An inside job or a set-up?

It is indisputably the case that genetic and biological factors play a non-trivial role in predisposing people to criminal behaviour. I firmly believe this because the past decade of biological research has revolutionised our knowledge of crime and clearly implicated genetic and biological factors in the development of criminal behaviour. Yet in the face of the clear scientific evidence, some individuals deny this conclusion and remain bitterly opposed to such research.

Let me describe why I find the notion that biology underpins crime to be so compelling. First and foremost, one has to consider the genetic evidence from twin and adoption studies. There have been more than 13 studies of adult crime, and all of them find that monozygotic or identical twins are more criminally similar than dizygotic twins. On average, where one member of an identical pair is criminal, 51.5 per cent of the co-twins are criminal compared to 20.6 per cent for fraternal twins (who are less alike genetically).

These data lend clear support to the notion that crime is partly heritable. They do not stand alone. Results from 14 adoption studies conducted in three different countries support this conclusion. Such studies find that if a child has a criminal biological parent, but is adopted at birth into a non-criminal foster-home, that child is more likely to become criminal than a child also adopted by non-criminal foster parents but who had a non-criminal biological parent.

Crime in general is partly heritable, but what about violent crime in particular? The evidence here is more conflicting. The landmark adoption study published by Sarnoff Mednick in 1984 found clear evidence for heritability for non-violent property crime, but not for violent crime per se. Other adoption studies have supported this finding. Yet in the past six months a different view has emerged. Findings from two twin studies reported at the American Society for Criminology last October reported strong heritability for violence. Han Brunner from the Netherlands reported in Science the identification of the type of genetic mutations which produce biochemical brain abnormalities associated with aggression and violence. At next week's CIBA meeting in London, Sarnoff Mednick will report new findings from a large adoption cohort showing substantial heritability for violent crimes in adults.

The notion that crime has a biological basis does not just rest on genetic studies. There are thousands of studies on factors from diverse fields of biological research which have also repeatedly demonstrated links with crime. In the field of neurochemistry, Markku Linnoila at the National Institute of Mental Health has demonstrated that impulsive violent offenders have low levels of serotonin, a brain chemical which underlies information processing and communication within the brain. Neuropsychologists such as Terrie Moffitt at Wisconsin have clearly shown that delinquent criminals have brain dysfunction. Localisation of this dysfunction remains a future challenge, but it seems that the left hemisphere (involved in verbal skills), the frontal cortex (self-control, intellectual flexibility, and personality regulation) and limbic structures (regulation of emotion) may be damaged in violent offenders.

Our own psychophysiological research has shown that low levels of physiological arousal (heart rate, EEG, skin conductance) measured at age 15 in normal schoolboys predicts at an accuracy rate of 75 per cent who becomes criminal nine years later. Hans Eysenck in London and Herbert Quay in Miami have speculated that these low levels of arousal may lead some individuals to seek out stimulation to increase their arousal levels back to normal; joining a gang, burgling a house, and robbing someone seem for some adolescents to be stimulating and exciting. Psychophysiological studies have also shown that criminals do not form conditioned fear responses easily, a finding which supports Eysenck's seminal theory of crime developed more than 30 years ago.

Yet perhaps the most exciting development in the past decade lies in the area of brain imaging. As one example, using positron emission tomography (PET) we recently found that 22 murderers are characterised by reduced glucose metabolism (indicating poorer functioning) in the prefrontal region of the brain relative to 22 matched controls. Brain imaging now allows us to document directly the types of brain dysfunctions which clearly exist in violent offenders, and such research could have major ramifications for the field.

Two decades ago, computerised tomography (another brain imaging technique) provided irrefutable evidence that schizophrenics have enlarged ventricles in the brain. This finding destroyed the myth that schizophrenia was largely a psychosocial disorder stemming from disturbed family communication and instead established schizophrenia as primarily a biological disorder. Furthermore, brain imaging coupled with cognitive remediation techniques may help us to both detect and compensate for brain damage at a fairly early age, thus helping to prevent delinquency and crime.

