It would obviously be very useful to know whether different kinds of life events tend to provoke different complaints. Unfortunately, this question (which is much more complex than the question of whether life events, in general, provoke episodes of psychosis) has rarely been studied. In one investigation, Brown and his colleague Robert Finlay-Jones found that events associated with loss (broadly defined to include not only the loss of a person but also the loss of a role, goal or cherished idea) were especially likely to lead to depression, whereas events associated with long-term threat were more likely to lead to anxiety.60 In later studies it was found that depression was also likely to follow humiliating events (for example, being deserted by a partner) or events that created a sense of entrapment (for example, the onset of a debilitating physical illness). It was also found that life events were particularly likely to have an emotional impact if they matched an ongoing difficulty (for example, a mother’s discovery that her son has been arrested, matching her ongoing concerns about their difficult relationship and his delinquent behaviour).61
As we saw in Chapter 12, Brown’s colleague Tirril Harris has suggested that the events most likely to lead to psychotic and especially paranoid symptoms can be characterized as intrusive, because they involve unwanted experiences being forced on the patient (examples include threats, burglaries and police investigations).62 This suggestion has been partially supported by research conducted by other investi-gators.63 In the case of mania, as we saw in Chapter 11, it appears that episodes of illness are often precipitated by events that disrupt social rhythms64 or which involve goal attainment.65 Each of these effects seems to be relatively easy to understand in the context of our knowledge of the psychological mechanisms involved in these kinds of symptoms. Those which bring to mind a sense of loss or failure are most dangerous to people who already doubt their self-worth; those with a persecutory element appear to bring out pre-existing fears about others’ intentions; and those which exacerbate subjective feelings of energy and excitement are particularly hazardous to people who are vulnerable to mania.
Despite these findings, some researchers have been sceptical about the link between life events and psychotic episodes. In part, this has been because they have been influenced by genetic evidence, which has suggested to them that there is little room for environmental influences on patients’ symptoms (a prejudice that I will address in the next chapter). Whereas I have tried to point to the overall pattern of the findings from many research projects, critics have focused on the methodological inadequacies of some investigations and some inconsistencies in the results from different studies. As a counter to these kinds of arguments, it is worth pointing out that there are at least two reasons to think that existing findings underestimate the role of life events in psychosis.
First, as Paul Bebbington has noted, some of the apparent inconsistencies in the research findings may reflect the extreme sensitivity of psychotic patients.66 Perhaps they are so easily affected by adverse experiences that even relatively trivial events affect them. Consistent with this suggestion, studies have shown a clearer link between life events and psychosis in patients who take anti-psychotic drugs or who live in low EE households, than in patients who refuse drugs or live in high EE environments. The implication of these findings is that even very minor events – which would not be picked up by the methods developed by Brown and Harris – may be capable of triggering symptoms in patients who refuse medication or are embroiled in stressful relationships.
Second, because most researchers have deliberately ignored life events that could be the consequence of symptoms, they have failed to record circumstances in which life events and symptoms exacerbate each other in an escalating spiral of psychopathology. Imagine a remitted patient who is not coping well with life, becomes depressed and, as a consequence, loses his job. Later, the patient becomes floridly psychotic. In this case, the loss of the job is clearly not independent of the patient’s prodromal symptoms, so it would not be counted in a study of life events. Nonetheless, sudden unemployment will almost certainly have been a further source of distress and may well have played a role in moving the patient from a prodromal phase into frank psychosis.
The Process of Relapse
Stressful relationships and life events do not exhaust the factors that can increase the likelihood that a person will once again experience psychotic complaints. There is evidence that the excessive use of either legal or street drugs, such as alcohol, amphetamine, cannabis or the hallucinogens, can increase the probability of relapse in both schizophrenia patients67 and patients with a diagnosis of bipolar disorder.68 Unfortunately, in Europe and North America at least, psychotic patients very commonly abuse alcohol or take drugs, partly because these substances often make them feel better in the short term.69
Keith Nuechterlein and his colleagues at UCLA have proposed an influential model of schizophrenic relapse, which is shown in Figure 16.7.70 Although at first glance it appears very complex, it encompasses just a few simple ideas. Nuechterlein argues that the risk of a psychotic episode is affected by a combination of personal and environmental factors. The personal factors include those (such as schizotypal personality traits and putative abnormalities in the dopamine system) that increase an individual’s likelihood of experiencing complaints, and also protective factors (such as coping skills and taking medication), which have the opposite effect. The environmental factors are similarly divided into those that protect (a supportive family, for example) and those that increase the risk of symptoms (stressful life events, for example). According to the model, some interactions between these factors can exacerbate existing cognitive deficits, leading the patient’s cognitive system to become overloaded and also to increased emotional arousal. These in turn are thought to lead to prodromal symptoms and then, eventually, to the full symptoms of a psychotic episode. Importantly, the UCLA group propose that the experience of an episode can feed back to change the earlier personal and environmental factors, so that further episodes are more likely (a sort of kindling effect).
