaviours

Twin studies allow the environmental component to be divided into shared (common) and unshared environments. The shared or common environment refers especially to the early years when the twins are being raised together and experience the world together. The unshared environment refers to situations which are not shared. These are more likely to occur later in childhood and in adolescence. One of the remarkable and surprising findings of twin studies, for both personality and childhood behavioural disorders, is that the contribution of the common environment approaches zero. Rowe has stated that, ‘Given the environmental emphasis on behavioural science theories, the idea that the common environment influences fail to impact on personality development is radical; but it is, nevertheless, supported by an extensive literature of twin and adoption data.’ Further support for the minimal role of shared environment comes from studies of unrelated siblings reared in the same household. Here the concordance rates for a number of personality traits are essentially zero.

Role of Genes on Environmental Factors

In addition to the independent roles of genes and the environment, there has been increasing awareness of a role of the genetic factors in what on the surface appear to be environmental factors. For example, Rowe utilized a twin study to examine the Family Environment scale, a widely used self-assessment to evaluate the family environment in which an individual was raised. He compared how identical versus fraternal twins assessed their family environment. This might seem a silly study since the family was the same for both twins and thus the assessments should be highly correlated. This was true for the assessment of some characteristics, such as firm versus lax control, and control versus autonomy. However, for acceptance versus rejection, by the father or the mother, the correlation was significantly higher for identical twins than for fraternal twins. This indicated that one's perception of what was thought to be purely an environmental factor was itself under genetic control. These findings were replicated in subsequent studies.

The influence of genetics on factors that appear to be environmental can explain why some psychological studies have implicated a role of early shared environment on behaviour, while twin studies have failed to confirm this. Thus, Reiss and colleagues reported that 60% of the variance in adolescent antisocial behaviour could be predicted by conflictual and negative parental behaviour directed specifically at the adolescent. However, this ignores the possibility that parental behaviour and the behaviour of the child may both be due to common genetic factors. When the data were reanalysed to include this possibility, it was found that most of the association between parental behaviour and adolescent outcomes was genetically mediated.

Behavioural Traits and Disorders are Polygenic

The major emphasis in human genetics since the early 1900s has been on disorders that are due to single gene defects. By the end of the century, the genes involved in all of the major single gene disorders, such as Huntington disease, cystic fibrosis, muscular dystrophy, neurofibromatosis and hundreds of others, had been cloned and sequenced. There are a few rare examples of a single gene being involved in a primary behaviour disorder, i.e. one that is not also associated with physical or neurological defects such as mental retardation or seizures. Such a disorder, involving the absence of function of the monoamine oxidase A gene, was described by Brunner and colleagues. This mutation completely destroyed the function of the gene. Since it was an X-linked gene, in males there was no normal gene on the other X chromosome to compensate for the defect. This was associated with aggressive and hypersexual behaviours involving only the males in the family. However, as typical of single gene disorders in general, these families are very rare.

Despite the great intellectual excitement and precision of these studies, all of the single gene disorders combined, involving not only behaviour but all other conditions, do not account for more than 1% of the total of all human disease. By contrast, the complex and common disorders and traits such as height, weight (obesity), intelligence quotient (IQ), learning disorders, blood pressure, osteoarthritis and others, including all behavioural disorders, are inherited in a polygenic fashion; that is, a number of genes are involved. The variants at the relevant genes are largely additive in their effect.

Because of the great emphasis on single gene disorders, there is a tendency to think of polygenic disorders as an extension of single gene disorders, simply involving perhaps 3–5 genes in a given individual instead of one. These are termed oligogenetic disorders. While it may be true that some disorders, such as Alzheimer's disease, may largely involve variants of a small number of genes, accumulated evidence from many sources suggest that most behavioural disorders are truly polygenic. Thus, dozens of different genes involving the regulation of neurotransmitters, neuropeptides, transporters, secondary messengers, gene activation factors and hormones, are involved. While a single individual may begin to show a phenotypic effect after inheriting only a modest number of the relevant variants, different combinations of 50–100 different genes may be capable of producing similar behavioural phenotypes. This is termed genetic heterogeneity, i.e. the ability of multiple different genes or combinations of genes to produce similar phenotypes.