The list of biological correlates of crime goes on. Criminals have a history of head injuries, some of which are probably caused by environmental accidents which could be preventable. They are more likely to suffer birth complications which can contribute to brain damage. They have minor physical anomalies (for example, low-seated ears, furrowed tongues) which are not visibly obvious but which reflect foetal maldevelopment during the second trimester of pregnancy. They are thought to have a mesomorphic-endomorphic body build (muscular and large) which through the process of instrumental learning may encourage fighting. Male criminals have lower levels of the hormone cortisol (possibly reflecting low anxiety) and higher levels of testosterone, a steroid hormone known to be causally linked to aggression in animals. Low progesterone has been linked to aggression and violence in women, although this finding remains controversial. Aggression, delinquency and criminality have been linked to hypoglycemia (low blood sugar levels), symptoms of which can include irritability and aggression and which can be corrected through dietary change.

In addition, there are many psychological factors that are strongly influenced by genetics. For example, verbal IQ has a heritability of at least 50 per cent and is also the best replicated psychological correlate of crime, being approximately ten points lower in criminals and delinquents relative to controls. High impulsivity and high psychoticism also have high heritabilities and are two of the many well-replicated personality correlates of crime.

What are the implications of a biological basis to crime? One conceptual implication is that it supports the notion that repeated, recidivistic criminal offending may be a clinical disorder. If one were to view recidivistic crime as such, offenders could be punished less harshly, vigorous attempts could be made to increase research and clinical efforts to understand and treat crime, and the criminal justice system would probably require revision. Until we can successfully treat crime, we must institutionalise repeat offenders to protect society. But perhaps we should not kill them and treat them inhumanely as many societies do.

These implications are not popular. But do we want to be driven by a retribution that originates from our evolutionary and sociobiological past, or do we want societies to take into account the constraints that early social and biological predispositions place on free will? Whatever the merits of the case, this "bleeding heart'' approach to crime is undoubtedly one of the sources of the opposition to biological research on crime; society is frightened that such research could be used to let offenders off the hook. And nobody wants to see that happen, particularly those who sit on the political right.

Paradoxically, those who sit on the political left have a radically different concern. They erroneously think that biology means destiny, and are worried that those with the biological markers for crime and violence will be either placed on mind-controlling drugs, or rounded up and imprisoned even before they have committed any crime.

To some extent I share this concern. I have low levels of physiological arousal (one of the biological predispositions to crime), and I have a brain scan ominously like that of a multiple murderer. Yet so far I have not murdered any of my critics, and I certainly do not want to be locked up for my crimes until I have committed them. But this honest political concern is based on a misunderstanding of biological processes. Biology is not destiny. Many biological factors are caused by the environment, not genetics, and by changing the environment we can change the expression of the biological and genetic predispositions to violence. A recent large study of 4,269 males found that birth complications (ie, forceps delivery, breech birth, lack of oxygen) combined with early maternal rejection (ie, mother not wanting the pregnancy, child institutionalised in its first year) predisposes to violent crime 18 years later. Only 4 per cent of the sample had both of these factors, but they accounted for 18 per cent of all the violent crimes committed by the entire sample. By providing better ante-natal and perinatal health care to underserved inner-city mothers we may be able to prevent these birth complications. This in turn would neutralise the biological ingredient essential to this recipe for violence and could reduce the rates of violent crime by about 18 per cent.

Birth complications are a biological factor, but they can be prevented. The same is true for genetic factors which critically require an environment to receive expression. Change the environment, and we may be able to prevent the latent genetical underpinnings of crime from taking effect. How do biological and genetic factors interact with environmental processes in producing crime? What are the psychological, psychosocial, cultural, and community factors that protect a person who has the biological profile of a criminal from actually becoming criminal? How can we develop benign biological interventions to help reduce violence in society? These are the critical questions we sorely need to answer in order to generate new advances in understanding and preventing crime. And we will only answer these questions if we adopt a biosocial approach which integrates research findings from the biological and social sciences.

If on the other hand society wants another five decades of failed attempts to prevent crime and violence in society, I strongly recommend that it continues to put politics before science and ignores biological processes.

Adrian Raine is professor of psychology at the University of Southern California.