BIPOLAR DISORDER AND CREATIVE POTENTIAL 1

The Manic Idea Creator? A Review and Meta-Analysis of the Relationship Between

Bipolar Disorder and Creative Cognitive Potential

Karin Kazcykowski1, Boris Forthmann2, Mathias Benedek3, and Heinz Holling1

1Institute of Psychology, University of Münster

2Institute of Psychology in Education, University of Münster

3Institute of Psychology, University of Graz

Author Note

Boris Forthmann https://orcid.org/0000-0001-9755-7304

Mathias Benedek https://orcid.org/0000-0001-6258-4476

Heinz Holling https://orcid.org/0000-0002-0311-3970

We would like to thank Celeste Brennecka for proofreading of an earlier version of this work.

We have no known conflict of interest to disclose.

Correspondence concerning this article should be addressed to Boris Forthmann,

Institute of Psychology in Education, University of Münster, Fliednerstrasse 21, 48149

Münster, Germany. Email: [email protected] BIPOLAR DISORDER AND CREATIVE POTENTIAL

Abstract

Even though a relationship between psychopathology and creativity has been postulated since the time of ancient Greece, systematic meta-analyses on this topic are still scarce. Thus, the meta-analysis described here can be considered the first to date that specifically focuses on the relationship between creative potential, as measured by divergent thinking, and bipolar disorder, as opposed to psychopathology in general. An extensive literature search of 4,670 screened hits identified 13 suitable studies, including a total of 42 effect sizes and 1,857 participants. The random-effects-model showed an overall significant, positive, yet diminutively small effect (d = 0.11, 95% CI: [0.002, 0.209], p = .045) between divergent thinking and bipolar disorder. A handful of moderators were examined, which revealed a significant moderating effect for bipolar status, as either euthymic, subclinical, manic, or depressed. We discuss further results, especially regarding the differences between subclinical and clinical samples.

Keywords: bipolar disorder, creativity, divergent thinking, meta-analysis BIPOLAR DISORDER AND CREATIVE POTENTIAL

The Manic Idea Creator? A Review and Meta-Analysis of the Relationship Between

Bipolar Disorder and Creative Cognitive Potential

Since the time of the ancient Greece philosophers, the idea of the mad genius has insinuated a connection between psychopathology and creativity; for most laypeople, this image still exists today (Kaufman, Bromley, & Cole, 2007). This putative association between psychopathology and creativity has also been the subject of empirical psychological research. However, most studies on this topic have examined the link between creativity and psychopathology in general rather than focusing on the link between creativity and a specific disorder (Silvia & Kimbrel, 2010). Furthermore, studies have not conceptualized creativity in a homogeneous way, as most have focused on creative achievement, but others on creative professions, or creative personality, and only few measured creativity with established assessments of creative cognitive potential (Thys, Sabbe, & De Hert, 2014).

Studies that have focused on the specific link between bipolar disorder and creative potential, so far, revealed inconsistent findings: While some studies found a positive relation between bipolar disorder and creative potential (e.g. Santosa et al., 2007; Simeonova, Chang,

Strong, & Ketter, 2005), some found no relation between the two constructs (e.g. Siegel &

Bugg, 2016; Von Stumm, Chung, & Furnham, 2011), and other studies even indicated a negative relationship between bipolar disorder and creative potential (e.g. Alici, Özgüven,

Kale, Yenihayat, & Baskak, 2019). To clarify the relationship between bipolar disorder and creative potential, this meta-analysis considered all relevant data and took relevant moderators into account that might explain these contradicting results.

Creativity, Creative Potential, and Divergent Thinking

Since creativity is multi-layered and complex concept, definitions of creativity usually focus on specific aspects of the construct. In the Standard Definition of Creativity, Runco and

Jaeger (2012) noted that finding a standard definition of creativity is not easy, partly due to BIPOLAR DISORDER AND CREATIVE POTENTIAL historical changes of the term. Interestingly, one of the most extensive and precise definitions of creativity was given as far back as 1956 by Drevdahl, who summarized creativity as the following: “Creativity is the capacity of persons to produce compositions, products, or ideas of any sort which are essentially new or novel, and previously unknown to the producer.”

(Drevdahl, 1956, p. 22). As an extension of this definition, Sternberg (2012) suggested in his investment theory of creativity that creativity is not only a capacity but in fact a habit that leads creative people to routinely approach problems in a novel way.

It has further been proposed to divide creativity into different domains or facets, such as the hierarchical six-P framework by Runco (2007), extending the 4-P model by Rhodes

(1961), who fundamentally differentiated between creative performance (further comprising product, persuasion, and interaction between person and environment) and creative potential

(further comprising person, process, and press). Creative performance is what laypeople generally understand as creativity: a creative profession, creative achievement, even reaching the level of a creative genius, as exemplified by famous creative individuals such as Vincent van Gogh, Robert Schumann, Ernest Hemmingway, or Virginia Woolf. Additionally, creative performance can further be separated into different areas of performance, such as artistic creativity in fields of art, music, and literature, or scientific creativity in scientific fields

(Feist, 1998). However, because not all creative people have (yet) reached the point of actually behaving or performing creatively, one must also take into account the potential to have creative ideas. Creative potential does not involve actual creative execution but rather examines a person’s potential to produce new and original ideas that may (but does not have to) lead to creative performance (Runco, 2007).

More contemporary approaches that are in line with separating creative performance and potential distinguish between Big-C, Pro-C, little-c and mini-c creativity (Beghetto &

Kaufman, 2007; Kaufman & Beghetto, 2009): while Big-C creativity is seen as creativity on BIPOLAR DISORDER AND CREATIVE POTENTIAL a groundbreaking level that is able to change a field, Pro-C creativity refers to skilled creative actions that make a contribution to a field without making a revolutionary change, little-c creativity describes everyday creativity in laypeople, and mini-c creativity refers to someone’s first creative spark, such as using creative learning processes in experiences, events, and actions. These four levels of creativity are directly connected, in that mini-c creativity may lead to little-c creativity, which in turn may lead to Pro-C creativity, which might eventually result in Big-C creativity (Beghetto & Kaufman, 2007).

Creative potential can be seen on a par with little-c, everyday creativity; this is also the most important domain for investigating creativity in the general population, as opposed to studying selective creative samples, which occurs when sampling from creative fields or even from the small pool of famous creative people. Since this meta-analysis aimed to study everyday creativity, the main variable of interest were assessments of creative potential.

The most commonly used way to measure creative potential is through divergent thinking (DT; Runco & Acar, 2012) tasks, which are open-ended tasks that require participants to generate ideas. DT performance can be scored for different aspects such as fluency, flexibility, originality, and elaboration (Guilford, 1950, 1957; Guilford & Hoepfner,

1971). Fluency can be understood as the number of different ideas or solutions generated to address a specific problem; flexibility relates to the number of different conceptual categories from which ideas were generated; originality can be defined as remoteness, novelty, and inventiveness of ideas; and elaboration describes the detailedness of produced ideas

(Guilford, 1950, 1957; Guilford & Hoepfner, 1971). These facets of DT are not independent but are intertwined, since generating a greater number of ideas, reflecting greater fluency, enhances the likelihood of generating more unique ideas, thus reflecting more originality (e.g.

Plucker, Qian, & Wang, 2011). Moreover, the relationship between facets depends on how scores are aggregated across responses: Summative scorings, which add scores across all BIPOLAR DISORDER AND CREATIVE POTENTIAL responses, result in huge correlations between fluency and other facets (i.e., fluency confound; Hocevar, 1979; Dumas & Dunbar, 2014; Forthmann, Szardenings, & Holling,

2020), whereas discriminant validity of facets can be maintained when scorings are independent of fluency. This can be achieved by dividing the number of original ideas by the total number of ideas (i.e., average scoring; Runco & Mraz, 1992; Yamamoto, 1964), or by scoring subsets of ideas rather than scoring every idea individually as in the snapshot scoring

(Silvia et al., 2009) or top scoring (Benedek, Mühlmann, Jauk, & Neubauer, 2013; Silvia et al., 2008). Common examples of DT tasks include the Alternate Uses Test (AUT; Guilford,

1967), and the Torrance Test of Creative Thinking (TTCT; Torrance, 1990, 1998, 2008).

These tests further intend to measure DT on various modalities such as the verbal, figural, or numeric content modality (Frasier, 1988).

In addition, time-on-task has been shown to influence performance in DT task (Said-

Metwaly et al., 2020). More lenient time limits imply more responses as well as more original responses. This can be explained by the serial order effect which refers to the observation that quality of responses is a function of time during testing (Beaty & Silvia, 2012; Kleinkorres et al., 2021). Typically, high-quality responses are generated at later stages of the idea generation process. In addition, rather strict time limits may introduce task speededness which potentially affects the measured construct (Forthmann, Lips et al., 2020; Preckel et al.,

2011). Another relevant factor for assessing creativity relates to the type of instruction participants are given to perform the DT task, since participants can be instructed to generate as many ideas as possible, only original ideas, or only appropriate ideas (Forthmann et al.,

2016; Runco, Illies, & Eisenman, 2005; Runco & Okuda, 1991). Different instructions, therefore, emphasize different DT facets, either stressing the quantity or quality of ideas, which influences DT performance (Forthmann et al., 2016). BIPOLAR DISORDER AND CREATIVE POTENTIAL

As demonstrated above, assessing creativity is complex and involves a variety of specifications. For that reason, the meta-analysis at hand will also include variables around the assessment of DT, like type of DT task (e.g. AUT, TTCT, other.), the DT facet that was scored (fluency, flexibility, originality, elaboration, total), the creative modality domain

(figural, verbal, numeric, mixed), the available time-on-task, the method of calculating the creativity score (summed, averaged, other), and the type of instruction (quantity, quality).

Bipolar Disorder

Bipolar Disorder (BD), also known as manic-depressive disorder, is a mood disorder involving extreme mood swings between depressive and elevated moods (American

Psychiatric Association [APA], 2013). Interestingly, in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), the chapter about bipolar and related disorders is placed between chapters about schizophrenia (and other psychotic disorders) and the chapter on depressive disorders. This already hints that BD is more multifaceted than a simple mood disorder and instead is considered a “bridge between the two diagnostic classes in terms of symptomatology, family history, and genetics” (APA, 2013, p. 123). Further, bipolar and related disorders is generally referred to in the plural rather than in the singular since their presentations are manifold. To be precise, the DSM-5 differentiates between bipolar I, bipolar II and cyclothymic disorder, varying in different lengths and intensities of affective episodes (APA, 2013). Regarding affective episodes, one can distinguish between depressive, manic, and hypomanic episodes.

A depressive episode is characterized by a distinct period of at least two weeks with either depressed mood and/or loss of interest, pleasure or both, along with accompanying symptoms (APA, 2013). Depressive mood states are usually accompanied by listlessness, ruminative thinking and hopelessness, manifesting itself in behaviours such as having trouble getting out of bed or engaging in activities (APA, 2013; MacGill, 2019). Manic and BIPOLAR DISORDER AND CREATIVE POTENTIAL hypomanic episodes can both be described as “a distinct period of abnormally and persistently elevated, expansive, or irritable mood” (APA, 2013, p.124), in which at least three (or four, if the mood is not expansive but only irritable) of seven criteria must be met

(for a full description of diagnostic criteria, see APA, 2013). The typical symptoms are inflated self-esteem, decreased need for sleep, talkativeness, racing thoughts, distractibility, increased psychomotor agitation, and high-risk activities (APA, 2013; MacGill, 2019). To name some examples, typical behaviours during a (hypo)manic episode are, for instance, seeking high-risk situations such as disproportional speeding in traffic, irresponsible sexual contacts, and unaccountable financial expenses (APA, 2013; MacGill, 2019). The difference between a hypomanic and a manic episode is the length of an episode as well as the level of functioning: A manic episode has to last at least one week, while a hypomanic episode lasts at least four days; and while a manic episode usually comes along with severe social or occupational impairment, such as hospitalization, harm to self or others, or psychotic symptoms, the functioning level in a hypomanic episode is not severe enough for hospitalization (APA, 2013).

To fulfill the diagnostic criteria for bipolar I disorder, the criteria for a manic episode must be fulfilled at least once during a person’s lifetime (APA, 2013; bipolar I, criterion A), independent of the prior existence of a depressive episode. To fulfill the diagnostic criteria for bipolar II disorder, the criteria for at least one hypomanic episode as well as for at least one depressive episode must be met during a person’s lifetime (APA, 2013; bipolar II, criterion A), without ever meeting the criteria for a manic episode (APA, 2013; bipolar II, criterion B). For both, bipolar I and bipolar II, the (hypo)manic and/or depressive episode(s) cannot be better explained by other disorders, such as schizoaffective, schizophreniform, delusional or other specified or unspecified disorders from the schizophrenia spectrum. While the manic episodes often cause the most significant impairment in individuals with bipolar I BIPOLAR DISORDER AND CREATIVE POTENTIAL disorder, for individuals with bipolar II, the depressive episodes or the unexpected switches between episodes are usually stated as the most constricting, and they would rarely consider hypomanic symptoms as a reason for treatment (APA, 2013).

Altogether, it is important to note that bipolar II cannot simply be seen as a milder form of bipolar I disorder; there are some general differences regarding symptomatology.

Namely, not only do individuals with bipolar II disorder a stronger tendency to experience chronic progression, but individuals with bipolar II disorder also more typically suffer from hypomanic and depressive symptoms at the same time, so called mixed features (APA, 2013;

Benazzi & Akiskal, 2001; Koukopoulos & Sani, 2014). It is commonly reported within bipolar II diagnoses that mixed features of both episodes can manifest themselves in a form of depression with increased energy and irritability; this can also occur in bipolar I, but much more rarely (APA, 2013; Koukopoulos & Sani, 2014).

Usually, individuals with bipolar I or II disorder experience phases without a

(hypo)manic or depressive episode and, in between episodes, return to their euthymic baseline level of mood (APA, 2013). If this baseline level is never met, yet hypomanic and depressive symptoms persist, cyclothymic disorder is taken into consideration. To meet the criteria for cyclothymia, there must be numerous periods with hypomanic and depressive symptoms that never meet the criteria for a full-blown hypomanic or major depressive episode within the course of at least two years (APA, 2013; cyclothymia, criterion A).

Furthermore, the depressive and hypomanic periods must be present at least 50% percent of the time during the two-year period without a phase in which there were no symptoms for longer than two months at a time (APA, 2013; cyclothymia, criterion B).

