Robert Lowell, Setting the River on Fire: A Study of Genius, Mania, and Character

by Kay Redfield Jamison

  What once was a field of speculation, case study, and biographical research has splayed out into a teeming field of cognitive and affective psychologists, linguists, neuroscientists, psychiatrists, geneticists, neurobiologists, and epidemiologists. Many studies have been published in recent years, and the number is rising sharply. Not everyone regards this as a good thing. Some have an aesthetic or intellectual aversion to scientific study of the artistic mind, or to work that suggests an association between psychopathology and creativity. It may appear to be reductionist in the highest degree to investigate the relationship of brain circuitry to something so importantly complex as creativity, or setting off on a fool’s journey to study the association between mental illness and creativity across populations of hundreds of thousands of people—but the ship has sailed. Scientists have been doing exactly this, and their research is narrowing in on increasingly specific questions about brain and mind. Such questions are of elemental and broad human importance.

  Skepticism about neuropsychological research and studies that use neuroimaging or genetic techniques remains appropriate; research findings are early, tentative, and inconsistently replicated. But existing methods, and the incrementally more sophisticated methods that will evolve from them, promise a different kind of insight into ancient questions. Skepticism about psychological science is essential and well directed toward the quality of the questions asked, how the findings are interpreted, and what the research means in the broader human context. At the moment, the scientific study of creativity and psychopathology is in an early stage, but the accumulating evidence for a connection between creativity and mood disorders, especially with manic-depressive (bipolar) illness, is proving to be even stronger than previously thought. And Robert Lowell was right: poets are uniquely marked and fretted.

  —

  Five lines of evidence make an increasingly persuasive case for a link between mental illness, especially bipolar disorders, and creativity: biographical studies of individuals that examine the rates of mental disorders in people distinguished by their creative eminence; diagnostic studies of mental disorders in living artists and writers; large population studies that look at the association between psychiatric diagnosis and intelligence, academic performance, and creative occupation; experimental studies that investigate the effect on creativity of cognitive and mood states associated with mania and depression; and, most recently, genetic and brain imaging studies that look at the biological mechanisms common to both creativity and psychopathology.

  Biographical studies investigate mental disorders in eminent writers, composers, and artists by reviewing autobiographical writings, accounts given by contemporaries, correspondence, medical records, journals, and court records. Studies done in this way are, of course, more subjective and involve far fewer individuals than is ideal. Diagnostic accuracy is affected not only by the availability and comprehensiveness of medical records but by the clinical experience and skill of the investigator. There may be bias in determining who is chosen to be included or excluded for study. By its nature, posthumous diagnosis is difficult: evidence of high energy during periods of sleeplessness, exalted mood, and increased productivity may reflect perfectly normal periods of creative excitement or they may be indicative of mania. They may overlap as the illness takes hold. Alcoholism, drug addiction, and medical conditions such as thyroid disease or seizure disorders may present as symptoms of mania or depression. If mildly elevated mood states have not been severe enough to be observed by other people, then major depression rather than bipolar illness may be incorrectly diagnosed. Still, enough is known about specific and defining characteristics of mania and depression—their symptoms; their course (the age at the onset of illness, a worsening of illness over time, seasonal patterns); a family history of mania, depression, or suicide; their close association with alcohol and drug abuse; the implication of a mood disorder in most suicides—to make it possible for experts to do meaningful biographical research. The caveats are many, but it is important to note the results.

  The first biographical studies, carried out in the late 1800s, found improbably high rates of insanity in those defined as artistic geniuses. By the middle of the twentieth century, studies of mental illness in eminent artists and writers had become more refined in their diagnostic techniques and clearer about the criteria they used to select the artists and writers to be studied. The studies remained problematic but less so than they had been.

  Despite the difficulties in research design and the wide variety of methods used, the findings are surprisingly consistent. They find a much higher rate of psychosis, usually mania, psychiatric hospitalization, depression, and suicide in writers and artists than in the general population. (Epidemiologic studies of the more severe form of bipolar illness, bipolar I, find rates of 0.6 percent in the general population and 0.4 percent for the less severe form, bipolar II. An estimated one person in a hundred, then, has bipolar illness. Individuals with milder forms of the disorder, those who do not reach the threshold for a diagnosis of bipolar illness, make up another 1 to 2 percent of the population.) This is true whether the research is conducted in the United States, Hungary, England, or France and whether the artists are painters, poets, or jazz musicians. Among types of artists, there is consistency: poets are the most likely to have a history of mania and the most likely to kill themselves. None were diagnosed with schizophrenia.

