AUTHOR COMMENTARIES - SPECIAL TOPICS

In Essential Science IndicatorsSM from Thomson Scientific, Professor Murray's citation record includes 163 papers cited a total of 4,865 times in the field of Psychiatry & Psychology, as well as 52 papers cited a total of 1,427 times in the field of Neuroscience & Behavior.

Professor Murray is a Professor of Psychiatry in the Institute of Psychiatry at the Maudsley, King's College, University of London; he is also a consultant psychiatrist at Maudsley Hospital.

How did you get into this field?

I started work in the 1970s. At that time some of the biological theories, or explanations, for schizophrenia were crazier than the condition itself. We were interested in the theory that people with schizophrenia were walking hallucinogenic factories, and we studied the hallucinogenic substance dimethyltryptamine. We found significantly more dimethyltryptamine in the urine of schizophrenic patients than in other psychotic patients or in healthy controls. So for a while we seemed to be onto something.

However, about this time the dopamine hypothesis became the pre-eminent neurochemical explanation for schizophrenia. Dysregulation of the dopamine system was subsequently strongly linked to psychosis and schizophrenia, with abnormally high striatal dopamine shown to underlie these conditions. So although the endogenous hallucinogenic theory was never disproved, it just died of fatigue and disinterest.

Could you tell me about your group?

I am at the Institute of Psychiatry, London, which is by far the largest psychiatric research center in Europe. We have about 1,000 academics and researchers, of whom about 150 concentrate on schizophrenia and bipolar disorder. Within our group there is a range of research mainly looking at etiological questions. In epidemiology we take advantage of the UK's National Health Service that has provided us with data across the population and over a large span of time. Epidemiology is important to us in part because our US colleagues cannot do it: there is no database of patient records at the national level in North America.

We work on the precursors of psychosis. For example, we ask questions such as: what are the children like who later develop schizophrenia, or, what characterizes children who will later have bipolar disorder? We look at which groups are particularly prone: city dwellers, migrants, and so forth. This research is impossible in countries with fragmented health care.

How disabling is it? Can the average person hold down their job for example?

No, mostly they cannot, and so the condition contrasts with most other psychiatric disorders. Sadly, the majority of people with schizophrenia will not recover to their previous level of functioning. Among sufferers, the luckiest 20% have just one episode, after which they may still have some problems, but are able to manage without medication. At the other extreme, the worst 20% will go into a nosedive and deteriorate: they become very sick and highly dependent. Then in the middle, we have 60% who will have one or more episodes, from which they recover to a certain extent, but after each episode they don't quite regain their previous level of functioning.

What are environmental risk factors?

There are a number of risk-increasing factors. One of the most important is living in a city. There is ample evidence from across the western world that being born and brought up in a city increases risk. The best evidence is from Denmark: people brought up in Copenhagen are about 2.5 times more likely to develop schizophrenia than people in the rural population.

Does this suggest it is stress related?

We do not know. Researchers have pointed their fingers at everything from lead poisoning and subsequent cognitive impairment through to the idea that it is easier to become isolated and paranoid in the inner city; another theory is that illicit drug use is more prevalent in cities. However, none of these theories have been proven.

Among your most-cited papers are the two, with Alastair Cardno as the first author, that use twin studies to investigate genetic relationships (Cardno AG, et al., "Heritability estimates for psychotic disorders - the Maudsley twin psychosis series," Arch. Gen. Psychiat. 56[2]: 162-8, February 1999, and Cardno AG, et al., "A twin study of genetic relationships between psychotic symptoms," Amer. J. Psychiat. 159[4]: 539-45, April 2002).

Here at the Maudsley Hospital, we are very lucky to have a twin register. Anybody who presents to our hospital is asked if they are a twin. The question sometimes perplexes the patients although it is useful for us!

The Cardno study was the first to look at the monozygotic co-twins of people with schizophrenia to see not only if they also had schizophrenia or alternatively were well, but also to see if they had suffered from mania. What was surprising was that when we relaxed the hierarchical system of classifying psychotic disorders, we found that while 40% of the co-twins had schizophrenia (which we expected), but also that 8% had had a manic episode (which was unexpected).

Thus we were able to show that schizophrenia and bipolar disorder are not totally genetically discrete. We calculated that probably about half the genetic liability is in common between the two disorders, a liability to psychosis in general. If an individual additionally suffers neurodevelopmental impairment (because he has additional genes that slightly deviate brain development or because he suffers some early hazard such as hypoxia at birth) this pushes the individual towards the schizophrenia end of the continuum of psychosis. On the other hand, if the individual has the predisposition to psychosis but is cognitively very able as a child then he is more likely to end up with the clinical picture typical of the bipolar disorder end of the continuum.

Why did the papers make such an impact?

The papers are highly cited because the Cardno study caused a lot of fuss in the US community. There is a big split between European psychiatry, where there is considerable skepticism about the validity of the diagnostic categories that we use, and American psychiatry, which has tended to believe that the categories in DSM-IV are actual disease entities. Therefore there was a heated debate about this study. American critics asked: could we diagnose schizophrenia correctly—was there a statistical artifact? But basically it was a reluctance of American psychiatrists to believe that there was a genetic relationship between the two conditions. Now, several years on, the link is much more widely accepted, and indeed there are some molecular studies suggesting that some of the risk genes for schizophrenia are also risk genes for bipolar disorder.