Related Behaviours Share Genes in Common

A major distinction between single gene and polygenic inheritance is that single gene disorders are due to a mutation of a single specific gene. Thus, Huntington's disease is due to a mutation of the huntington gene, cystic fibrosis is caused by a mutation of the cystic fibrosis gene, etc. By contrast, in polygenic inheritance different disorders may share genes in common; for example, ADHD, CD, ODD and substance abuse may share genes in common. The family histories of single gene and polygenic disorders are also different. Because single gene disorders are uniquely caused by specific genes, the relatives only have that disorder. Thus, the father of a patient with Huntington disease may also have Huntington disease, but he is unlikely to have a different single gene disorder such as neurofibromatosis. By contrast, it is common for an individual with a polygenic disorder, such as ADHD, not only to have relatives with ADHD, but also with CD, ODD, antisocial personality disorder, alcoholism, drug abuse, learning disabilities and other interrelated entities. This is illustrated in the pedigree shown in Figure 1. These are referred to as hereditary spectrum disorders, as a wide spectrum of phenotypes can occur both in the proband and in the family members.

Figure 1. Example of a spectrum of disorders in the relatives of a proband with attention deficit hyperactivity disorder (ADHD). The proband (arrow) had ADHD and obesity. Her older brother had Tourette syndrome and alcoholism (A) and her younger brother had ADHD, phobias and alcoholism. Disorders in other relatives are shown.

Polymorphisms are Used to Identify Specific Genes in Behavioural Genetics

While twin and adoption studies have been responsible for most of the shift in thinking in the later part of the twentieth century about the importance of genetic factors in behaviour, to truly understand this relationship will require an understanding of which genes are involved in which traits. The ability to do this was dramatically accelerated by the development of the polymerase chain reaction (PCR) technique. This allows deoxy ribonucleic acid (DNA) segments to be amplified so the polymorphisms (as variants) they contained could be tested in a given individual. The type and location of polymorphisms used in behavioural genetics are shown in Figure 2. One type of polymorphism, the single nucleotide polymorphisms (SNP), can be located in exons, where they can cause a change in the amino acid sequence or can be silent (no change in amino acid sequence). Alternatively, they can occur in the introns, or in the 5¢ or 3¢ noncoding sequences. A second common type of polymorphism is the short tandem repeat polymorphism (STRP).

Figure 2. Type and location of polymorphisms used in association studies for the identification of specific genes involved in behavioural disorders (see text). Solid blocks represent exons. The numbers, as in C 1456 T, represent the nucleotides, counting from the start site of the gene. A, adenine; G, guanine; C, cytosine; T, thymine; Cys, cysteine; Lys, lysine.

Genetic Variants in Polygenic and Single Gene Disorders are Fundamentally Different

The genetic changes that cause single gene disorders are called mutations. Mutations typically cause severe disruption in the function of the gene they affect, which is why only a single such mutation is required to cause disease. It makes more sense to call the genetic changes that cause polygenetic disorders variants, as the effect they have on the function of genes is minor. Table 1 lists some of the important differences between single gene and polygenic disorders.

Table 1. Single gene versus polygenic inheritance

Single gene disorders             Polygenic disorders

Number of genes                                    1                                              5 to dozens

Frequency of the disorder                      Rare (< 1%)                          Common

Frequency of mutation/variant               Rare (< 1%)                          Common (5–95%)

Type of mutation/variation   Exons, splice junctions, deletion        Introns, 5¢ and 3¢ sequences, repeat polymorphisms, mild if in exons

Effect on gene function                          Major                                       Minor

Direction of effect                   Usually decreased expression           Increased or decreased expression

Identification by                                       Linkage analysis                 Association analysis

Type of subjects studied                       Large pedigrees                 Probands versus controls

Variance explained per gene               100%                                     0.5–5%

Specificity                                                 Disease specific                Spectrum of disorders

The mutations in single gene disorders cause such a severe defect in the function of the gene that only one is required to cause disease. By contrast, the variants involved in polygenic disorders cause such a mild disruption in gene function that the additive effect of many genes is required to produce a phenotypic effect. Since the frequency of single gene disorders is so low, the frequency of the mutations that cause them is also low. By contrast, since the frequency of polygenic disorders is high, and since multiple variants are necessary to cause each disorder, the frequency of the variants must be very high. For example, if a polygenic disorder present in 3.3% of the population was caused by five equally important genes, each variant would have to be present in at least 50% (0.5) of the population (0.5 × 0.5 × 0.5 × 0.5 × 0.5 = 0.033). In single gene disorders, in order to cause complete disruption of the function of a gene, the mutations must be in a critical part of the gene, such as in the transcribed portion (exon) causing a change in the amino acid sequence of the gene, in the promoter region to alter the expression of the gene, or at exon–intron junctions to disrupt splicing. The mutations may also cause deletions of critical parts of the gene or involve insertions that cause a frameshift in the reading sequence. By contrast, for polygenic disorders, since the variants must cause only minor changes in gene function, they are likely to occur in the same regions as above but have minor effect on gene function, or occur outside the gene and produce only minor effects. One theory that explains these requirements is that the alleles of microsatellite polymorphisms or STRPs, which are very common in the genome and tend to especially occur in five regions, may be responsible for the variations involved in polygenic inheritance. These tandem repeats commonly form an altered configuration of DNA which affects the function of nearby genes. The individual genes involved in polygenic disorders account for only a small percentage of the variance of the respective phenotypic trait, rarely more than 5% and usually less than 2%.