In addition, there is another important specification regarding the switching between episodes: If there are four or more mood episodes (hypomanic, manic, or depressive), the specifier applied is rapid cycling, and when a rhythm of 48-hour switches occurs, the BIPOLAR DISORDER AND CREATIVE POTENTIAL specifier applied is ultra rapid cycling (APA, 2013; Kupka, Luckenbaugh, Post, Leverich, &

Nolen, 2003).

Taking all various types of bipolar and related disorders together, the lifetime- prevalence can be estimated at around 3.9 - 4.4% (APA, 2013; Merikangas et al., 2011). As in most psychiatric diagnoses, risk factors are not univocal but originate from a combination of biological, social, and psychological factors (APA, 2013; Alloy et al., 2005). However, family history of BD seems to be one of the greatest risk factors, resulting in a 10- to 15- times higher risk to fall ill with BD if relatives are known to suffer from it (APA, 2013;

Pavuluri, Henry, Nadimpalli, O’Connor, & Sweeney, 2006).

Furthermore, in BD, comorbidities are the norm rather than the exception: Of those with a bipolar diagnoses, 60% to 75% have at least one other co-occurring mental disorder

(APA, 2013; Krishnan, 2005). By far the most common comorbidities are anxiety disorders

(75%), substance abuse (37%) and eating disorders (14%). For these comorbidities, episodic differences can also be observed, since anxiety or eating disorder symptoms are more commonly related to depressive episodes, while substance abuse is more commonly linked to

(hypo)manic episodes (APA, 2013; Krishnan, 2005). Besides comorbidities, the risk for suicidal actions is heightened as well. Nearly one third of individuals with bipolar I or II disorder have a history of suicide attempts during their lifetime, and 17% to 24%, complete suicide (APA, 2013; Rihmer & Kiss, 2002).

For this reason, appropriate treatment is not only essential but also lifesaving. Usually, a combination of psychotherapy and medication is used, showing good remission rates (Scott,

Colom, & Vieta, 2007). In many countries, the main medication used to treat BD is lithium

(APA, 2013; MacGill, 2019), followed by other so-called mood stabilizers or antipsychotic medicines (APA, 2013; Sachs, Printz, Kahn, Carpenter, & Docherty, 2000). Even though unipolar depression and BD-related depression episodes might manifest themselves very BIPOLAR DISORDER AND CREATIVE POTENTIAL similarly, the medication still fundamentally differs, because antidepressants that are given for unipolar depression might trigger a manic episode in BD (APA, 2013; Henry, Sorbara,

Lacoste, Gindre, & Leboyer, 2001).

Interestingly, a paragraph about the relationship between BD and creativity can be found in the DSM-5:

There may be heightened levels of creativity in some individuals with a bipolar

disorder. However, that relationship may be nonlinear; that is, greater lifetime creative

accomplishments have been associated with milder forms of bipolar disorder, and

higher creativity has been found in unaffected family members. The individual's

attachment to heightened creativity during hypomanic episodes may contribute to

ambivalence about seeking treatment or undermine adherence to treatment (APA,

2013, p.136).

Notably, this is the only paragraph in the whole DSM-5 that deals with the topic of creativity, thus hinting at BD’s special significance in this regard. Also, as some bipolar patients might not undergo treatment due to the fear of, in a sense, losing their heightened creativity, the nature of this relationship must be clarified, which will help to reduce prejudice towards treatment and encourage everybody who needs treatment to seek it.

Because of the high prevalence rates and the high risk for suicidal attempts and suicides, appropriate diagnosis and treatment of BD are essential, in both psychological research and practice. When taking a closer look on the specifications of BD, the complexity of the disorder as well as the influencing factors become evident. For that reason, the meta- analysis conducted here aimed to include all the variables intertwined with BD. To cope with the diversity of specifications around BD, the moderators examined in this meta-analysis include specific type of BD (e.g., bipolar I, mixed sample of bipolar I and bipolar II, or subclinical BD), current episode (euthymic, depressed, hypomanic, manic), diagnostic BIPOLAR DISORDER AND CREATIVE POTENTIAL assessment of the disorder, comorbidities, duration, onset, hospitalization, number of episodes in the past, as well as medication.

Mad Genius Theory

“No great genius has existed without a strain of madness” (Seneca the Younger, as cited in Clark & de la Motto, 1992; p. 193). This famous quote by Seneca the Younger illustrates that the idea of a link between mental illness and creativity is not a modern phenomenon but even existed in ancient societies such as in ancient Greece. Apart from

Seneca, other Greek philosophers also promoted a mad genius link by suggesting that creative individuals like philosophers, poets, or politicians all seem to have suffered from what Aristotle called melancholia and Plato called divine madness (Carson, 2011).

When examining the biographies of famous creative individuals from different creative fields like the arts, music, or science, the mad genius link also seems to be ubiquitous. Examples like Vincent van Gogh, who, presumably suffering from BD, cut off his ear and eventually committed suicide; Robert Schumann, who created 12.3 hymns on average during his manic episodes, while on average creating only 2.7 during his depressive episodes (Weisberg, 1994); or John Forbes Nash, one of the most important mathematicians in the 20th century, who suffered from paranoid schizophrenia (Dietrich, 2014). Self- evidently, many famous and genius artists, musicians, and scientists do or did not suffer from mental illness – yet, these examples do not seem to win over the media. Stories of the suffering artist or the crazy scientist are often portrayed in books, films, and TV series (e.g.

Harbour, 2015; Ponterotto & Reynolds, 2013), popularizing the idea of the mad genius link in the general public.

Psychological research has contributed empirical evidence to inform the debate. One of the first studies to examine overall mental illness in a creative sample, also known as the landmark study, was conducted in 1997 by Andreasen, stating that in the study, reportedly BIPOLAR DISORDER AND CREATIVE POTENTIAL

80% of creative writers (compared to 30% of nonwriters) suffered from a mood disorder.

Even greater public attention was given to Jamison’s book Touched with Fire in 1993, in which the author examined mental illness in a sample of people in creative fields such as poets, writers, and visual artists, stating that probability of mental illness in creative individuals was 30 times higher than in non-creative people (Jamison, 1993; 1996).

Following this path, in The Price of Greatness, Ludwig (1995) reviewed 1,004 biographies of individuals published in the New York Times between 1960 and 1990, most of which were in creative professions such as art, music, and poetry, concluded that great creativity comes at the price of madness.

All studies by these authors, namely Andreasen, Jamison, and Ludwig, are often cited as evidence for a mad genius link, not only in psychological research but also in popular media. However, these studies have been harshly criticized, especially by Schlesinger in

2009, pointing to poor methodological quality. First, the participants in these studies were not diagnosed with mental disorders; instead, they simply self-reported their mental well-being.

Furthermore, most of the studies did not have a control group, which is why no in-depth statistical analyses could be conducted and only percentages of mental illness are stated. On top of that, due to the long duration over which data was collected, the small number of participants, as well as the personal connection between the participants and the researchers,

Andreasen’s (1997) study in particular was even more harshly criticized as not only being unscientific but also biased.

Overall, the evidence from psychological research for a mad genius link is subject to debate and still based heavily on anecdotal rather than empirical grounds. Although the connection between psychopathology and creativity is not well supported, the idea that mental illness coincides with giftedness or great creativity seems to be something considered as fact by the general population. Studies on stereotypes about gifted individuals show that BIPOLAR DISORDER AND CREATIVE POTENTIAL people often view them as having negative socioemotional characteristics, such as isolation

(Solano, 1987), lower agreeableness, higher introversion, and higher neuroticism (Baudson &

Preckel, 2013). Even though the stereotype of the sad loner as well as the stereotype of the mad genius may not be empirically true, they still ‘form a reality in people’s minds and shape how they perceive and behave towards gifted individuals’ (Baudson, 2016, p. 1).

Theoretical Framework for a Relationship Between Creativity and Bipolar Disorder

Shared phenotype. Even though there is no clear support of the mad genius theory by scientific research, it makes sense to explore a possible connection between BD and creativity based upon shared composition. To begin with, typical symptoms of BD and the abilities needed in creative thinking may be related, especially the following symptoms that appear during a (hypo)manic episode: racing thoughts, distractibility, and increased goal-directed activity (APA, 2013; criterion B 4 - 6). Racing thoughts may mean that an individual may have many ideas, which appears in line with the concept of fluency; distractibility might imply frequent shifts between different topics and, therefore, may connect to the concept of flexibility. Jamison (1993) also drew a connection between hypomanic episodes and creative potential, saying that the

[t]wo aspects of thinking in particular are pronounced during creative and hypomanic

thought – fluency, rapidity, and flexibility of thought, on the one hand, and the ability

to combine ideas or categories of thought in order to form new and original

connections, on the other. […] Speed per se, that is, the quantity of thoughts and

associations produced in a given period of time, may be enhanced. The increased

quantity and speed of thoughts may exert an effect on the qualitative aspects of

thought as well; that is, the sheer volume of thought can produce unique ideas and

associations. (Jamison, 1993, p.105). BIPOLAR DISORDER AND CREATIVE POTENTIAL

To summarize this idea, the characteristics of a (hypo)manic episode of racing and shifting thoughts might support the ability to have many ideas as evidenced in high idea fluency, which further allows for creating unique and novel ideas, as related to originality.

Indeed, this link between quantity (fluency) and quality (originality) in idea generation is implied by Simonton’s equal odds baseline (Forthmann et al., 2021; Mouchiroud & Lubart,

2001; Simonton, 2010) and also by the dual pathway model of creativity (Nijstad et al.,

2010). Murray and Johnson (2010) even went a step further by developing a schematic representation between the BD phenotype and creativity. Of note, they use different terminology around the concept of creativity, distinguishing between generativity/novelty and consolidation/usefulness. Generativity relates to the production of various ideas, as in fluency, and consolidation relates to editing and polishing of ideas, mostly related to originality.

The core idea in their theory is that elevated positive affect during mania might lead to greater DT, topped up by impulsivity, which might boost this creativity without being inhibited by self-censorship (Murray & Johnson, 2010). In this scheme, a number of variables overlap either between BD and both facets of creativity (A), BD and generativity (B), and BD and consolidation (C). The three-way overlap between BD, generativity, and consolidation

(A) includes the following variables: positive affectivity, openness, extraversion

(performance), neuroticism (arts), neuroticism (science), high goal setting, romantic aesthetic, familial creativity, and evening preference (arts), which all have a positive direction. The direction of the association is more ambiguous for the two two-way relationships between generativity and BD (B) and (consolidation and BD C): While (B) is assumed to go along with psychoticism, impulsivity, loose associations, and imagery which are all connected positively, (C) has mixed associations, since impulsivity, intrinsic motivation, substance BIPOLAR DISORDER AND CREATIVE POTENTIAL misuse, and anxiety disorder have a negative connotation, and only drive to achieve has a positive association.

In conclusion, there is a variety of behavioural, cognitive, and motivational variables that BD and the two facets of creativity have in common. However, the scheme portrays a rather flat net of connectivity, for instance neglecting dynamic relationships such as the influence of depressive episodes in BD.

Shared vulnerability. Alongside the aforementioned idea of the link between BD and creativity on the level of phenotype or behaviour, one could also step back and consider similarities in the underlying factors that lead to BD or creativity. Based on evidence from molecular genetics, neuroscience, and biological determinants, Carson (2011) suggested a model of shared vulnerability, postulating that there are overlapping vulnerability factors between creativity and psychopathology that, in combination with either protective or risk factors, result in either enhanced creativity or psychopathology (psychopathology here refers to BD, schizophrenia, and substance abuse only). Derived from the fact that both creativity and BD are heritable (Kendler et al. 1993; Lichtenstein et al., 2009; McGuffin et al., 2003) and to some extent polygenetic (Berrettini, 2000; Whitfield et al., 1998), the idea of a genetic linkage stands to reason. Indeed, polygenetic risk scores for BP were found to predict creative professions to a modest degree (Power et al., 2015). As such, Carson (2011) generally assumes that creativity and psychopathology are similar in that they give individuals enhanced access to subconscious material. While the risk factors for psychopathology (low intelligence, working memory deficits, perseveration) lead to being overpowered by this unconscious material, resulting in pathological symptoms, the protective factors (high intelligence, working memory skills, cognitive flexibility) ensure the necessary meta- cognitive control to turn one’s bizarre and unusual thoughts into a creative advantage. BIPOLAR DISORDER AND CREATIVE POTENTIAL

The three vulnerability factors identified by Carson (2011) that may either result in enhanced creativity or psychopathology (depending on protective versus risk factors) are attenuated latent inhibition, preference for novelty, and hyperconnectivity. First, latent inhibition (LI) refers to how well sensory stimuli are selected and filtered (Kasprow,

Catterson, Schachtmann, & Miller, 2007; Lubow & Moore, 1959). Having a low LI implies that irrelevant information enters consciousness more easily (Carson, 2011; Carson, Peterson,

& Higgins, 2003). A low LI has been found not only in individuals with psychosis (mainly schizophrenia but also BD; Baruch, Hemsley, & Gray, 1988; Lubow, Ingberg-Sachs,

Zalstein-Orda, & Gewirtz, 1992) but also in individuals with high openness to experience, which, in turn, is the personality trait that best predicts creative potential (Peterson & Carson,

2000; Peterson, Smith, & Carson, 2002). Furthermore, a low LI can be induced by psychoactive substances, which again have been linked to elevated levels of creativity (Gray et al., 1992; but see also Benedek & Jauk, 2019).

Second, the factor preference for novelty or novelty seeking is an intrinsic motivational variable of heightened interest for exploring and approaching new stimuli (e.g.

Ebstein et al., 1996). Novelty seeking is related to both creativity and BD, since it was found that creative individuals prefer novel and complex stimuli over familiar and simple stimuli

(Martindale, 1999; Reuter et al., 2005), and people with BD were found to show higher novelty seeking, especially during episodes of mania or hypomania (Haro et al., 2007).

Finally, the factor hyperconnectivity, described as abnormal neural linkage between brain areas that typically have no functional connection (Carson, 2011), has been detected both in people with mental disorders and in highly creative people performing a creative task

(Chai et al., 2011; Takeuchi et al., 2011; Whitfield-Gabrieli et al., 2009). Further, hyperconnectivity is a common phenomenon in people with synesthesia, namely cross-modal sensory associations such as connecting colors with musical tones, which is prevalent more BIPOLAR DISORDER AND CREATIVE POTENTIAL often in creative individuals (Hubbard & Ramachandran, 2005). As speculated by

Ramachandran and Hubbard (2010), patterns of hyperconnectivity may even be connected to human metaphorical thinking, a thinking style often seen in creative people as well as people experiencing hypomania, psychotic episodes, or drug intoxication (Carson, 2011;

Ramachandran & Hubbard, 2001).