  High rates of mood disorders have been found not only in biographical studies but in investigations of living artists and writers as well. In 1987, Nancy Andreasen, a psychiatrist at the University of Iowa, used standardized diagnostic interviews to study thirty writers and thirty control subjects (individuals who were matched for education, gender, and age but whose professional work was not in the arts). The writers were participants in the Iowa Writers’ Workshop, some but not most of whom were nationally acclaimed. Dr. Andreasen’s primary research specialization is schizophrenia, a disease characterized by psychotic breaks but not by mania or bipolar ranges in mood. She expected that she would find a correlation between schizophrenia and creativity; that is, that the writers would have relatively high rates of schizophrenia. She found instead that fully 80 percent of the writers met the diagnostic criteria for a mood disorder; most strikingly, nearly one-half of the writers met the criteria for bipolar disorder. Indeed, writers were more than ten times as likely as the general population to be diagnosed with bipolar I disorder, the more severe form of the illness. They were also more likely to kill themselves. None were diagnosed with schizophrenia.

  Biographical Studies of Depression, Mania, and Suicide in Eminent Writers, Composers, and Artists

  Credit 35

  Adapted from K. R. Jamison, “Creativity in Manic-Depressive illness,” in F. K. Goodwin and K. R. Jamison, Manic-Depressive Illness: Bipolar Disorders and Recurrent Depression (New York: Oxford University Press, 2007).

  A few years after Andreasen’s study of writers, I published a paper on forty-seven eminent British artists and writers I had studied while on sabbatical leave in England. They were selected on the basis of having won at least one of several major awards in their fields; for example, the painters and sculptors were Royal Academicians or Associates of the Royal Academy, an institution established by King George III in 1768 to honor a limited number of British artists and architects. Literary prizes used as criteria included the Queen’s Gold Medal for Poetry and other prestigious British awards in fiction, nonfiction, and poetry. At the time of my study, one-half of the poets had already been anthologized in The Oxford Book of Twentieth-Century English Verse; since then, most of the remaining poets have been included as well. The playwrights who participated in my study were recipients of the major awards in their field; for example, the New York Drama Critics’ Circle Award, the Tony Award, or the London Evening Standard Theatre Award.

  My research focused not so much on mental illness in these artists and writers as on the influence of mood and seasonal changes on their creative wo
rk, but I also inquired whether they had been treated for depression or mania. A significant percentage, more than one-third, reported that they had. Because most people who meet the diagnostic criteria for mood disorders never seek treatment, this is likely a low estimate of the true rate of depression and mania in this group. Of artists and writers who had been treated for a mood disorder, three-quarters of them had been prescribed lithium or antidepressants and/or admitted to a hospital for psychiatric care. All of those who had been treated for mania were poets. One-half of the poets had been treated for bipolar illness or depression and, of the artistic groups, it was poets who most often reported that they had experienced extended periods of elated mood states. Most of them reported that their intense moods were essential to their creative work.

  Arnold Ludwig, a psychiatrist who had conducted a large biographical study of psychopathology in eminent scientists, artists, writers, and military and civic leaders, then turned to the study of living writers. He compared fifty-nine women writers with fifty-nine women nonwriters who had been matched for age, educational level, and their fathers’ occupational status. He found that the writers in his study were four times more likely than the nonwriters to meet diagnostic criteria for depression, five times more likely to have attempted suicide, and six times more likely to meet diagnostic criteria for mania; these findings were consistent with those from his own earlier biographical research and the studies by Dr. Andreasen and me. These studies, like the biographical ones, are hampered by methodological problems, such as possible inclusion bias and small sample size. They are illustrative and suggestive but far from definitive.

  More recent investigations of creativity, intelligence, leadership, entrepreneurship, and mental disorders have used more sophisticated research designs applied to very large numbers of individuals. Six population studies have been carried out in Sweden and one in Denmark, where the governments maintain comprehensive medical and psychiatric records, as well as information on occupation, intelligence, military service, and educational attainment. Three studies were conducted elsewhere, one each in New Zealand, the United States, and the United Kingdom.

  In the case of studies linking mental disorders to intelligence and other measures of cognitive ability, it emerges that lower childhood intelligence is associated with a higher risk of developing schizophrenia or major depression, but that the relationship is more complicated for bipolar disorder. A study of fifty thousand individuals, published in 2004, found no association between childhood intelligence and the development of bipolar disorder. Two others, one of more than one million men, found that high measured IQ in childhood was significantly associated with subsequent hospitalization for bipolar disorder. At the other extreme, individuals with the lowest IQ were also more likely to develop bipolar illness.

  Population Studies of Intelligence, Achievement, and Creativity in Individuals with Bipolar Disorder

  Credit 36

  In a study of academic performance, as opposed to simple IQ, in more than seven hundred thousand individuals, those sixteen-year-olds who excelled at school, especially in music and language, were four times more likely to be hospitalized later for bipolar disorder than were those who performed at an average level. As was the case for intelligence, not only were those with excellent school performance more likely to develop bipolar disorder but so too were those with the lowest IQ and worst school performance.