The investigations led by Louise Arseneault on cannabis and schizophrenia have also attracted a huge amount of interest (Arseneault L, et al., "Causal association between cannabis and psychosis: examination of the evidence," Brit. J. Psychiat. 184: 110-7, February 2004).

Beginning in the 1990s it was apparent to clinicians like myself that a lot of patients with psychosis smoked cannabis. However, initially (being children of the '60s and '70s) we thought, "Well, poor souls, they have a lot of problems, they are miserable, they are poverty-stricken, they are hearing voices – what's wrong with them having a few puffs?"

Then families started saying to me "Could this have been caused by cannabis? He's been smoking it since he was 13..." Gradually we began to take this question seriously, and cohort studies were initiated, following up adolescents who did and did not smoke cannabis to see how many of each group developed psychosis.

In 2002 the UK government decided to lower the illegality of cannabis, making its use subject to lighter penalties. Their timing was very unfortunate because within months there appeared several papers linking cannabis use by young people with later onset of schizophrenia. The Arseneault paper was one of the first of these. Of course, the relationship is nothing like as strong as that connecting smoking with lung cancer. Nevertheless, the more cannabis you smoke in adolescence the greater the risk of schizophrenia: for the heaviest users the risk increases about five-fold. Thus instead of a risk of under 1%, about 5 in every 100 heavy-using teenagers will develop psychosis.

I think the high citation rates are a partly a result of the UK political parties turning cannabis classification into a political football. Policymakers opposed to the decriminalization of cannabis used our results to support their case, while those in favor pointed out that our studies show that the vast majority of cannabis users come to no harm.

Your group's high-impact papers include striking results on auditory hallucinations, in which patients complain of "hearing voices."

Patients with schizophrenia believe the voices are external. They are adamant that the voices are real. If you put patients in an fMRI scanner and ask them to indicate when they are hearing voices, you see activation of Broca's area where speech is produced. This confirmed Chris Frith's theory that the voices are a misinterpretation of inner speech. There does not seem to be anything wrong with the production of inner speech itself in sufferers. But, in normal individuals the production of inner speech somehow turns off the neural system that normally processes external speech.

In contrast, Suhki Shergill used MRI studies to show that when patients are hallucinating they are activating temporal lobe areas that are normally involved in the processing of external speech (Shergill SS, et al., "Mapping auditory hallucinations in schizophrenia using functional magnetic resonance imaging," Arch. Gen. Psychiat. 57[11]:1033-8, 2000). People with schizophrenia show activation of this auditory processing system when they are hearing voices; no wonder they are convinced the voices are real.

The outstanding question is how do we relate these observations to the neurochemical abnormalities; we do not know that yet. We have the dopamine hypothesis and we have the explanation for the voices, but how is it that an excess of striatal dopamine somehow facilitates the misinterpretation of inner speech?

It's also a striking feature of your list of hot papers that the group has found that obstetric complications are a risk factor for the later development of schizophrenia.

There is indeed an increased risk with pre- and perinatal events. The effect is not large: it's a doubling of risk. My interest goes back to 1978 when I went to Scandinavia for a couple of weeks. In Denmark and Sweden I discovered that there was a lot of research on the role of obstetric complications leading to schizophrenia. This was all totally new to me.

The 2002 paper led by Mary Cannon (Cannon M, Jones PB, Murray RM, "Obstetric complications and schizophrenia: historical and meta-analytic review," Amer. J. Psychiat. 159[7]: 1080-92, July 2002) reviewed all the sound epidemiological studies, and confirmed that there is a risk-increasing effect of pre- and perinatal events: for example, mother having had a severe viral infection, or the baby being small for its gestational age (which would indicate nutritional problems), or hypoxia at birth. The factor that is best established is hypoxia: babies who suffered oxygen starvation at birth tend to have small hippocampi. This is postulated to lead to a dopamine system that is vulnerable to becoming over-activated after puberty.

Are we getting closer to a scientific understanding of schizophrenia?

Schizophrenia is not a total enigma any longer; rather it's like a huge jigsaw puzzle. We have a reasonable idea of the big picture, and we have some sections where the bits fit together, but we do not know how to match up all the different parts and fill in the missing areas.

My view is that the final common pathway is dopamine dysregulation in the striatum. People developing psychosis are releasing excess dopamine. One pathway to this is neurodevelopmental: you either inherit genes that cause subtle impairment of cortical development, or you have some brain insult which results in you having a neural system which is compromised; this renders you prone to dysregulation of dopamine after you go through adolescence. A second pathway is abuse of drugs such as amphetamines, cocaine, and cannabis, which mess up the dopamine system. I believe that a third pathway is through severe chronic social stress, which also can impact on the dopamine system. However, this latter notion is not yet widely accepted.

In summary, we know about the effect of genes and several environmental risk factors. What we do not know is how they interact or what are the intervening mechanisms that link the etiological risk factors to the final common pathway of dopamine dysregulation. These are the critical questions right now.

Professor Robin M. Murray, MD DSc FRCP FRCPsych FMedSci
Institute of Psychiatry at the Maudsley Hospital
King's College London
London, United Kingdom