Taken together, these shared vulnerability factors offer theoretical grounds to assume a link between BD and creativity. Yet, the model does not imply that bipolar individuals are more or less creative; it only indicates that BD and creativity share certain dispositions but it depends on yet other (protective versus risk) factors to eventually enhance either BD or creativity, but not necessarily both. As such, the meta-analysis at hand will try to incorporate variables that act as either protective factors or risk factors and thus include intelligence as a moderator in the analysis.

Inverted U-shaped relationship. Apart from the general idea of a relationship between creativity and BD, some theories further deal with the type of the relationship. One of the most prevalent theories regarding the type of relationship is the idea of an inverted U- shaped relationship (e.g. Abraham, 2014; Richards, Kinney, Benet, & Merzel, 1988;

Simonton, 2019). Here, a curvilinear relationship between DT and BD is hypothesized, such that creativity is expected peak at a certain, rather subclinical level of BD, whereas individuals without any symptoms of BD as well as severely ill individuals show lower creativity (Abraham, 2014; Simonton, 2019). Greenwood (2016) also postulated an inverted

U-shaped model, in which creativity and other desirable traits are expected to be enhanced by genetic loading, but at some point they cross over into undesirable regions as signs of psychopathology appear. This non-linear relationship is further foreshadowed by the above quote from the DSM-5 on the relationship between DT and BD, saying that greater creative BIPOLAR DISORDER AND CREATIVE POTENTIAL accomplishments are related to milder forms of BD and that creativity might be enhanced in healthy relatives of people suffering from BD.

Following the hypothesis of an inverted U-shaped relationship, it might also be interesting to study the relationship between DT and BD in the context of subclinical expressions of BP. Hence, studies using a subclinical sample will also be included in the meta-analysis, and examined as moderator. One of the most commonly used diagnostic tools for assessing subclinical symptoms of BD is the Hypomanic Personality Scale (HPS; Eckblad

& Chapman, 1986), which was found to correlate highly with hypomanic symptoms (Kim,

2016). Hence, to more accurately map the different levels of impairment due to BD, the meta- analysis will include severity as a moderator, thus allowing us to investigate a possible inverted U-shaped relationship.

Previous Findings on the Link Between Bipolar Disorder and Creativity

Since BD can primarily be seen as a disruption of mood (Mansell, Morrison, Reid,

Lowens, & Tai, 2007), the role of mood is noteworthy when investigating the relationship between BD and creativity. Furthermore, since BD usually manifests itself in a pattern of depressive and manic or hypomanic episodes, it also makes sense to consider the relationship between major depression and creativity. On these grounds, the following paragraphs first investigate the relationship between creativity and non-psychopathological affective states, such as a positive and negative mood, and then dive deeper into existing data on the connection between depression and creativity, before finally reviewing current results on the relationship between BD and creativity.

Mood and creativity. When breaking mood down to its basic spectrum, it can be seen as a two-poled scale or hedonic tone (Baas, De Dreu, & Nijstad, 2008) between positive and negative affect. Mood can have a variety of influences on our lives, impacting emotions, behaviours and cognitions as well as cognitive processes associated with creativity BIPOLAR DISORDER AND CREATIVE POTENTIAL

(Chermahini & Hommel, 2012; Russ, 1999). To date, most studies have found that a positive mood facilitates creativity, in that a positive mood enhances DT (e.g. Isen, 1999; Hirt,

Melton, McDonald, & Harachiewicz, 1996). For instance, Isen, Daubman, and Nowicki

(1987) found that positive mood resulted in enhanced performance in the Duncker candle task (Duncker & Lees, 1945), which is assumed to measure creative problem solving, yet in a convergent rather than divergent manner (see also Cropley, 2006). These results were replicated by Greene and Noice (1988) using also Duncker’s candle task. Further, Ashby and

Isen (1999) showed that positive affect results in greater cognitive flexibility, which, in turn, results in better performance in DT fluency and flexibility. But apart from enhancing fluency and flexibility, positive mood states also seem to enhance originality, as suggested by

Lyubomirsky, King, and Diener (2005), who assume that positive mood leads to richer associations within one’s existing knowledge structures and, therefore, to more originality. In addition, Yamada and Nagai (2015) found that, compared to a neutral-mood control group, an induced positive mood enhanced the production of unconventional and untypical ideas but did not increase performance on CT tasks.

However, a few contradictory results exist, reporting a lower creative performance associated with positive mood (e.g. Kaufman & Vosburg, 1997; Martin, Ward, Achee, and

Wyer, 1993). When taking a closer look at these studies, the contradictory results might be explained by some very crucial deviations in how DT was assessed. For example, Martin and colleagues (1993), participants were instructed to stop the task at any point if they no longer enjoyed it.

The reason why a positive mood might enhance DT can be explained by the association between positive mood and memory: Positive mood is connected to greater memory retrieval relative to other mood states, and retrieving memories is necessary for generating ideas (e.g. Vosburg, 1998; Isen, 1999). This can partly be explained by increased BIPOLAR DISORDER AND CREATIVE POTENTIAL dopamine release during positive mood, which, in turn, can stimulate cognitive abilities (e.g.

Ashby & Isen, 1999). Furthermore, positive mood is associated with more rapid decision- making (Isen & Means, 1983), which might also be relevant during creative problem-solving tasks. Further, relevant specifications on the relationship between positive mood and creativity were made by Chermahini and Hommel in 2012 regarding the level of positive mood induction: In their study, they found that the dopamine release induced by a positive mood facilitates DT performance – but only to a certain extent. The highest cognitive flexibility is reached at a certain amount of dopamine release, but it decreases if this sweet spot is exceeded by too much dopamine release in prefrontal regions. Therefore, a non-linear relationship between mood and DT can be assumed, again referring to the idea of an inverted

U-shaped relationship.

Regarding negative mood, research results are far more contradictory compared to positive mood: Some studies found that negative mood was associated with decreased creative performance (e.g. Kaufman & Vosburg, 1997), some found no association (e.g.

Göritz, & Moser, 2003; Verhaeghen, Joormann, & Khan, 2005), and others even found it was associated with enhanced performance (e.g. Adaman & Blaney, 1995; Carlsson, Wendt, &

Risberg, 2000). In one study by Kaufman and Vosburg (1997), experiments were performed with induced and naturally occurring negative moods, which both caused worse performance on the insight task. Yet, in a study using the same insight task, Isen et al. (1987) found that negative mood had no influence on creativity. Further, in another study, even though they hypothesized the opposite, Clapham (2001) found that negative affect, operationalized by trait and state anxiety, had a positive correlation with figural DT. In the context of these conflicting findings, Zenasni and Lubart (2002) tried to broaden the spectrum of the mood- creativity relationship by including various mood-related moderators. They found that valence and intensity of mood as well as playfulness of the task must be considered: In their BIPOLAR DISORDER AND CREATIVE POTENTIAL study, negative mood generally seemed to reduce performance in two assignments of the

Abbreviated Torrance Test for Adults (ATTA; Goff & Torrance, 2002), but performance could be boosted if the negative mood was highly intense and the task was fun and playful

(Zenasni & Lubart, 2002). Taken together, the effect of negative mood on DT is very ambiguous, and many yet uncovered moderators might be at play.

A possible explanation for these diverging results might due to the nature of negative affect itself. Negative affect states are assumed to have a far more complex dimensional structure than positive affect states (Clapham, 2001; Ellsworth & Smith 1988). Also, the various types of negative emotions, such as anxiety, anger, or sadness, seem to be more distinct from each other than positive emotions (Clapham, 2001; Ellsworth & Smith 1988).

Therefore, the induction of negative mood might lead to a variety of negative mood states that might not be consistent among participants, leading to varying results in creative performance.

Overall, the results for how positive mood influences creativity seem quite clear, in that positive mood leads to enhanced creativity. For negative mood, results are inconclusive, and a handful of important processes still remain uncovered. A recent meta-analysis on the relationship between mood and creativity is in line with this conclusion (Davis, 2009).

However, another interesting idea is that the influence of mood on creativity might not be unidirectional but in fact be bidirectional, as one’s performance on a creativity task might also influence one’s current mood state (Bujacz et al., 2014). So far, not enough research has addressed the reverse direction, namely the influence of creativity tasks on mood, but first indications foreshadow that high motivation, a love for the task (Russ, 1999), and a broader attentional focus (Fredrickson & Branigan, 2005; Yamada & Nagai, 2015) during DT tasks might also brighten one’s mood. The exploration of causal mechanisms falls beyond the BIPOLAR DISORDER AND CREATIVE POTENTIAL scope of the current meta-analysis, but it is important to keep in mind, since it would have practical implications for treating mood disorders.

Depression and creativity. Despite its high prevalence, depression has not been heavily studied in creativity research (Silvia & Kimbrel, 2010). And, similar to the aforementioned paragraph in the DSM-5 about the mad genius link, empirical studies on the link between depression and creativity have mostly focused on creative professions, creative achievement, or creative personality. For example, several studies have investigated whether individuals pursuing a creative career, such as art students, had higher rates of depression, which was generally supported (e.g. Lauronen et al., 2004; Papworth & James, 2003; Young,

Winner, & Cordes, 2013). Also, on the eminent level, biographical and historical analyses of the mental health of eminent creatives suggest that eminent creatives have a greater prevalence of mood disorders, particularly depression (Schildkraut, Hirshfeld, & Murphy,

1994; Wills, 2003). Further, the literature review by Feist (1998) also came to the conclusion that depressive affective illness leads to greater creative accomplishments, primarily regarding artistic (not scientific) creativity.

However, results from research on depression in eminent creativity cannot be transferred to the current meta-analysis, which focuses on creative potential, since the relation between DT and creative achievement is rather small (Kim, 2008). Regarding DT, studies on the relationship between DT and depression are still scarce. In one recent study, Silvia and

Kimbrel (2010) conducted confirmatory factor analyses and multivariate structural equation models (SEMs) on the relationship between depression and creativity, including a variety of different creativity measures, such as DT, creative self-concepts, everyday creativity, and creative achievement. Interestingly, the effect sizes were small and inconsistent in their direction, and the SEMs showed that depression only explained a small amount of the variance in creativity (roughly 3%), independent of the type of creativity measured. BIPOLAR DISORDER AND CREATIVE POTENTIAL

On the subclinical level, Schuldberg (2001) found that higher scores on the MMPI-2

(Minnesota Multiphasic Personality Inventory-2; Hathaway, McKinley, & MMPI

Restandardization Committee, 1989) for depression were related to lower creativity in various measurements, including the AUT. All effects were small yet significant, with rs around -.15 and -.20, suggesting that a subclinical tendency towards depression worsens one’s performance in DT (Schuldberg, 2001). Interestingly, another study found that current depression measured by the CED-D (Center for Epidemiological Studies Depression Scale;

Locke & Putman, 1971) was neither linked to creative interest nor DT as measured by the

ATTA (Verhaeghen et al., 2005). Even though this study showed that current depression had no link to DT, past depressive symptoms, as measured by a 10-item checklist based on the

DSM-IV, predicted higher DT scores, and this was mediated by self-reflective rumination

(Verhaeghen et al., 2005). Furthermore, Sánchez, Martín-Brufau, Méndez, Corbalán, and

Limiañan (2010) found zero-correlations between depression and creativity, operationalized by the CREA task (Corbalán, Martínez, Donolo, Tejerina, & Limiñana, 2003) measuring creative cognitive flexibility in a subclinical sample of university students.

On the pathological level, creativity in individuals diagnosed with depression appears to be domain specific: Santosa and colleagues (2007) compared patients diagnosed with a major depressive episode to people in two control groups (healthy controls and creative controls) using a variety of creative measurements (creative interest, creative personality, and

DT) and found no difference between the depressed group and the healthy control group in the verbal domain of DT. However, individuals with depression performed worse on the figural domain of DT compared to the control groups, indicating that individuals with depression might either perform just as well or even slightly worse on certain domains of DT compared to healthy people. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Taken together, the link between creativity and depression is not fully understood, and available findings are inconsistent. So far, psychological research indicates that depression might be connected to greater creative accomplishment on the eminent level (Lauronen et al.,

2004; Papworth & James, 2003) or at least to the greater probability of having a creative profession (Schildkraut et al. 1994; Wills, 2003), but depression does seem to reduce people’s everyday creativity, such as in DT (Santosa et al., 2007; Silvia & Kimbrel, 2010; Verhaeghen et al. 2005). Even though a depressive episode within a major depressive disorder and within

BD are distinct from each other (Mitchell et al., 2001), these results are relevant to consider when investigating the relationship between BD and creativity.

Bipolar disorder and creativity. As stated before, BD seems to hold a special position in the relationship between creativity and psychopathology, as seen in DSM-5’s paragraph specially dedicated to creativity as being related to BD (APA, 2013). Regardless of its presumable significance, once again empirical studies on the relationship between BD and creativity are rare, and have focused heavily on creative achievement, creative professions, or creative personalities rather than creative potential. To find out whether individuals with BD are creative, the most obvious approach would be to simply ask them about their self- observed creative actions during affective episodes. This is exactly what McCraw, Parker,

Fletcher, and Friend (2013) did in their study with a sample of 213 patients diagnosed with either bipolar I or II disorder. Results showed that, indeed, 82% of patients confirmed that they were creative during hypomanic and manic episodes. A following qualitative analysis on a subsample of the patients showed that bipolar patients mostly engaged in writing, painting, work, or generating business ideas. However, the study did not include a control group, so the results only indicate that bipolar patients show creative actions during (hypo)mania; no proposition could be made as to whether bipolar patients show more creative behaviour compared to healthy individuals. BIPOLAR DISORDER AND CREATIVE POTENTIAL

On a bigger scale, Kyaga et al. (2011) conducted a nested case-control study on

1,742,684 people registered in the Swedish population register between 1969 and 2006 to examine a connection between field of occupation and mental illness. Diagnoses were retrieved from the Hospital Discharge Register, and occupations were divided into scientific, visual artistic, non-visual artistic, and other occupations. In total, 29,644 individuals with BD could be identified, and analyses showed that individuals with BD as well as their healthy siblings were overrepresented in artistic and scientific creative occupations (Kyaga et al.,

2011).