For single gene disorders, the best method of identifying the genes involved is linkage analysis. Since a large number of polymorphisms localized to specific chromosomes and chromosome segments are available, families with a large number of affected individuals can be genotyped at hundreds of different markers. Computer programs are then used to determine if there is a significant tendency for a given marker to be transmitted with the trait. If this occurs, the marker and the trait are said to be linked. Since the chromosome location of the marker is known, this indicates the chromosomal location of the gene responsible for the trait. While this powerful technique has been used to identify the chromosomal location of the genes for many different single gene disorders, linkage analysis loses power when multiple genes are involved.

The genes associated with polygenic disorders are best identified by comparing the frequency of alleles of polymorphisms at known candidate genes in severely affected probands versus unrelated, unaffected population controls of the same ethnic background. These are called population-based association studies. Since the frequency of the alleles of many genes vary in different racial and ethnic groups, one of the potential problems with population-based association studies is that hidden differences in the ethnic make-up of the controls versus test subjects may be responsible for the differences in the frequency of alleles. One method to correct for this is to examine both parents and the proband. These are called family-based association studies. If one parent is heterozygous for a two-allele polymorphism and the other is homozygous, one can test if the percentage transmission of the alleles to the proband is significantly greater than the expected 50%. These two types of association studies are responsible for identifying most of the genes presently known to be associated with human behavioural disorders.

An additional aspect of the theory that STRPs, and SNPs that have only a minimal effect on gene function, may provide the genetic variations that are responsible for polygenic inheritance is that there may be no such thing as a ‘normal’ gene. For example, as shown above, the genetic variants for polygenic disorders must be very common, their effect on gene function must be mild, and the number of alleles associated with increased gene expression are similar to the number of alleles associated with decreased gene expression. This concept is important because it suggests that two different alleles of any common SNP in or near a given gene will tend to sort into two groups containing hypofunctional versus hyperfunctional alleles. The hypofunctional alleles may be associated with one phenotype, while the hyperfunctional alleles may be associated with a different phenotype.

Specific Genes for Specific Disorders

The two major psychotic disorders, schizophrenia and manic-depressive disorder, have been the focus of many attempts to identify the specific genes involved. While linkage analysis has identified many possible linkages, replication across different studies has been difficult. Manic-depressive disorder has been linked to 14 different chromosomal regions, none of which have been widely replicated. Similar problems have been encountered with linkage studies of many other behavioural disorders, including Tourette syndrome, autism, major depression, panic disorder, phobias and others. With the possible exception of the Golf gene on chromosome 18 in manic depressive disorder, as of this writing, linkage analysis has yet to identify a single gene for any common behavioural disorder. These difficulties are due to the polygenic nature and genetic heterogeneity of these disorders.

Association studies are better suited to detecting the role of genes that account for only a small percentage of the variance; however, because of the small effect of each gene and genetic heterogeneity, it is unusual for the majority of subsequent studies to replicate an initial significant association. Despite this, there are many examples of multiple replications of the association of specific genes with specific behaviours. Only one example below is presented: the role of the dopamine transporter gene in ADHD.

Dopamine Transporter Gene and Attention Deficit Hyperactivity Disorder

The most commonly used medications to treat ADHD are methylphenidate and dextroamphetamine. These drugs exert their effect on ADHD by binding to and inhibiting the action of the dopamine transporter, a protein on the surface of dopamine neurons that transports dopamine from the synaptic space into the presynaptic neuron. This makes the dopamine transporter gene (DAT1) a good candidate gene for ADHD. This is strengthened by the observation that knockout mice, missing the DAT1 gene, are very hyperactive. Using family-based association techniques, Cook showed that the 10 allele of a polymorphism of the human DAT1 gene was associated with ADHD. This has been confirmed in multiple additional studies using both the family- and population-based association techniques, and it has been found that different alleles of the DAT1 gene identify those individuals who do or do not respond to methylphenidate.