The UCLA group has marshalled a lot of evidence for some of the core assumptions of their model. As I explained in Chapter 7, they have carried out experiments to show that working memory deficits become more severe as patients become ill. However, notice that, despite its apparent complexity, the model remains very general in its scope. In fact, although it is proposed as a theory of schizophrenic relapse, with a few modifications it could equally well serve as an account of relapse in depression or mania. This is because the main weakness of the model is that it does not explain how cognitive changes lead to specific complaints such as hallucinations, delusions or mania. What appears to be missing is any reference to the social-cognitive processes that play a role in prodromal dysphoria, and which we have directly implicated in these symptoms.
It is easy to see how the model might be tinkered with to overcome some of these limitations, and I will leave this task to the reader. Of course, as different cognitive processes are implicated in different complaints, the protective and vulnerability factors associated with them are also likely to differ. If this is the case, a truly comprehensive model of psychotic relapse will consist of a series of inter-related and more specific models, each specifying pathways involved in the exacerbation of one particular type of psychotic behaviour or experience.
Figure 16.7 The UCLA model of psychotic relapse (from Nuechterle in et al., 1992).
17
From the Cradle to the Clinic
The childhood shows the man,
As morning shows the day.
John Milton1
Any attempt to explain the origins of madness must address an important puzzle. Since Kraepelin’s time, it has been known that schizophrenia symptoms most often appear for the first time during late adolescence or early adulthood. It was for this reason that Kraepelin christened his disorder dementia praecox. Modern studies have generally supported this observation. Figure 17.1 shows a fairly typical age d
istribution for the onset of psychosis, which I have obtained from a sample of 254 first-episode schizophrenia or schizoaffective patients who took part in a clinical trial of psychological treatment for early psychosis carried out at my own University.2 It is obvious from the graph that most people became ill in their early twenties; the age of maximum risk seems to be about 22 years. However, notice also that a few patients were in their sixties when they were first affected, so that the mean (average) age of onset was exactly 29 years. However, a better indication of the typical age of onset is the median (the age at which half of the patients are younger and half are older), which is 26.5 years. When the data are broken down by sex, we find that the median age of onset for women (29.3 years) is a few years older than that for men (25.6 years). This earlier onset for men than for women has been found in nearly every other study that has examined this issue.3
It is sometimes forgotten that diagnoses of bipolar disorder follow similar age trends. The period of peak risk for developing bipolar
Figure 17.1 Ages of onset of illness for 254 first-episode schizophrenia and schizoaffective patients recruited to the SoCRATES (Study of Cognitive-Realignment Therapy in Early Schizophrenia) project.
symptoms is during adolescence and early adulthood and the distribution of ages of onset is very similar to that of schizophrenia patients; interestingly, there is evidence that those who first show symptoms early in life are especially likely to become floridly psychotic.4 There is also some evidence that women are slightly more likely than men to develop bipolar symptoms later in life, although the data pertaining to this difference are relatively sparse and not entirely consistent.5
These observations raise the possibility that psychosis is the end-point of a developmental process. But what kind of process? Once again, the ground has been staked out by biological researchers, who have proposed various versions of what I term the biological time-bomb hypothesis. According to this theory, schizophrenia and bipolar disorder are consequences of defects in the developing brain, which only ‘explode’ into frank illness as or after the afflicted person reaches maturity. The exact nature of the hypothesized time-bomb varies from theorist to theorist, and includes genetically determined neurodevelopmental deficits,6 faulty brain maturation,7 and late manifestations of a virus contracted by the pregnant mother.8 In a twist to this tale, it has been suggested that the average age at which women develop schizophrenia symptoms is later than the average age for men because the female hormone oestrogen confers protection against psychosis (apparently, oestrogen shares some of the pharmacological properties of the neuroleptics, the class of drugs most widely used in the treatment of psychotic patients).9
Clearly, a developmental approach has much to recommend it, but it is pretty obvious that there is something missing from any account that portrays madness exclusively as the product of processes unwinding within the individual. What is missing is any recognition that psychosis might be influenced by the trials of life.
Rethinking the Genetics of Psychosis
Writing in 1939, Robert Gaupp, one-time Professor of Psychiatry at the University of Tübingen, made the following comment:
We are, of course, clearly aware of the fact, which we don’t deny even for a second, that the greater part of all genetic work in psychiatry would immediately collapse like a house of cards if Kraepelin’s theory was shown to be altogether mistaken.10
In Chapter 4, I discussed family, twin and adoption studies that have pointed to the role of genetic factors in the origins of psychosis. I suggested that genetic researchers have often exaggerated the weight of the evidence obtained from these kinds of studies, and have tended to ignore serious methodological weaknesses in their own work. I think this observation would be widely accepted by many fellow psychologists and psychiatrists who are familiar with the data. However, what is less widely recognized is that even if (and it is a big if) very substantial genetic contributions to psychosis could be proven, this proof would not preclude the possibility that the environment also plays a major role in the development of symptoms.