The study by Simeonova et al. (2005) used the Barron-Welsh Art Scale (BWAS,

Welsh & Barron, 1959) to examine a relationship between BD and creativity. The BWAS contains 86 figures of varying complexity, whereby complex and asymmetrical figures are assumed to reflect a higher degree of creativity. Simeonova et al. (2005) found that individuals with BD scored 120% higher on the two subscales of the BWAS compared to healthy controls, indicating that BD might indeed enhance interest in more complex stimuli and, thus, creativity. Interestingly, the study not only investigated adults suffering from BD but also considered their offspring, who also scored higher on the BWAS compared to healthy control children (Simenonva et al., 2005). This finding was extended by Soeiro-de-

Souza, Dias, Bio, Post, and Moreno (2011), who further observed differences related to bipolar patients’ current episode: BWAS total scores were significantly higher during mania than during mixed or depressed episodes.

In the context of little-c creativity, Ruiter and Johnson (2015) found that the risk for

BD as measured by the Creative Achievement Questionnaire (CAQ; Carson, Peterson, &

Higgins, 2005) as well as heightened scores on the HPS were not related to performance on the Compound Remote Associates test (CRA; Bowden & Jung-Beeman, 2003), a derived version of the Remote Associates Test (RAT; Mednick, 1968) that measures creative problem BIPOLAR DISORDER AND CREATIVE POTENTIAL solving in a convergent manner. In contrast to these findings, Kim and Kwon (2017) found a large correlation between hypomania risk and self-reported everyday creativity, and this correlation was moderated by the behavioural activation system, a neurobiologically based system that is associated with goal-directed and reward-relevant stimuli (e.g. Johnson, Edge,

Holmes, & Carver, 2012).

Focusing on studies that used DT as a measure of creativity to examine the relationship between creativity and BD in clinical and subclinical samples, the results appear to be just as diverse. For instance, Siegel and Bugg (2016) found no significant relationship between the risk for hypomania as measured by the HPS and DT as measured by the ATTA.

In fact, none of the four subscales (fluency, flexibility, originality, elaboration) were connected to measures of hypomania risk (Siegel & Bugg, 2016). On the other hand, other studies such as the study by Alici et al. (2019) showed a negative relationship between BD and DT. In their study, remitted patients with BD were compared to healthy controls, whereby BD patients showed significantly worse performance in fluency and originality in

AUT. In contrast, other studies have shown increased DT performance in bipolar patients: In a study by Srivastava and colleagues (2010), for example, bipolar patients scored better in verbal and figural DT tasks than healthy controls and depressed patients. In line with this,

Claridge and Blakey (2009) found that hyperthymic and cyclothymic temperament could predict heightened performance in DT tasks. In summary, the body of empirical studies on the relationship between BD and DT is limited, and existing studies show diverging results, calling for a systematic meta-analytic examination of the average effect and potential moderators.

Previous Meta-Analytic Findings

Meta-analytic approaches on the connection between creativity and psychopathology are still scarce. A recent meta-analysis by Paek, Abdulla, and Cramond (2016), which BIPOLAR DISORDER AND CREATIVE POTENTIAL focused on an overall mad genius connection by investigating the relationship between three common psychopathologies – ADHD, anxiety, and depression – and little-c creativity, did not find a substantial connection between both (overall correlation: r = .06). However, when taking a closer look, differences in strength of the relationship were found between the three psychopathologies, and ADHD and creativity even had a negative association (r = - .17). This highlights the relevance to examine different disorders separately. So far, only two meta- analyses have explicitly included BD: First, the systematic review and meta-analysis by

Taylor (2017) examined the connection between creativity and mood disorders, and, second, the meta-analysis by Baas, Nijstad, Boot, and De Dreu (2016) contrasted creativity in approach-based vs. avoidance-based psychopathologies, including the risk for BD in the cluster of approach-based psychopathologies.

The systematic review and meta-analysis by Taylor (2017) examined the connection between creativity and mood disorders, including unipolar depression, dysthymia, and BD.

Three separate meta-analyses were conducted to understand the direction of the relationship, investigating (1) mood disorders in creative individuals, (2) creativity in individuals with mood disorders, and (3) a covariance of creativity and mood disorders. The first group, focusing on mood disorders in creative individuals, included 1,805 participants from 10 studies and 24 effect sizes, finding a moderate to large effect (g = 0.64). The second analysis, focusing on creativity in people with mood disorders, included 8,316,598 participants from

13 studies and 76 effect sizes yielding a small and not statistically significant connection (g =

0.08). And the third analysis on the covariance of both included 3,965 participants from 15 studies and 59 effect sizes suggesting a small yet significant correlation of r = .09. Taken together, the results by Taylor (2017) point towards the idea that creative people in fact show a greater prevalence for having a mood disorder than non-creative people, while people with a mood disorder do not seem to be more creative in general. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Even though this study gives insightful results – not only regarding a connection between mood disorders and creativity but also regarding the direction of the effect – there are still some important limitations. First and foremost, the effect sizes were not calculated for specific disorders but as an overall effect for mood disorders. When taking a closer look at the results from the moderation analysis, especially from the first meta-analysis, it is noticeable that the level of creativity differs between the different types of disorders: While dysthymia is not linked to creativity at all across all analyses, BD seems to have the strongest connection, which might have biased the strength of the overall effect, once again making it clear that disorders should be analysed individually. Moreover, Taylor (2017) used several indicators of creativity, ranging from everyday creativity over creative performance to creative achievement, such that DT was only one of many variables used to operationalize creativity. For instance, DT and creative accomplishments only have a small (r = .22) correlation (Kim, 2008), which is not sufficient to merge these variables into an overall creativity score. Therefore, the current meta-analysis solely focuses on DT, since it is the most commonly used indicator of creative potential (Runco & Acar, 2012).

Furthermore, calculating three separate analyses to investigate the direction of the effect is an interesting and insightful approach, but it results in a smaller number of studies per analysis. Rather than conducting three separate analyses with smaller sample sizes over various disorders, the current meta-analysis rather conducts one meta-analysis that only focuses on BD and incorporates the assessment method as a moderator together with other relevant factors.

The meta-analysis by Baas et al. (2016) also included BD when investigating creativity but took a slightly different approach. Instead of focusing on a clinical sample,

Baas and colleagues investigated the vulnerability or risk for psychopathologies in non- clinical samples. Also, the relationship between creativity and psychopathology was not BIPOLAR DISORDER AND CREATIVE POTENTIAL investigated by specific disorders but was separated into two clusters based on motivational approach and avoidance systems: Approach-based psychopathologies included positive schizotypy and risk for BD, and avoidance-based psychopathologies included anxiety, negative schizotypy, and depressive mood. Results showed that approach-based psychopathologies were indeed related to higher creativity (positive small effect), while avoidance-based psychopathologies were related to lower creativity. Taking a closer look, results also showed that risk for BD was associated with higher creativity (k = 28; r = .22), indicating that people with a greater risk of suffering from BD are also more creative. Yet again, these results highlight the need for a disorder-specific approach regarding the connection between creativity and psychopathology, but they still must be taken with caution.

First and foremost, Baas et al. (2016) explicitly examined a non-clinical sample, whereas the connection between creativity and BD in the whole spectrum of clinical and subclinical samples is relevant. Therefore, rather than looking at bipolar risk via family history or related diagnostic assessment methods like temperament, such as in Baas et al. (2016), the current meta-analysis focuses on clinically diagnosed BD, including presentations of this diagnosis as euthymic, depressed, or manic, as well as subclinical BD. Furthermore, just as in the meta- analysis by Taylor (2017), Baas et al. (2016) used many variables to operationalize creativity, roughly separating them into clusters of creative performance (DT is included here) and self- rating of creativity (in terms of personality and creative behaviour). Apparently, DT was defined rather broadly, since the RAT test or tasks of verbal fluency were also included in the category of DT; however, the RAT corresponds to convergent thinking, and verbal fluency cannot be used synonymously with fluency in view of Guilford’s (1950, 1957) classification, which means these measures are not equivalent to DT. To avoid ambiguity in the conceptualization of creativity, the current meta-analysis solely focuses on creative potential BIPOLAR DISORDER AND CREATIVE POTENTIAL measured by DT tasks and excluded studies that made exclusive use of other measures such as the RAT or verbal fluency.

Aims of the Present Study

Considering all the different presentations of BD, BD is not only a common mental illness (APA, 2013; Merikangas et al., 2011), but, due to its associated high suicide rates, it is also one of the deadliest (APA, 2013; Rihmer & Kiss, 2002). Therefore, more in-depth research is needed on the protective factors and risk factors of BD in order to build a better understanding of the disorder and continuously work on better treatment methods. Since it is still debated whether creativity is unrelated, positively related, or negatively related to BD, this meta-analysis aims to offer systematic research on variables connected to BD. Further, this meta-analysis aims to challenge myth-based conceptions about the relationship between psychopathology and creativity that appear to be driven by anecdotal evidence. The idea of a mad genius often romanticizes people struggling with mental illness, making it sound desirable to have a mental disorder in order to benefit from the creative genius that comes along with it. On the other hand, mental disorders are still stigmatized in today’s society (e.g.

Corrigan & Bink, 2016), such that considerable negative consequences come along with these diagnoses. To neither romanticise a mental illness nor to deny possible positive implications, the present meta-analysis aims to assemble empirical facts to enable a better understanding of the connection between BD and creativity.

To do so, the meta-analysis makes use of distinct definitions and assessments of both

BD and creativity. As opposed to existing meta-analyses that have commonly focused on psychopathology as a whole, the current meta-analysis solely focuses on BD, which arguably shows the strongest theoretical and empirical relationship to creativity (Carson, 2011; Murray

& Johnson, 2010). Specifically, in comparison to the meta-analysis by Taylor (2017), the current study only examines BD rather than a whole cluster of mood disorders and, different BIPOLAR DISORDER AND CREATIVE POTENTIAL from the meta-analysis by Baas et al. (2016), the current meta-analysis focuses on clinical and subclinical BD rather than just the risk for BD.

On top of that, the present meta-analysis focuses on creative potential as assessed with

DT tasks, since we were interested in everyday, little-c creativity rather than Pro-C or Big-C creativity, and aimed to avoid broad, ambiguous conceptualizations of creativity (e.g., creative achievement, professions, personality, or convergent thinking).

To identify factors that affect the relationship between BD and creative potential, the current meta-analysis also considers relevant moderators. They include general moderators on study characteristics such as study age (i.e., 2019 minus year of publication), study country, participants’ mean age, use of a matched control group, as well as study design

(correlational vs. mean differences). Due to the diverse assessment of creative potential, it also investigates moderators related to the kind of DT assessment including type of DT test, scoring of DT test, domain of DT test, instruction, and calculation method. As stated above,

BD is a complex psychopathology, which is why the current study also analyses moderators on specific presentation of BD, status of BD, BD assessment tool, duration, onset, severity, medication, and comorbidities.

Method

Study Collection

Inclusion and exclusion criteria. Studies were included in the meta-analysis if the following criteria were met: (a) The study assessed creative potential in terms of little-c creativity by using measurements of DT (e.g. AUT, TTCT, ATTA). Studies measuring creativity on the Pro-C or Big-C level, such as focusing on creative achievement, or creative profession were excluded. Similarly, studies using other measures of creative potential than

DT, such as verbal fluency or convergent thinking creativity (e.g. RAT), were also excluded.

(b) The study used established measurements to assess clinical or subclinical BP (e.g. BIPOLAR DISORDER AND CREATIVE POTENTIAL assessment by a clinician, diagnostic interviews, or questionnaires especially designed to measure symptoms of BD). Studies only using self-reported psychopathology without a prior diagnosis were excluded. (c) The study had to include quantifiable measures of creativity and psychopathology. Therefore, theoretical reviews, case reports, and similar works were excluded. (d) Studies comparing mean differences had to have a healthy control group (CG).

Studies only comparing two pathological groups or only using a preselected group as CG

(e.g. creative controls) were excluded. For correlational studies, no CG was needed. (e) The study either had to report an effect size or had to report sufficient statistical information to calculate an effect size. In studies comparing mean differences, at least the means and standard deviations had to be reported. Correlational studies had to report the r-value for zero-order bivariate relationships. The reported effect size had to be convertible to the effect size used in the study, which is Cohen’s d; (f) The study had to focus primarily on BD.

Because the majority of patients suffering from BD also have at least one other comorbid diagnosis (APA, 2013; Krishnan, 2005), studies also incorporating patients with comorbidities were not automatically excluded. However, BD had to be the main diagnosis; studies with a sample suffering primarily from another psychological or physiological condition were therefore excluded. (g) The study had to be conducted on a human sample; animal studies were excluded. (h) In the case that the same sample was used in several studies, only the published version or the more substantial version was used. (i) The study had to be published in a language known to the authors (German, English, Polish, or

Norwegian).

Literature search. A literature search for potentially relevant studies was conducted using the electronic databases PsychINFO, PSYNDEX, and Medline. The following search term was used for PsychINFO and Medline: (bipolar* OR mood disorder OR manic* OR mania OR cyclothym* OR hyperthym*) AND ((divergent think* OR creative think* OR BIPOLAR DISORDER AND CREATIVE POTENTIAL creativ* OR convergent think* OR little-c OR fluency OR flexibility OR originality) OR

(alternate use* OR masselon* OR insight* OR remote associat* OR torrance test)) and was adjusted for German language within PSYNDEX. Notably, the search term was more general as implied by the relationship between BD and DT (i.e., it includes convergent think*, insight*, or remote associate*). This choice is based on the fact that sometimes studies with a focus on a different measure of creative thinking (e.g., the RAT) may also assess a measure of DT as a complementary variable. Such studies can still be eligible for a meta-analysis and, hence, the search term was in accordance with a comprehensive search strategy. There were no restrictions regarding date or publication location.

Ultimately, 5,989 articles were found via database searching (as of 10 July 2019).

Further, other resources, such as a forward- and backward-searches on the basis of the two meta-analyses by Baas et al. (2016) and Taylor (2017), as well as a keyword search in Google

Scholar, yielded another m = 59 studies. After removal of duplicates, a total of m = 4,670 studies were screened for eligibility. On the basis of the title and abstract, m = 4,541 studies were excluded for not fitting the subject matter. Hereupon, m = 129 full-text articles were further screened for eligibility, applying the inclusion and exclusion criteria mentioned above. As a result, m = 13 studies, with 13 samples and 42 effect sizes were included in the meta-analysis.

Figure 1 shows a flowchart with more detailed information on the inclusion process.