Genetic Variants are Additive in their Effect

One of the characteristics of polygenic disorders is that each gene has only a modest effect on the phenotype. In studies of the effect of over 45 separate genes on a range of different behaviours, each gene usually accounted for less than 2% of the variance. The combination of these small effects plus genetic heterogeneity often make replication of an association between a behaviour and a specific gene difficult. However, since a major characteristic of polygenic inheritance is that the phenotype is due to the additive effect of multiple genes, these problems with replication can be decreased by examining the additive effect of multiple genes. Figure 3 is an example. In this study the TaqI A1 allele of the dopamine D2 receptor gene (DRD2), the TaqI B1 allele of the dopamine beta-hydroxylase gene (DBH), and the 10/10 genotype of the dopamine transporter gene (DAT1), were found to be associated with ADHD. Each association has been verified in independent studies. Subjects were divided into four groups, those who carried 0, 1, 2 or 3 of the relevant alleles. There was a progressive increase in the mean ADHD score in individuals carrying 0, 1, 2 or 3 of such alleles. The resultant significance level was much greater than for any individual gene. We have extended this concept to examine 20 genes associated with dopamine, serotonin and noradrenaline (norepinephrine) metabolism. Combined, these accounted for over 10% of the variance of ADHD, P < 0.006. When the same group of individuals is genotyped for multiple genes, it is also possible to examine the relative importance of the three functional groups of genes by comparing the additive variance of each group. This shows that four genes involved in noradrenaline (norepinephrine) metabolism accounted for almost twice the variance accounted for by the dopamine or serotonin genes.

Figure 3. Additive effect of dopamine genes on attention deficit hyperactivity disorder (ADHD) in Tourette syndrome probands. From Comings DE et al. (1996) American Journal of Medical Genetics (Neuropsychiatric Genetics) 67: 264–288, by permission.

Genetic Testing will Play an Important Role in the Diagnosis and Treatment of Behavioural Disorders in the Twenty-first Century

It has been generally assumed that, since polygenic behavioural disorders are due to the additive effect of multiple genes, and that since they do not produce a highly specific phenotype, genetic tests will be of no value for diagnosis. However, the development of the ability to genotype polymorphisms at hundreds of genes using DNA arrays has eliminated many of the technical problems of examining large numbers of genes at a reasonable cost. As the number of genes identified for given phenotypes increases and the percentage of variance explained increases, genetic testing will become more and more valuable as a means of providing more precise and accurate psychiatric diagnoses. It is likely that one of the most important and powerful uses of genetic testing in the coming century will be to identify precisely those medications that are the most effective for a given individual. Such testing could reduce the time it takes to find effective medical treatment from months, and sometimes years, to days.


The presence of more than one disorder in a given individual.

An estimate of the relative importance of the role of genetic factors in a trait, expressed as a fraction, i.e. heritability of 0.5 suggests that genetic factors account for 50% of the trait.

A region of DNA containing short (1–6 base pairs) repeat sequences such as CACACACACA.

The physical or mental characteristics that define a given syndrome or trait.

A mechanism of inheritance due to the additive and interactive effect of many different genes.

Common variations in the genetic sequence present in at least 1% of the population.

The initial member of the family to be studied for a genetic disorder.

Short tandem repeat polymorphisms
Polymorphisms due to the presence of microsatellites or minisatellites.

Spectrum disorder
A disorder commonly associated with a wide range of other comorbid behaviours.

A statistical term relating to the proportion of a trait that is due to a given factor such as a gene; for example, gene X accounts for 2% of the variance of a behavioural trait.

Further Reading

  • Bouchard TJ Jr (1994) Gene, environment, and personality. Science 264: 1700–1701.
  • Bouchard TJ Jr, Lykken DT, McGue M et al. (1990) Sources of human psychological differences: The Minnesota study of twins reared apart. Science 250: 223–228.
  • Comings DE (1996) Polygenetic inheritance of psychiatric disorders. In:BlumK, NobleEP, SparksRS et al.. (eds) Handbook of Psychiatric Genetics, pp. 235–260. Boca Raton, FL: CRC Press
  • Loehlin JC (1992) Genes and Environment in Personality Development. Newbury Park: Sage.
  • Plomin R, Owen MJ and Mcguffin P (1994) The genetic basis of complex human behaviors. Science 264: 1733–1739.
  • McGue M and Bouchard TJ Jr (1998) Genetic and environmental influences on human behavioral differences. Annual Review of Neuroscience 21: 1–24.
  • Seeman P and Madras BK (1998) Anti-hyperactivity medication: methylphenidate and amphetamine. Molecular Psychiatry 3: 386–396.  
  • Sherman DK, McGure MK and Iacono WG (1997) Twin concordance for attention deficit hyperactivity disorder: a comparison of teachers’ and mothers’ reports. American Journal of Psychiatry 154: 532–535.
  • Slutske WS, Heath AC, Dinwiddie SH et al. (1997) Modeling genetic and environmental influences in the etiology of conduct disorder: a study of 2682 adult twin pairs. Journal of Abnormal Psychology 106: 266–279.
  • Recommend Us