  A recent study from Stanford University and Aalborg University in Denmark found a similar split distribution in those with bipolar disorder, between those who exceeded the population performance—in this case, in wages and being an entrepreneur—and those who fell far below it. The researchers, who used Danish registry data from more than 3.4 million individuals, found, on the one hand, that individuals with bipolar disorder earned 43 percent less in wages; on the other hand, they were more likely to be entrepreneurs (for example, they were 33 percent more likely to be incorporated, one measure of entrepreneurial willingness to take risk) and 8 percent more likely to enter the ninetieth percentile of the wage distribution. Access to lithium treatment eliminated the difference between the population wage level and that earned by those with bipolar disorder; it also increased the chances that those with bipolar disorder would enter the ninetieth percentile of the wage distribution. Access to lithium treatment decreased the probability of becoming an entrepreneur, however; the investigators suggest that this may be due to lithium’s effect of decreasing the willingness of individuals with bipolar disorder to take risks in the same way that they would if they were unmedicated. A recent Swedish total population study of leadership found that bipolar patients without comorbidity (that is, without any additional psychiatric diagnosis), and their healthy siblings, demonstrated superior leadership traits. Bipolar patients, but not their healthy siblings, were overrepresented in the lower strata of leadership as well. Healthy siblings of the bipolar patients were more likely to hold executive positions than the general population (particularly in political professions).

  The brain vulnerable to mania and depression, it would seem, can veer to illness from origins of unusual strength or weakness. This is consistent with what we know from clinical experience. Some people become ill with almost no warning, their lives to that point characterized by good academic performance, by good relationships and psychological stability. In others, the onset of illness is more insidious and clinical symptoms develop earlier. Life even before the first manic or depressive break is more troubled, the personal and academic toll more lasting.

  Patients with bipolar illness often show marked deficits in specific types of intellectual functioning, some much more than others. Neuropsychologists have found, for example, that the cognitive abilities most dependent on language are the ones least affected by bipolar disorder. Thus, verbal intelligence—language and verbal skills—tends to be normal or increased in bipolar patients. However, many if not most patients with bipolar illness display deficits in performance intelligence when tested on measures of visuospatial ability, perceptual reasoning, attention, and concentration. Often the deficits in attention and concentration, usually measurable before the first manic attack or depressive episode, are severe.

  Additional characteristics interact with bipolar disorder to determine a patient’s life story. Temperamental, genetic, and environmental factors may protect some patients more than others. Alcohol and drug use and too little sleep make the illness worse. A young developing brain assailed by mania or severe depression is usually more adversely affected than a brain that is older and has a longer history of stability.

  The interplay between an individual’s illness and temperament exerts a decisive role in how those with bipolar disorder ultimately do. It is clear that the prevailing mood during mania—whether paranoid and irritable or expansive and elated—influences the flow and the use of ideas during mania. This is important in creative work. But the difference in an individual’s predominant mood may color the capacity for friendship and marriage as well; exuberance is more attracting and binding than suspiciousness and a short fuse. Certainly the age at which the illness begins and the number, duration, and severity of episodes of mania and depression are critical. Those who become ill when young or who remain untreated are less likely to enter successfully into the stress and scramble of life. And, after an attack of illness, they are less able to reenter life and regenerate what they need to thrive. Bipolar disorder is by no means a unitary diagnostic category. Hippocrates, Aretaeus, and Kraepelin knew this from their patients; modern scientists who study mania and depression provide evidence to buttress their observations. The variability of mania and depression remains constant.

  —

  In addition to the large population studies of the association between bipolar disorder, intelligence, leadership, and academic and entrepreneurial performance, three studies have looked more specifically at the association of bipolar disorder with creativity. In each, the researchers examined the association of the disorder
with creative occupation rather than with creativity itself. They are not the same, but occupation is a pragmatic measure to use when studying very large numbers of individuals. Studies of eminent creators necessarily must involve relatively small numbers of individuals. Each kind of research, whether biographical or population-based, has its limitations and assets. An American study of more than twenty thousand individuals published in 2010 found that those with bipolar illness were disproportionately concentrated in the most creative occupations such as writing, the visual arts, and music. This finding is consistent with two much larger Swedish studies that were published shortly after. The first looked at three hundred thousand patients, their first-degree relatives, and controls, and the second at more than one million individuals; both investigations found that those who had been hospitalized for bipolar illness were overrepresented in the creative professions. This was not true for those who had been hospitalized for schizophrenia or depression. The first-degree relatives of individuals with bipolar disorder or schizophrenia were also more likely than controls to hold creative jobs. In the larger of the studies, the researchers found that writers had a greatly elevated rate of suicide.