For simplicity, an excluded study was only listed in one of the exclusion categories, even though multiple reasons for exclusion may have applied. Therefore, the number of studies in each exclusion category is only estimated. Some studies fit inclusion criteria and collected data on DT and BD but unfortunately did not report an effect size or necessary data to calculate an effect size; the authors of these studies were contacted via e-mail and asked whether it was possible to obtain the effect size or data needed to calculate an effect size. If BIPOLAR DISORDER AND CREATIVE POTENTIAL the authors did not answer within the course of one week, a reminder e-mail was sent, including a time limit of two weeks to deliver effect sizes in order to be included in the meta- analysis. Of the authors contacted, three were able to obtain the missing effect size(s), so their studies were included in the meta-analysis. Two authors reported that the information of interest was no longer accessible, and one author did not reply. The literature search is fully documented in a file that is openly available in the Open Science Repository

(https://osf.io/tnksa/). BIPOLAR DISORDER AND CREATIVE POTENTIAL

records identified through records identified through records currently not database searching other sources obtainable (m= 5,989) (m= 59) (m= 30)

records after duplicates removed (m= 4,670)

records screened records excluded (m= 4,670) (m= 4,541)

full-text articles assessed full-text articles for elibibility excluded (m= 129) (m= 116)

studies included in no data (e.g. commentary, essay, meta-analysis editorial, manual) (m= 13) (m= 14)

review or meta-analysis (m= 14)

single-case study (m= 2)

unsufficient assessment of DT (m= 66)

unsufficient assessment of BD (m= 11)

no control group (m= 2)

redundant sample (m= 2)

unfamiliar language (m= 2)

effect size not obtainable (m= 3)

Figure 1. PRISMA flow diagram (Moher, Liberati, Tetzlaff, & Altman, 2009) illustrating the selection process.

Coding Procedure

The coding procedure was developed and defined by communication between the first, second, and fourth author of this work. The first author coded all the variables alone but BIPOLAR DISORDER AND CREATIVE POTENTIAL discussed coding decisions with other members of the department in case of uncertainty or inconsistent information. Detailed information on the coding scheme can be found in Table

A1 in the Appendix. Coding was executed on several levels, namely the study level, sample level, and information on BD as well as information on DT.

On the study level, the following variables were coded: authors, publication year, study age, country of study, type of publication (published vs. unpublished), and type of study

(mean difference vs. correlational). On the sample level, variables coded were total sample size, gender (in % female), mean age, and age group. Age group was coded as school children if the sample comprised of participants younger than 18, as young adults if participants were between 18 and 25, as adults when participants were over 25, and as mixed when several age groups were represented in the sample. Further, it was coded whether or not the CG was matched. For both EG and CG, the group size, mean age, IQ test as well as mean

IQ were coded.

Concerning BD, the following variables were coded: age at first affective episode, number of affective episodes, times of hospitalization due to affective episodes, as well as total duration of the illness. Further information such as BD presentation, BD status, medication, severity, and comorbidity were also coded. Presentation was coded as either bipolar I, bipolar mixed, or subclinical. Bipolar I was coded if all participants suffered from bipolar I disorder, and bipolar mixed was coded if the sample comprised several types of bipolar disorder, such as bipolar I, bipolar II, or unspecified bipolar disorder (please note that no samples were found in which all participants had exclusively bipolar II). Subclinical was coded in case of a subclinical sample, where participants did not have a diagnosis but reported subclinical symptomatology of the BD spectrum. Current status was coded as euthymic when patients were currently in remission or in-between episodes; as manic or depressed depending on the current episodes they were in during the task; or, in case of a BIPOLAR DISORDER AND CREATIVE POTENTIAL non-clinical sample, as subclinical. Medication was coded as all when all the participants took medication; none when none of the participants took medication; or mixed when some of the participants took medication at the time of data collection. For severity, either the value of a specific severity indicator or the average of available assessments were used. Comorbidity was coded as yes when at least some participants of the sample had comorbidities or as no when none of the participants was reported to have any comorbidities. On top of that, assessment of diagnoses was coded separately for BD assessment tool, BD tool category, and whether the assessment was clinical. Information on the specific BD assessment tools can be found in Table A2 in the Appendix. Tool category was coded as DSM when diagnostic tools were based on the DSM, such as the DSM-5, DSM-IV, or diagnostic interviews using the

SKID, or a prior clinical diagnosis given by a clinician; HPS was coded when the tool used was based on the hypomanic personality scale, or as other when none of the other two categories were suitable.

Concerning DT, information on the DT test, DT scoring, DT domain, DT time (i.e., time per item in minutes), DT instruction, and DT calculation were coded. Since most of the tests were either based on Guilford’s AUT (Guilford, 1967) and Torrance’s TTCT (Torrance,

1974), the DT test was classified as AUT, TTCT, or other. Modifications of the AUT (such as the Unusual Uses Task, UUT; Torrance, 1974) or the TTCT (such as the ATTA; Goff &

Torrance, 2002) were also coded within the categories of AUT and TTCT. Information on the specific DT assessment tools can be found in Table A3 in the Appendix. For scoring the DT facets, we coded fluency, flexibility, originality, elaboration, and total. Some studies used slightly expanded measures of originality, such as divergence or uniqueness, which were still coded as originality due to the relatedness of the construct (e.g. Smith & Yang, 2004). Total was coded when studies calculated a total score for creativity or when no information on the specific type of scoring was obtainable. Further, a coding variable was added to be able to BIPOLAR DISORDER AND CREATIVE POTENTIAL differ between studies where the total score was the only scoring used and studies that reported a total score in addition to individual scoring of the DT facets. The DT domain was coded as either figural, verbal, numeric, or mixed. Mixed was coded if test scores were calculated over more than one domain, such as in a total score of the TTCT, which includes a verbal and figural domain. In addition, the available time-on-task was coded as minutes per item. On top of that, DT instruction was either coded as quantity when participants were instructed to generate as many ideas as possible or as quality when participants were instructed to be creative or to generate original ideas. DT calculation was either scored as summed when DT scores were calculated by only including the number of ideas (e.g. a score for originality that was not adjusted for fluency), as averaged when quality of ideas was divided by number of ideas (e.g. a score for originality that was adjusted for fluency), or as other when other calculation methods were used such as the CAT (Consensual Assessment

Technique; Amabile, 1982; Zabelina, Condon, & Beeman, 2014), where the number of ideas was not relevant.

Of note, some presentations of moderators occurred infrequently. To name an example, the presentation England in the moderator country only occurred in two studies with three effect sizes, making the interpretation of the moderator analysis unreliable. As suggested by Fu and colleagues (2011), at least four or more effect sizes per presentation are recommended to calculate subgroup analyses, even though no universal minimum number of effect sizes is required per presentation. Therefore, to control for infrequent moderator presentations in categorial moderators, contrast analyses between moderator subgroups were only calculated for presentations with at least four effect sizes.

Further, not enough information was obtainable from the studies to further analyze some of the variables coded, including IQ test, mean IQ, number of affective episodes, age during first affective episode, duration of hospitalization, total duration of the illness, and BIPOLAR DISORDER AND CREATIVE POTENTIAL severity of BD. For instance, we could not further analyse the number of affective episodes, age of first affective episode, times of hospitalization due to BD, and total duration of the illness, since only one study reported information on these parameters related to the illness.

The same was true for the doses of medication, since only one study reported information on the lithium dose. Even though six studies gave direct or indirect information on severity of

BD, several instruments to calculate severity were used, making it impossible to work with a common scale. Since the model of shared vulnerability (Carson, 2011) named intelligence as one of the most important mediating variables in the relationship between BD and creativity, we also planned to examine IQ as a moderator. However, only two studies gave information on IQ, one using the Wechsler Adult Intelligence Scale (WAIS; Wechsler, 2008) and one using the Baddeley Reasoning Test (Baddeley, 1968), which could not be transformed to a common scale, making the moderator analysis for IQ impossible. Therefore, these variables could not be analyzed further.

Extraction of Effect Sizes and Model Fitting

The current meta-analysis was conducted using the statistics program R (Version

3.1.3; R Core Team, 2018) using the extension package metafor (Version 2.1; Viechtbauer,

2010). The analysis script and the data are openly available in the project’s OSF repository

(https://osf.io/tnksa/). Even though most studies were based on correlational data

(correlational: m = 9, mean differences: m = 5), more effect sizes were based on mean differences (correlational: k = 19; mean differences: k = 23), which is why the effect size used for the meta-analysis was Cohen’s d. In studies calculating mean differences, Cohen’s d was calculated based on means and standard deviations reported in the studies. In correlational

2r studies, Pearson’s r was converted into Cohen’s d by using the formula d= , as √1−r 2 BIPOLAR DISORDER AND CREATIVE POTENTIAL introduced by Borenstein, Hedges, Higgins, and Rothstein (2011), with the variances being

2 2 4V r (1−r ) calculated by V = , with V = (Borenstein et al., 2011). d (1−r2)3 r n−1

Due to the nature of the studies being conducted in different locations and by different researchers, the idea of a random-effects model as opposed to a fixed-effects model seems plausible (Borenstein et al., 2011). To see whether using one source of variance was already sufficient compared to a multilevel model (Konstantopoulos, 2011) that used both sources of variance, such as between-study and within-study variances, likelihood ratio tests were conducted. Since the between-study variance was estimated to be zero, the full model using both sources of variance did not uncover any additional information on variance compared to the model using only the within-study variance (χ2= 0.00, p = 1.00). Thus, a model accounting for the between-study variance was not needed, but within-study variance was necessary (χ2= 17.16, p < .001). Therefore, all further analyses used the simple random- effects model, not the multilevel model using both sources of variance.

Importantly, some studies reported an additional total creativity score besides the individual scorings for fluency, flexibility, originality, and elaboration. Since the total score might be confounded with the individual scorings (Forthmann, Szardenings, et al., 2020), it might be inappropriate to include the total scores into the analysis if the total score was not the only scoring used. To control for the possible confounding, analyses were also calculated with a subset of effect sizes (m = 13, k = 37) excluding total scores whenever other scores were also reported (k = 5 excluded). Results only differed slightly (confidence intervals were strongly overlapping) between the entire group of effect sizes (d = 0.11, 95% CI: [0.00, 0.21], p = .045) and the subset of effect sizes without the additional total scores (d = 0.09, 95% CI:

[-0.02, 0.20], p = .124). Thus, for clarity and simplicity, we used the entire number of extracted effect sizes in all the following analyses. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Results

In total, the meta-analysis included 1,857 participants from 13 studies with 42 effect sizes. The mean age of participants was 28.01 years (SD = 7.54, range = 16.06 - 38.06) with an average of 60.11% female participants (SD = 7.01, range = 47.66 - 72.22). On average, there were 142.85 participants per study (SD = 98.06, range = 58 - 437) across all included studies. Taken together, nine studies (69.23% of studies) with 19 effect sizes (45.24% of effect sizes) used a correlational design, while five studies (38.46% of studies) with 23 effect sizes (54.76%) used mean differences. Here, the total percentage of studies equals more than

100%, since one study reported both correlational effect sizes and effect sizes for mean differences on the basis of different assessment tools. Of the 13 studies, only one study

(7.69% of studies) with two effect sizes (4.76% of effect sizes) was unpublished; all other studies were published.

Regarding the assessment tools for creativity, five studies (38.46% of studies) with 12 effect sizes (28.57% of effect sizes) used tools based on Torrance’s TTCT, seven studies

(53.85% of studies) with 14 effect sizes (33.33% of effect sizes) used tools based on

Guilford’s AUT, and one study (7.96% of studies) with 16 effect sizes (38.10% of effect sizes) used other tools. Namely, this study used the inventiveness scale from the Berlin

Intelligence Structure Test (BIS; Jäger, Süß, & Beauducel, 1997), which, next to an AUT task, also includes several other tasks. The study also included a task, where participants had to give explanations for a specific scenario, and a masselon task, where participants had to form sentences from three given words; this is why the study could not be coded as only

AUT. Regarding the assessment of BD, five studies (38.46% of studies) with 23 effect sizes

(54.76% of effect sizes) used tools based on the DSM, six studies (46.15% of studies) with 12 effect sizes (28.57% of effect sizes) used the HPS, and three studies (23.08% of studies) with seven effect sizes (16.66% of effect sizes) used other measures. More detailed information on BIPOLAR DISORDER AND CREATIVE POTENTIAL study, sample, and effect sizes, including moderators, and Q-statistics, can be found in Table

B1 in the Appendix.

Overall Effect Size on the Relationship Between BD and DT

In total, 42 effect sizes were included in the meta-analysis. Of these, 30 (71.43%) were positive, 12 (28.57%) negative. A histogram for the distribution of effect sizes is shown in Figure 2. The overall effect across all studies in the random-effects model was significantly positive yet diminutively small (d = 0.11, 95% CI: [0.00, 0.21], p = .045). A forest plot for the overall effect can be found in Figure 3.

Since the main fundamental difference between the included studies lies in the two different study types, we conducted further subset meta-analyses for study type (mean differences vs. correlational). In all subclinical samples correlational data was used, while clinical samples reported data based on mean differences; only one clinical study reported both mean differences and correlational data. The random-effect meta-analysis for the subset of studies using mean differences was not significant (d = 0.05, k = 23, 95% CI: [-0.13, 0.22], p = .602), but the subset meta-analysis for studies using correlations showed a significant positive result (d = 0.18, k = 19, 95% CI: [0.08, 0.27], p < .001). Forest plots for the correlational and mean differences subsets can be found in the Appendix (see Figure B1 for correlational and Figure B2 for mean difference forest plots).

The conducted overall random-effects model indicated considerable heterogeneity (Q

= 116.57, df = 41, p < .001, I 2 = 67.13%), which means that random variance does significantly differ from zero. To account for the found heterogeneity, moderator analyses were conducted on the level of study and sample characteristics as well as coded variables regarding DT and BD. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Figure 2. Histogram illustrating the distribution of effect sizes using Cohen’s d values

(Cohen, 1988). BIPOLAR DISORDER AND CREATIVE POTENTIAL

Figure 3. Forest plot for the overall effect including all 42 effect sizes from 13 studies.

Analysis of the Moderation Effects by Study and Sample Characteristics

Study age. All studies were from 2006 (13 years old) or newer, except two studies published in 1990 (29 years old) and 1997 (22 years old). Study age did not significantly account for any amount of the overall effect size (Q = 1.02, df = 1, p = .313,  = 0.28).

Country. There was a moderation effect of the country the study was conducted in that was significant by trend (Q = 7.75, df = 3, p = .051,  = 0.26), such that one study from

Turkey ended up having large influence (d = -0.97; 95% CI: [-1.75, -0.19]; p = .015).