Geneticists commonly use the data from their studies to calculate heritability estimates, which describe the amount of variability in some phenomenon (say, intelligence or mental illness) that can be attributed to genes.* Typically, researchers investigating schizophrenia or bipolar symptoms obtain estimates that are very high – as much as 80 or 90 per cent. Values of this sort seem to leave little room for environmental influences. However, this conclusion is based on the assumption that the amount of variability attributed to genes and that attributed to the environment are independent of each other, and therefore can be simply added to total 100 per cent. In real life, genes and environments are almost always correlated, so that genetic and environmental influences cannot be separated out in this way. This happens because individuals (not only human beings, but even plants and small animals)11 influence their environments in ways that depend on their genetic endowment. When this happens, environmental influences can bring about dramatic changes in behaviours or abilities, even when the behaviours and abilities are under strong genetic control.
A formal mathematical analysis of this kind of interaction† has recently been published by William Dickins of Harvard University and James Flynn of the University of Otago.12 However, rather than attempting to summarize their difficult proof here, I will offer a hypothetical case study that may help the reader to understand how gene–environment correlations can have this effect. Imagine that a young man, who is destined to become psychotic by virtue of inheriting genes that make him vulnerable to negative symptoms, has poor social skills. Such a person might be more likely than most to be rejected when asking a potential girlfriend for a date. Moreover, because he does not know how to seek emotional support from friends, or because his parents (who also have poor social skills) do not know how to reassure him, he may become more socially anxious and self-conscious afterwards. Consequently, he may become withdrawn, making him more likely to be ridiculed by unkind peers. As time goes by, he may become progressively more isolated until, perhaps reacting to some further difficulties, he becomes psychotic. In this imaginary, but not implausible, scenario, genes have predisposed the unlucky young man to an environmental trauma (being rejected by a potential girlfriend), but the trauma has had a major impact on his future life. Had he decided to approach the one girl in his neighbourhood who was attracted to quiet young men, the outcome might have been very different.
Evidence that gene–environment interactions of this sort happen in psychiatric conditions has emerged from genetic studies of depression, which have shown that the close relatives of depressed patients experience a higher rate of unpleasant life events than the general population.13 Presumably because of their genetically determined characteristics they are more likely to find themselves in circumstances in which these kinds of events are likely to occur. If these kinds of effects occur across the spectrum of patients suffering from severe psychiatric disorders (and I would be amazed if they did not) environmental events may have a much greater ability to influence psychosis than most psychiatric geneticists have supposed.
From genes to madness
Perhaps the central argument of this book is that the problems involved in explaining ‘schizophrenia’ and ‘bipolar disorder’ will disappear once we have adequately explained the complaints that lead to these diagnoses. An implication of this argument is that we must consider genetic influences on particular kinds of psychotic behaviours and experiences. This task is especially important because, in their critique of the symptom-orientated approach, Ramin Mojtabai and Ronald Rieder have argued that no such influences can be discerned.14
According to Mojtabai and Rieder, twin and family studies show that a diagnosis of schizophrenia is more heritable than any of the symptoms that contribute to the diagnosis. American psychologists Howard Berenbaum, Tom Oltmanns and Irving Gottesman analysed data from a twin study to show that the concordance rates for positive or negativ
e symptoms were lower than the concordance rate calculated for a global diagnosis.15 In a more recent study of ‘high density schizophrenia families’ living in Ireland, conducted by American researcher Kenneth Kendler and his colleagues, sibling pairs in which both had received a diagnosis of schizophrenia were examined to see whether they experienced similar symptoms; in general the concordance between the symptoms of the affected pairs were low.16 According to Mojtabai and Rieder, these observations are impossible to square with a symptom-orientated approach.
It is worth unpacking Mojtabai and Rieder’s argument, using the functional analysis framework that I introduced in the last chapter. They have implicitly contrasted two models, which I have shown in Figure 17.2. According to their preferred model (which is a version of the Kraepelinian paradigm), genes influence the development of a disease, which in turn causes the affected individual to experience symptoms. The alternative model supposes that genes influence symptoms, without the mediating disease step. Mojtabai and Rieder argue that this alternative model must be rejected because the empirical evidence shows that genes have only a relatively weak effect on each of the symptoms but a strong impact on the overall diagnosis.
In fact, there is evidence of a specific genetic contribution to at least one psychotic symptom. It has been known for some time that thought disorder runs in families. In the 1960s, American researchers Margaret Singer and Lyman Wynne carried out a series of studies in an attempt to show that this happened because thought-disordered parents communicate inadequately with their children. (In essence, they proposed that thought disorder is learnt.)17 The comments of patients’ parents were recorded as they completed a series of psychological tests, which were also administered to the patients.18 A detailed rating procedure was then used to analyse the comments for unusual styles of communication (particularly, abnormal ways of handling attention and meaning which might ‘impair a growing child’s capacity for selective attention, purposive behaviour, and subjectively meaningful experiences’). They also rated unusual ways of talking about relationships, abnormal emotional responses (especially comments thought to reflect ‘underlying feelings of pervasive meaninglessness’), and statements suggesting that
Madness Explained Page 50