Contrast analyses were only conducted between country presentations with at least four effect sizes, namely between the USA and Poland. The contrast between studies from those countries was not significant. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Study type. The type of the study did not account for significant variance in the overall effect size (Q = 1.28, df = 1, p = .259,  = 0.27); yet, there was a significant, positive effect for correlational studies (d = 0.17, 95% CI: [0.02, 0.32]; p = .027), but no significant effect for difference studies (d = 0.05, 95% CI: [-0.09, 0.19]; p = .503).

Mean age. The mean age of participants did not explain significant variance of the overall effect size (Q = 2.88, df = 1, p = .090,  = 0.31). One study did not report the mean age of participants.

Age group. The age group of participants had no significant effect on the overall variance in effects (Q = 1.64, df = 3, p = .651,  = 0.28). Yet, it is important to note that most participants were classified as adults (m = 5, k = 25) or young adults (m = 6, k = 15), while only one study with one effect size was classified as being mixed and including school children.

Gender. There was no significant effect of gender ratio on the overall variance of effect sizes (Q = 0.78, df = 1, p = .377,  = 0.29). One study with two effect sizes did not report gender information.

Matched CG. Whether or not the CG was matched did not explain significant variance of the overall effect size (Q = 0.01, df = 1, p = .910,  = 0.38). Notably, of the studies having a CG, only two used a matched CG; three studies used no matched CG.

Analysis of the Moderation Effects Regarding DT

DT test. Categorization of DT tests as either AUT, TTCT, or other did not have a significant moderating effect on the overall variance of the effect size (Q = 0.27, df = 2, p

= .872, = 0.28). When examining the DT measures, namely either AUT (m = 5, k = 8),

UUT1 (m = 2, k = 6), ATTA (m = 3, k = 8), TTCT-F (m = 1, k = 2), TTCT-V (m = 1, k = 2), or BIS (m = 1, k = 16), no significant effect on the overall variance of effect sizes could be

1 To give a more detailed insight into differences between DT tests used, here AUT and UUT are listed separately, even though they are often used synonymously. BIPOLAR DISORDER AND CREATIVE POTENTIAL found either (Q = 4.78, df = 5, p = .443,  = 0.28). Interestingly, all effects were positive except for the effect of TTCT-F (d = -0.37, p = .117).

DT scoring. The scoring of DT tests used, namely either fluency, flexibility, originality, elaboration, or total, did not explain any variance in effect sizes (Q = 1.07, df = 4, p = .898, = 0.30). The direction of the effect for flexibility (d = -0.05, p = .889) was negative, while it was positive for fluency (d = 0.09, p = .302), originality (d = 0.05, p

= .628), elaboration (d = 0.03, p = .933), and total (d = 0.18, p = .068). Once again, it is important to mention that flexibility and elaboration each only comprised one effect size from one study.

DT domain. Whether the domain of the DT test was verbal, figural, mixed, or numeric had no significant influence on the variance of the effect size (Q = 5.12, df = 3, p

= .162,  = 0.26). Interestingly, the direction of the effect differed between domains: While the figural (d = -0.12, p = .390) and numeric domains (d = -0.07, p = .705) had a negative effect size estimate, the verbal (d = 0.14, p = .056) and mixed domains (d = 0.22, p = .026) had a positive effect size estimate (verbal by trend). Notably though, the two domains with a negative connection contained a small amount of studies and effect sizes (figural: k = 3, m =

6; numeric: k = 1, m = 4). Contrast analyses showed a significant contrast between the figural and mixed domain (d = -0.33, p = .047), but this no longer remained significant after Holm- adjusting the p-value for multiple testing (p = .283).

DT time. The time to perform the DT test had no significant moderating effect (Q =

0.10, df = 1, p = .748,  = 0.28).

DT instruction. Instruction classified as either quantity or quality of ideas did not significantly affect the overall variance of effects (Q = 0.07, df = 1, p = .789,  = 0.28).

DT calculation. The calculation of the DT test as either summed, averaged, or other did not significantly explain any variance in the overall effect size (Q = 1.48, df = 2, p = .477, BIPOLAR DISORDER AND CREATIVE POTENTIAL

 = 0.28). However, the majority of studies and effect sizes reported summed scores (k = 8, m

= 33), while only a handful of studies reported either averaged (k = 4, m = 7) or other (k = 2, m = 2) scorings. One study with one effect size did not report information on how DT scores were calculated and it was not obtainable through other context information regarding the specific DT test.

Analysis of the Moderation Effects Regarding BD

BD presentation. Presentation of BD as bipolar I, bipolar mixed, or subclinical did not show a significant moderating effect (Q = 1.10, df = 2, p = .576,  = 0.28). The effect for the subclinical presentation was significant (d = 0.17, p = .039), but not for the clinical presentations.

BD status. The BD status showed a significant moderating effect (Q = 18.86, df = 3, p < .001,  = 0.20). While a depressed status had a significant, negative effect (d = -0.51, p

< .001), a status of euthymic (d = 0.14, p = .043) or subclinical (d = 0.17, p = .001) had a significantly positive effect. A manic status had a positive yet not significant effect (d = 0.25, p = .097). Significant contrasts were found between depressed and all other status presentations (depressed and euthymic: d = -0.65, p < .001; depressed and subclinical: d = -

0.68, p < .001; depressed and manic: d = -0.76, p < .001) that remained significant after

Holm-adjusting for the p-value (p < .001; p < .001, p = .001). Importantly, all of the four effect sizes for depressed and for manic statuses stemmed from the same study.

BD assessment. The tool of assessment for BD showed no significant effect on the variance of the overall effect size (Q = 1.77, df = 2, p = .413,  = 0.28). Interestingly, assessment using the HPS yielded a significant positive effect (d = 0.21, p = .029).

Clinical diagnosis. BD being diagnosed by a professional clinician (either fully, partly, or not at all) did not have a moderating effect on the overall variance of the effect size BIPOLAR DISORDER AND CREATIVE POTENTIAL

(Q = 1.86, df = 2, p = .394,  = 0.28). However, only no clinical diagnosis was significant (d

= 0.17, p = .036).

Comorbidity. The existence of comorbidities had no moderating effect on the overall effect (Q = 0.10, df = 1, p = .755,  = 0.35). However, in total, many studies (k = 8, m = 17) did not report whether participants had comorbidities; of these, two were clinical studies with four effect sizes, and the rest were subclinical studies.

Medication. Whether all or some participants in the sample took medication for BD had no significant moderating effect (Q = 0.84, df = 1, p = .360,  = 0.37). Importantly, the group of all participants taking medication only consisted of one study with 16 effect sizes, while the mixed group was made up of four studies with nine effect sizes. Overall, six studies

(13 effect sizes) did not report any information on the use of medication for BD.

Associations Between the Moderators

For an overview about frequencies of all characteristics of moderators, see Tables B2 for categorial moderators and B3 for continuous moderators in the Appendix. To examine possible confounding between moderators, their associations were also computed. As indicated by Lipsey (2003), confounding between moderators in meta-analyses is rather common, since moderator variables of interest are frequently related. Therefore, it is important to also take confounding of moderators into account to make more precise and non- misleading interpretations. In the current meta-analysis, moderators of different levels of measurements were used, which is why interrelationships were calculated using Pearson’s correlations, point biserial correlations, Eta (), and Cramer’s V. Pearson’s correlations were used for interrelationships between continuous moderator variables; point biserial correlations were used for interrelationships between continuous and categorial

(dichotomous) variables;  was used for continuous and categorial (polytomous) variables; and Cramer’s V was used for interrelationships between categorial variables. BIPOLAR DISORDER AND CREATIVE POTENTIAL

As anticipated, associations were found between most of the moderators. See Table

B4 in the Appendix for an overview of all interrelationships between moderators. However, because of missing data and uneven distributions, not all associations between moderators were computed. For these reasons, interrelationships also have to be interpreted with care.

Regarding the bivariate relationships, most of the results in the above-mentioned moderator analyses should be handled with care, especially those showing numerous highly significant intercorrelations, such as country, study type, mean age, age group, DT test, DT time, BD presentation, BD assessment, and clinical diagnosis.

Influencing Studies

To see whether any studies or effect sizes had particularly high influence, a Baujat plot for the random-effects meta-analysis was conducted; it can be found in Figure 4. In the

Baujat plot, three effect sizes from two studies seem to be influential. These three effect sizes were negative and relatively high regarding absolute values (d = -0.97; d = -0.77; d = -0.64).

However, an outlier analysis did not show a significant influence of these three or any other effect sizes. Nonetheless, to control for the possible influence of these three effect sizes apparent in the Baujat plot, an overall effect size was computed once again without the potentially influential effect sizes. When excluding the potentially influential effect sizes from the overall analysis, only small changes in the overall effect size could be observed (d =

0.16, 95% CI: [0.08, 0.15], p < .001, Δdwith-without influential effect sizes = 0.05), self-evidently resulting in a larger effect size. Nevertheless, the non-significant outlier analysis as well as the results from the adapted random-effects meta-analysis indicate that the influential studies can be included in further analyses, since the difference in effect size due their exclusion was negligible. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Figure 4. Baujat plot to identify influential studies. Vertical axis shows contribution to the overall effect, and horizontal axis shows heterogeneity.

Publication Bias

To test the occurrence of a publication bias, we conducted a funnel plot, trim-and-fill- test, Egger’s test, and rank correlation test. Visual analysis reveals symmetry of the funnel plot, as can be seen in Figure 5. The trim-and-fill test also indicated no missing studies on the left side. Further, the moderator analysis of publication status as either published or unpublished yielded no significant difference in publication status (Q = 4.25, df = 2, p = .119,

 = 0.28). However, only one unpublished study was included, and most unpublished studies were not obtainable within the time the meta-analysis was conducted. In line with this, the

Egger’s test (z = -1.51, p = .132) as well as the rank correlation test ( = -0.09, p = .458) also indicated no publication bias. BIPOLAR DISORDER AND CREATIVE POTENTIAL

In addition, Rosenthal’s Fail-safe N was calculated to see how many additional studies would be needed to reduce the observed significant effect (p < .001) to a target significant effect (p = .05), indicating that 167 studies would be needed. As introduced by Mullen,

Muellerleile, and Bryant (2001), the Fail-safe ratio calculated by N / ( 5×k+10 ), in which k refers to the number of experiments or studies, is a useful post hoc rule of thumb to interpret the Fail-safe N.If the ratio score exceeds 1, it can be assumed that the weight of the included studies is tolerant for future results (Mullen et al., 2001). In this case 167/(5×13+10)=2.23 is larger than 1, indicating that the meta-analysis is robust against a publication bias.

When examining the appearance of the funnel plot for the whole sample of effect sizes, it seems evident that most of the standard errors are placed on few vertical lines within the plot. This might be a result of the use of different scorings for DT tasks; as such, we conducted more in-depth analyses for the individual subsets of scoring methods for a possible publication bias. Regarding the scoring, a publication bias for fluency (z = -0.67, p = .504; 

= -0.11, p = .591) seems unlikely. However, for originality (z = -2.67, p = .008;  = -0.55, p =

.026), the Egger’s test and the rank test indicate that a publication bias cannot be ruled out, though the trim-fill test shows that no study is missing on the left side of the funnel. Funnel plots for fluency and originality can be found in Figures B3 and B4 in the Appendix.

Estimates of publication biases for the scorings of flexibility and elaboration could not be computed due to the small number of effect sizes.

Estimates of publication biases were also conducted for the two subsets of the moderator study design. Publication biases seem unlikely for the two subsets including the mean difference subset (z = -0.98, p = .326;  = 0.03, p = .870) and the correlational subset (z

= -1.11, p = .266;  = -0.24, p = .182). The trim-fill test also yielded no additional need for further studies on the left side for the mean difference subset, but it did indicate the need for two additional studies on the right side for the correlational subset. It can therefore be BIPOLAR DISORDER AND CREATIVE POTENTIAL assumed that some studies with positive mediocre or large effects within the correlational subset are missing; this is unlike an ordinary publication bias, where negative or unfavorable effects usually remain unpublished (e.g. Easterbrook, Gopalan, Berlin, & Matthews, 1991).

Funnel plots for the two subsets can be found in Figure B5 and B6 in the Appendix. Further, the funnel plot with additional studies after the trim-fill test for the correlation subset can be found in Figure B7 in the Appendix. In sum, a publication bias does not seem to be an important issue in the current meta-analysis.

Figure 5. Funnel plot for the overall effect size for analyzing publication bias.

Discussion

The aim of the current meta-analysis was to get an overview of all existing empirical research on the relationship between creative potential using DT and BD as well as to BIPOLAR DISORDER AND CREATIVE POTENTIAL estimate the overall mean effect size of this relationship. While the view of a “mad genius” often romanticizes a possible link between BD and creativity, people with BD suffer not only from their illness itself but usually also from stigma and a negative view on the disorder. To simultaneously get away from the popular view of a mad genius and to also to allow for dismantling the stigma around bipolar disorder, an empirical meta-analysis on the relationship between BD and creativity was needed.

In the current meta-analysis, the overall effect size was significantly positive yet diminutively small. Regarding the conventions for Cohen’s d values, an effect size with d =

0.2 upwards can be considered small (Cohen, 1988), which means that the found effect cannot even be regarded as small. However, the results suggest that people with diagnosed

BD and people with higher subclinical scores related to BD indeed show marginally better performance in DT tasks. An important implication of this finding is that BD does not exclusively come with impairments; instead, some areas such as creative thinking might even be slightly improved, yet not on a grand scale. However, the improvement of DT is certainly not as drastic as suggested by the mad genius theory; therefore, having BD should not be seen as something desirable that comes along with a genius level of creative thinking. In fact, the difference in DT performance between healthy and currently manic individuals as well as between euthymic and currently manic individuals was small and not significant, indicating that DT performance is not significantly better during mania. Further, DT performance during a depressive episode was even worse compared to people with a healthy or a euthymic status, which indicates that BD may even temporarily dampen DT performance. Hence, in sum, benefits of BP for creative thinking are practically very limited.

Some other interesting results were found in the moderator analyses. Considering the study and sample level, the only significant moderator was the country in which the study was conducted. Here, one study conducted in Turkey (by Alici et al., 2019) differed from the BIPOLAR DISORDER AND CREATIVE POTENTIAL others in showing a relatively large, negative effect size. When analyzing influential studies, this particular study was also found to appear striking in the Baujat plot. Even though cultural differences cannot be ruled out, another reason for this study to stand out may be because this study was the only to use a neuroimaging paradigm. While all other studies implemented DT assessments in the usual manner using pen-and-paper versions, Alici and colleagues (2019) used functional near-infrared spectroscopy to simultaneously measure brain activity.

However, a mechanism based solely on the laboratory setting of the study design appears to be not most reasonable from a theoretical point of view. Hence, sound explanations for this outlier study seems to be rather hard to find.

Regarding the moderators on the level of DT and BD, the DT-related moderators as studied in this work did not seem to be of special significance, while moderators for BD yielded some interesting findings. Regarding DT, the only variable with a significant influence was the classification of mixed regarding to the DT domain, which was associated with a slightly stronger positive association between BP and DT. Most studies classified as mixed used the ATTA, and thus merged the verbal and figural domains of the TTCT, or used total creativity scores which may have increased the reliability of the DT assessment.

However, the explanatory power of a mixed domain is questionable, since it can be seen as an umbrella term for studies using non-distinct or confounding categorization of their DT tasks.

Unfortunately, no difference between the verbal and the figural domains was found, which would have been of greater interest than the mixed domain. It would further be interesting whether the BP-DT relationship may be moderated by the type of originality scoring. It was proposed that mental illness may especially related to uncommon responses as measured with uniqueness scores but maybe less so to creative responses as assessed by creativity ratings

(Fink et al., 2014). This question may be addressed in future research when more studies with these different DT originality scorings are available. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Regarding BD moderators, a handful of variables seem to be of importance: First, only BD presentation as subclinical as opposed to bipolar I and bipolar mixed showed a significant positive effect. These findings are consistent with the notion of a U-shaped association between psychopathology and creativity (Abraham, 2014), as only subclinical expressions of BP, located in the middle of the continuum of psychopathological severity between healthy and clinical illness, were associated with higher creativity. This corroborates the notion that mild deviations of thought processes can support creative thinking, whereas full-blown clinical psychopathologies are usually associated with mental impairments.

Moreover, BD status as either euthymic, subclinical, depressed, or manic had a moderating effect, indicating that a patient’s current affective episode might influence the strength of the effect, with especially depressive episodes implying reduced creative potential. Further, non-clinical assessments of BD explained a significant amount of variance as well as the HPS assessment tool for BD, in contrast to the DSM and others. This might be explained by the intercorrelation of the moderators, since correlational studies mainly made use of the HPS as an assessment tool, and this approach can further be classified as non- clinical.

Another striking finding of the present meta-analysis is the fact that empirical research on the relationship between DT and BD is still scarce: Even though we conducted a thorough literature search with over 6,000 hits, in the end no more than 13 articles fit the research subject and met the inclusion criteria. Although strict inclusion criteria were applied, which might have reduced the total number of studies, these inclusion criteria were necessary to assure a certain level of empirical quality. Given the small number of included studies, the estimated effect size has to be interpreted with caution, and the implications of the moderator analyses are also limited. Hence, there is a need for a larger body of work surrounding creativity and BD, and future research can consider several aspects to improve study quality. BIPOLAR DISORDER AND CREATIVE POTENTIAL

For instance, severity, duration of the illness, onset, number of episodes, hospitalization, and detailed information on patients’ medication were rarely reported, which is why these moderators could not be included and analyzed in this meta-analysis. Only one of the five studies examining mean differences gave extensive information on patients’ history with BD, listing the mean age during the first manic and depressive episodes, number of manic and depressive episodes, information on hospitalization due to manic or depressive episodes as well as the specific dosage of various medications (Johnson, Tharp, & Holmes, 2015).

Further, only one study systematically examined the differences in DT during an affective episode as opposed to euthymic affective states: Rybakowski and Klonowska (2011) measured DT twice in their sample, once during a manic or depressive episode and once in remission after a manic or depressive episode. They found that DT performance was significantly lower during depressive episodes compared to manic episodes and compared to remission. This allows for a more detailed insight, making clear that BD is far too complex to only compare bipolar patients to healthy controls; one should also examine the changes across different episodes of BD, or at least consider the current episode status.

Another issue that came up regarding study quality was the assessment of DT. Even though all studies oriented their DT scoring on the four scorings of fluency, flexibility, originality, and elaboration, as suggested by Guilford (1950, 1957), calculations of these scorings were rather discordant. Due to high intercorrelations of the four scorings, especially between fluency and originality (e.g. Forthmann, Szardenings, et al., 2020), the scorings are heavily confounded. Confounding between fluency and originality can have a conceptual reason (the equal-odds rule assumes that there is a greater likelihood for original ideas when there is a greater number of overall ideas; Forthmann et al., 2021; Forthmann, Szardenings, et al., 2020), but it is commonly just due to summative scoring across responses, which makes all DT facets necessarily dependent on idea fluency. Indeed, a handful of studies did not BIPOLAR DISORDER AND CREATIVE POTENTIAL elaborate on their method of calculating originality, and over half of the studies chose to calculating originality by simply summing all original ideas. Several scoring methods have been proposed to avoid a necessary fluency confound (Reiter-Palmon et al., 2019) including the scoring (Forthmann, Szardenings, et al., 2020), snapshot scoring (Silvia et al., 2009), or top-scoring (Benedek et al., 2013). Average scoring was done for only seven effect sizes from four studies, but this should not only be more commonly applied when investigating creativity and BD, but it should be used more generally in all areas of creativity research.

The same issue arose for studies using a total score, where either the assessment of

DT domains was not specified at all or sub-domains were simply added up together into a total score for creativity. The problem with this type of total score also lies in the high intercorrelations between DT domains, since the explanatory power of such a total score remains obscure. As fluency is the number of ideas, flexibility is the number of categories of ideas, and originality is the uniqueness and remoteness of ideas, what is the meaning of the total score? In the meta-analysis, the total score was used as the only scoring method for DT in five studies with eight effect sizes; these studies did not report individual sub-scores for fluency, flexibility, originality, or elaboration. This undermines a better understanding of how

BP relates to quantitative versus qualitative facets of creative ideation. Another two studies with five effect sizes reported total scores in addition to scorings of the individual facets.

Hence, in sum over half the studies used total scores as a scoring type. The current meta- analysis did control for the use of total scorings beforehand, since analyses were also calculated with a subsample of effect sizes in which additional total scores were excluded; yet, their exclusion did not yield any remarkable differences. Thus, for simplicity, the analyses were performed using all the extracted effect sizes. Overall though, given that the quality of DT assessment tended to be problematic, since confounded methods were used for BIPOLAR DISORDER AND CREATIVE POTENTIAL total scores, the mixed domain, and summation indices, it is hardly surprising that no significant moderators for DT assessment were observed.

Even though the main focus here is to examine the studies included in the meta- analysis, it also makes sense to examine studies that were not included: Some studies were excluded for not meeting the inclusion criteria but can still add to the discussion on the relationship between DT and BD. For example, Ghadirian, Gregoire, and Kosmidis (2001) did not meet the inclusion criterion of having a healthy CG, but they had a noteworthy approach by comparing bipolar patients to patients with other psychopathologies. The study accounted for the severity of the illness, which was neglected by most of the included studies.

Even though the difference in DT scores between bipolar patients and patients with other psychopathologies was not significant, severity seems to play an important role in DT performance (Ghadirian et al., 2001). Interestingly, moderately ill patients showed the best creativity scores, even better than mildly ill or recovered patients, while severely ill patients showed the worst performance. These results are again in line with the idea of an inverted U- shaped relationship between psychopathology and creativity, which has often been put forward in the literature (Abraham, 2014; Richards et al., 1988; Simonton, 2019). However, more studies that make use of specific diagnostic instruments to measure severity of BD are needed to better understand the role of severity in DT performance.

Even though the current meta-analysis makes an important contribution to the more elaborated research on a possible psychopathology-creativity link, specifically on a link between BD and DT, it has some limitations. First, not all studies on the relevant topic were obtainable at the time the meta-analysis was conducted, which might have resulted in an incomplete sample of studies. Most studies that could not be obtained were unpublished papers or doctoral dissertations. Even though the meta-analysis showed no indication for a publication bias, including more unpublished studies would have helped obtain a full picture BIPOLAR DISORDER AND CREATIVE POTENTIAL of all existing studies. Further, despite contacting the authors, three studies could not be included due to the lack of an effect size. Including these studies would have strengthened the statistical and informational power of the current meta-analysis, as the final number of 13 studies is rather small.

Second, interpretational insights were limited as a result of unevenly distributed and highly confounded moderators. Noticeably, most moderators were strongly associated with each other, revealing covariations in study design factors, but they may be partly overestimated due to a small number of included studies and effect sizes, as well as infrequent and missing data on moderator presentations. Some presentations of moderators were represented fewer than four times, making the interpretation of their moderating effects rather difficult (Fu et al., 2011). More precisely, this concerned the presentations England and Turkey in the moderator country, since only three effect sizes from England and one from

Turkey were present; the presentations school children and mixed in the moderator age group, since the two presentations were only represented with one effect size each; the presentations flexibility and elaboration in the moderator DT scoring, since only one effect size was included for each; as well as the presentation other in the moderator DT calculation method, which had only two effect sizes. Furthermore, for some moderators, studies generally did not provide a great amount of information, which resulted in a lot of missing values within these moderators. Of concern were the moderators comorbidity and medication since only five of the 13 studies reported information on these two moderators. Taken together, this might possibly result in higher intercorrelations between moderators.

Additionally, some intercorrelations can also be explained by the characteristics of the studies and moderators themselves. All included studies were relatively similar in terms of the number of participants, with similar mean age and a similar ratio of male and female participants. Also, study designs in the included studies can all be regarded as quite similar BIPOLAR DISORDER AND CREATIVE POTENTIAL because most of the studies made use of similar assessment tools such as the DSM or HPS for

BD and the AUT or TTCT for DT. Therefore, it makes sense that other moderators related to the assessment tool are also associated with one another, such as DT time, DT domain, or DT instruction being related to the DT test.

Another similar aspect to consider, which was mentioned before, is the fact that this meta-analysis included both clinical and subclinical samples. When taking a closer look at this, interrelationships also make sense: All the subclinical studies used correlational designs, and all the correlational studies used non-clinical assessment methods in the form of BD questionnaires, mainly the HPS. Therefore, it comes as no surprise that a correlational design was associated with a subclinical presentation of BD, with a BD status as healthy, with a BD assessment tool using the HPS or other questionnaires, as well as with no information on comorbidities and medication. Certain converse associations were also apparent, then, for clinical studies, as most of the clinical studies used mean differences (apart from one that reported both mean differences and correlations). The studies using mean differences usually also made use of clinical assessment tools for BD, such as the DSM, had BD presentations as either bipolar I or mixed, had BD status as either euthymic, manic, or depressed, as well as had information on comorbidities and medication. Taken together, intercorrelations between moderators make sense and are also commonly reported as a problem of meta-analyses

(Lipsey, 2003). Nonetheless, given the high intercorrelations between moderators, findings from moderator analyses need to be interpreted with caution.

Third, the current meta-analysis was not able to directly test predictions regarding an inverted U-shaped relationship between DT and BD. By separating clinical and subclinical samples this study made a first attempt to map different levels of impairment to investigate the shape of the relationship, but it is questionable whether solely separating between clinical and subclinical samples is sufficient to illustrate the whole spectrum of impairment. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Nonetheless, studies from correlational designs, which represented subclinical studies, showed larger effect sizes than studies comparing mean differences, which represented clinical studies. This alone is not sufficient to illustrate an inverted U-shaped relationship between impairment due to BD and performance in DT, but it does indicate that subclinical samples, which can be considered to have less impairment, tended to show better performance on DT than clinical samples. It would have been of great interest to use severity as a variable to parametrically investigate differences in DT performance as a function of severity levels of BD.

Even though including clinical and subclinical samples was insightful, using both clinical and subclinical studies could also be considered as a limitation, since a full-blown clinical BD diagnosis is not equivalent to heightened subclinical scores on BD-related questionnaires. However, the number of studies would have been even smaller if only clinical studies were included, making results from the meta-analysis and moderator analyses less meaningful. Importantly, most correlational studies used diagnostic instruments based on the

HPS, which has been found to correlate highly with current hypomanic symptoms (Kim,

2016). Therefore, it can be assumed that both the clinical and subclinical studies measured the same construct in regards to BD.

Nonetheless, to control for this crucial difference in the study samples (subclinical vs. clinical), we conducted moderator analyses for study design as either correlational

(subclinical samples) or mean differences (clinical samples). The moderator analysis of study design yielded no significant differences for the overall moderator, but it showed that a significant BP-DT effect was only observed in subclinical studies. We therefore cannot rule out that differences between clinical and subclinical samples existed but remain obscured due to missing power. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Despite the limitations listed above, the current meta-analysis helps to shed light onto the relationship between BD and creativity. This was the first meta-analysis to date that used a theoretical framework other than the “mad genius” hypothesis as a basis to assume a relationship between DT and BD in one direction or the other. Further, it is the first meta- analysis to date that focused on BD exclusively, linking it to little-c creativity as measured by

DT, without pooling them with related but different constructs such as convergent thinking creativity or verbal fluency. The results generally indicate that individuals with diagnosed BD

(outside of a depressive episode) or higher subclinical ratings of BD indeed perform slightly better on DT tasks, indicating increased creative potential, but not to the magnitude that is sometimes postulated. Even though the small number of studies and the qualitative deficits of some studies limited statistical possibilities, the moderator analyses also indicate that the relationship might be even more complex, since some moderators on the study and sample level as well as moderators around BD had moderating effects.

Overall, this meta-analysis could not identify a succinct “yes-or-no” answer to the question of whether people with BD are more creative than healthy people. Even though the overall effect size in the current meta-analysis indicates that individuals with BD might indeed have a slight advantage (if any) in DT, it still underlines the need for future quality studies on the link between BD and creativity.

Open Practice Statement

The literature search documentation, data, and analysis script for this meta-analysis are openly available at https://osf.io/tnksa/ and the meta-analysis was not preregistered . BIPOLAR DISORDER AND CREATIVE POTENTIAL

References

Abraham, A. (2014). Is there an inverted-U relationship between creativity and

psychopathology? Frontiers in Psychology, 5(750). doi:10.3389/fpsyg.2014.00750

Adaman, J. E., & Blaney, P. H. (1995). The effects of musical mood induction on

creativity. The Journal of Creative Behavior, 29(2), 95-108. doi:10.1002/j.2162-

6057.1995.tb00739

*Alici, Y. H., Özgüven, H. D., Kale, E., Yenihayat, I., & Baskak, B. (2019). Prefrontal

Activity Measured by Functional Near Infrared Spectroscopy During Divergent and

Convergent Thinking in Bipolar Disorder. Archives of Neuropsychiatry, 56(2), 86.

doi:10.29399/npa.23203

Alloy, L. B., Abramson, L. Y., Urosevic, S., Walshaw, P. D., Nusslock, R., & Neeren, A. M.

(2005). The psychosocial context of bipolar disorder: Environmental, cognitive, and

developmental risk factors. Clinical Psychology Review, 25(8), 1043-1075.

doi:10.1016/j.cpr.2005.06.006

Altman, E. G., Hedeker, D., Peterson, J. L., & Davis, J. M. (1997). The Altman self-rating

mania scale. Biological Psychiatry, 42(10), 948-955. doi:10.1016/S0006-

3223(96)00548-3

Amabile, T. M. (1982). Social psychology of creativity: A consensual assessment

technique. Journal of Personality and Social Psychology, 43(5), 997.

doi:10.1037/0022-3514.43.5.997

American Psychiatric Association. (2000). Diagnostic and Statistical Manual of Mental

Disorders (4th ed.). Washington, DC: Author.

American Psychiatric Association. (2013). Diagnostic and Statistical Manual of Mental

Disorders (5th ed.). Washington, DC: Author. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Andreasen, N. C. (1997). Creativity and mental illness: Prevalence rates in writers and their

first-degree relatives. Eminent Creativity, Everyday Creativity, and Health, 144(10), 7-

18. doi:10.1176/ajp.144.10.1288

Angst, J., Adolfsson, R., Benazzi, F., Gamma, A., Hantouche, E., Meyer, T. D., ... & Scott, J.

(2005). The HCL-32: towards a self-assessment tool for hypomanic symptoms in

outpatients. Journal of Affective Disorders, 88(2), 217-233.

doi:10.1016/j.jad.2005.05.011

Ashby, F. G., & Isen, A. M. (1999). A neuropsychological theory of positive affect and its

influence on cognition. Psychological Review, 106(3), 529. Retrieved from

http://www.factorhappiness.at/downloads/quellen/S2_Ashby.pdf

Baas, M., De Dreu, C. K., & Nijstad, B. A. (2008). A meta-analysis of 25 years of mood-

creativity research: Hedonic tone, activation, or regulatory focus? Psychological

Bulletin, 134(6), 779. doi:10.1037/a0012815

Baas, M., Nijstad, B. A., Boot, N. C., & De Dreu, C. K. (2016). Mad genius revisited:

Vulnerability to psychopathology, biobehavioral approach-avoidance, and

creativity. Psychological Bulletin, 142(6), 668. doi:10.1037/bul0000049

Baddeley, A. D. (1968). A 3 min reasoning test based on grammatical

transformation. Psychonomic Science, 10(10), 341-342. Retrieved from

https://link.springer.com/content/pdf/10.3758%2FBF03331551.pdf

Baruch, I., Hemsley, D. R., & Gray, J. A. (1988). Latent inhibition and “psychotic proneness”

in normal subjects. Personality and Individual Differences, 9(4), 777-783.

doi:10.1016/0191-8869(88)90067-0

Baudson, T. G. (2016). The mad genius stereotype: Still alive and well. Frontiers in

Psychology, 7, 368. doi:10.3389/fpsyg.2016.00368 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Baudson, T. G., & Preckel, F. (2013). Teachers' implicit personality theories about the gifted:

An experimental approach. School Psychology Quarterly, 28(1), 37.

doi:1037/spq0000011

Beaty, R. E., & Silvia, P. J. (2012). Why do ideas get more creative across time? An

executive interpretation of the serial order effect in divergent thinking tasks.

Psychology of Aesthetics, Creativity, and the Arts, 6(4), 309–319.

doi:10.1037/a0029171

Beghetto, R. A., & Kaufman, J. C. (2007). Toward a broader conception of creativity: A case

for" mini-c" creativity. Psychology of Aesthetics, Creativity, and the Arts, 1(2), 73.

doi:10.1037/1931-3896.1.2.73

Benazzi, F., & Akiskal, H. S. (2001). Delineating bipolar II mixed states in the Ravenna–San

Diego collaborative study: the relative prevalence and diagnostic significance of

hypomanic features during major depressive episodes. Journal of Affective

Disorders, 67(1-3), 115-122. doi:10.1016/S0165-0327(01)00444-X

Benedek, M., & Jauk, E. (2019). Creativity and cognitive control. In J. Kaufman & R.

Sternberg (Eds.), Cambridge handbook of creativity (pp. 200-223). Cambridge:

Cambridge University Press. doi:10.1017/9781316979839.012

Benedek, M., Mühlmann, C., Jauk, E., & Neubauer, A. C. (2013). Assessment of divergent

thinking by means of the subjective top-scoring method: Effects of the number of top-

ideas and time-on-task on reliability and validity. Psychology of Aesthetics, Creativity,

and the Arts, 7(4), 341. doi:10.1037/a0033644

Berrettini, W. H. (2000). Susceptibility loci for bipolar disorder: overlap with inherited

vulnerability to schizophrenia. Biological Psychiatry, 47(3), 245-251.

doi:10.1016/S00006-3223(99)00226-7 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Borenstein, M., Hedges, L. V., Higgins, J. P., & Rothstein, H. R. (2011). Introduction to

Meta-Analysis. Hoboken, NJ: John Wiley & Sons.

Bowden, E. M., & Jung-Beeman, M. (2003). Normative data for 144 compound remote

associate problems. Behavior Research Methods, Instruments, & Computers, 35(4),

634-639. doi:10.3758/bf03195543

Bujacz, A., Dunne, S., Fink, D., Gatej, A., Karlsson, E., Ruberti, V., & Wronska, M. (2014).

Does creativity make you happy? The influence of creative activity on hedonic and

eudaimonic well-being. Journal of European Psychology Students, 5(2).

doi:10.5334/jeps.by

Carlsson, I., Wendt, P. E., & Risberg, J. (2000). On the neurobiology of creativity.

Differences in frontal activity between high and low creative

subjects. Neuropsychologia, 38(6), 873-885. doi:10.1016/S0028-3932(99)00128-1

Carson, S. H. (2011). Creativity and psychopathology: A shared vulnerability model. The

Canadian Journal of Psychiatry, 56(3), 144-153. doi:10.1177/070674371105600304

Carson, S. H., Peterson, J. B., & Higgins, D. M. (2003). Decreased latent inhibition is

associated with increased creative achievement in high-functioning

individuals. Journal of Personality and Social Psychology, 85(3), 499.

doi:10.1037/0022-3514.85.3.499

Carson, S. H., Peterson, J. B., & Higgins, D. M. (2005). Reliability, validity, and factor

structure of the creative achievement questionnaire. Creativity Research Journal, 17(1),

37-50. doi:10.1207/s15326934crj1701_4

Chai, X. J., Whitfield-Gabrieli, S., Shinn, A. K., Gabrieli, J. D., Castanón, A. N., McCarthy,

J. M., ... & Öngür, D. (2011). Abnormal medial prefrontal cortex resting-state

connectivity in bipolar disorder and schizophrenia. Neuropsychopharmacology, 36(10),

2009. Retrieved from https://www.nature.com/articles/npp201188.pdf BIPOLAR DISORDER AND CREATIVE POTENTIAL

Chermahini, S. A., & Hommel, B. (2012). Creative mood swings: divergent and convergent

thinking affect mood in opposite ways. Psychological Research, 76(5), 634-640.

doi:10.1007/s00426-011-0358-z

Clapham, M. M. (2001). The effects of affect manipulation and information exposure on

divergent thinking. Creativity Research Journal, 13(3-4), 335-350.

doi:10.1207/S15326934CRJ1334_11

Claridge, G., & Blakey, S. (2009). Schizotypy and affective temperament: Relationships with

divergent thinking and creativity styles. Personality and Individual Differences, 46(8),

820-826. doi:10.1016/j.paid.2009.01.015

Clark, J. R., & de la Motto, A. L. (1992). [Summary]. The paradox of genius and madness:

Seneca and his influence. Cuadernos de filología clásica: Estudios latinos, (2), 189-

200. Retrieved from https://dialnet.unirioja.es/servlet/articulo?codigo=165794

Cohen, J. (1988). Statistical power analysis for the behavioral sciences. Hillsdale, NJ:

Lawrence Erlbaum Associates.

Corbalán, F. J., Martínez, F., Donolo, D., Tejerina, M., & Limiñana, R. M. (2003). CREA

Inteligencia Creativa. Una medida cognitiva de la creatividad. [Creative Intelligence: A

cognitive measure of creativity.] Madrid: TEA Ediciones.

Corrigan, P.W., & Bink, A. B. (2016). The Stigma of Mental Illness. In H. S. Friedman (Ed.),

Encyclopedia of Mental Health (Second Edition) (pp. 230-234). Oxford: Academic

Press.

Cropley, A. (2006). In praise of convergent thinking. Creativity research journal, 18(3), 391-

404. doi:10.1207/s15326934crj1803_13

Davis, M. A. (2009). Understanding the relationship between mood and creativity: A meta-

analysis. Organizational Behavior and Human Decision Processes, 108(1), 25-38.

doi:10.1016/j.obhdp.2008.04.001 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Dietrich, A. (2014). The mythconception of the mad genius. Frontiers in Psychology, 5, 79.

doi:10.3389/fpsyg.2014.00079

*Drapeau, C. W., & DeBrule, D. S. (2013). The relationship of hypomania, creativity, and

suicidal ideation in undergraduates. Creativity Research Journal, 25(1), 75-79.

doi:10.1080/10400419.2013.752231

Drevdahl, J. E. (1956). Factors of importance for creativity. Journal of Clinical Psychology,

12(1), 21-26. doi:10.1002/1097-4679

Dumas, D., & Dunbar, K. N. (2014). Understanding fluency and originality: A latent variable

perspective. Thinking Skills and Creativity, 14, 56-67. doi:10.1016/j.tsc.2014.09.003

Duncker, K., & Lees, L. S. (1945). On problem-solving. Psychological Monographs, 58(5),

1-113. doi:10.1037/h0093599

Easterbrook, P. J., Gopalan, R., Berlin, J. A., & Matthews, D. R. (1991). Publication bias in

clinical research. The Lancet, 337(8746), 867-872. doi:10.1016/0140-6736(91)90201-Y

Ebstein, R. P., Novick, O., Umansky, R., Priel, B., Osher, Y., Blaine, D., ... & Belmaker, R.

H. (1996). Dopamine D4 receptor (D4DR) exon III polymorphism associated with the

human personality trait of novelty seeking. Nature Genetics, 12(1), 78.

doi:10.1038/ng0196-78

Eckblad, M., & Chapman, L. J. (1986). Development and validation of a scale for hypomanic

personality. Journal of Abnormal Psychology, 95(3), 214.

doi:10.1037//0021-843x.95.3.214

Ellsworth, P. C., & Smith, C. A. (1988). From appraisal to emotion: Differences among

unpleasant feelings. Motivation and Emotion, 12(3), 271-302. Retrieved from

https://link.springer.com/content/pdf/10.1007/BF00993115.pdf BIPOLAR DISORDER AND CREATIVE POTENTIAL

Feist, G. J. (1998). A meta-analysis of personality in scientific and artistic

creativity. Personality and Social Psychology Review, 2(4), 290-309.

doi:10.1207/s15327957pspr0204_5

Fink, A., Benedek, M., Unterrainer, H.-F., Papousek, I., & Weiss, E.M. (2014). Creativity

and psychopathology: Are there similar mental processes involved in creativity and in

psychosis-proneness? Frontiers in Psychology, 5, 1211. doi: 10.3389/fpsyg.2014.01211

Forthmann, B., Gerwig, A., Holling, H., Çelik, P., Storme, M., & Lubart, T. (2016). The be-

creative effect in divergent thinking: the interplay of instruction and object

frequency. Intelligence, 57, 25-32. doi:10.1016/j.intell.2016.03.005

Forthmann, B., Lips, C., Szardenings, C., Scharfen, J., & Holling, H. (2020). Are speedy

brains needed when divergent thinking is speeded—or unspeeded?. The Journal of

Creative Behavior, 54(1), 123-133. doi:10.1002/jocb.350

Forthmann, B., Szardenings, C., & Dumas, D. (2021). On the conceptual overlap between the

fluency contamination effect in divergent thinking scores and the chance view on

scientific creativity. The Journal of Creative Behavior, 55(1), 268-275.

doi:10.1002/jocb.445

Forthmann, B., Szardenings, C., & Holling, H. (2020). Understanding the confounding effect

of fluency in divergent thinking scores: Revisiting average scores to quantify artifactual

correlation. Psychology of Aesthetics, Creativity, and the Arts, 14(1), 94–112.

doi:10.1037/aca0000196

Frasier, M. M. (1988). Torrance Verbal and Figural Tests: Measuring General Creative

Thinking Processes. Retrieved from ERIC database. (ED304466)

Fredrickson, B. L., & Branigan, C. (2005). Positive emotions broaden the scope of attention

and thought‐ action repertoires. Cognition & Emotion, 19(3), 313-332.

doi:10.1080/02699930441000238 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Fu, R., Gartlehner, G., Grant, M., Shamliyan, T., Sedrakyan, A., Wilt, T. J., ... & Santaguida,

P. (2011). Conducting quantitative synthesis when comparing medical interventions:

AHRQ and the Effective Health Care Program. Journal of Clinical

Epidemiology, 64(11), 1187-1197. doi:10.1016/j.jclinepi.2010.08.010

*Furnham, A., Batey, M., Anand, K., & Manfield, J. (2008). Personality, hypomania,

intelligence and creativity. Personality and Individual Differences, 44(5), 1060-1069.

doi:10.1017/j.paid.2007.10.035

Ghadirian, A. M., Gregoire, P., & Kosmidis, H. (2001). Creativity and the evolution of

psychopathologies. Creativity Research Journal, 13(2), 145-148.

doi:10.1207/S15326934CRJ1302_2

Goff, K., & Torrance, E. P. (2002). The Abbreviated Torrance Test for Adults (ATTA).

Bensenville, IL: Scholastic Testing Service.

Göritz, A. S., & Moser, K. (2003). Mood and flexibility in categorization: A conceptual

replication. Perceptual and Motor Skills, 97(1), 107-119.

doi:10.2466/pms.2003.97.1.107

Gray, N. S., Pickering, A. D., Hemsley, D. R., Dawling, S., & Gray, J. A. (1992). Abolition

of latent inhibition by a single 5 mg dose of d-amphetamine in

man. Psychopharmacology, 107(2-3), 425-430. Retrieved from

https://link.springer.com/content/pdf/10.1007/BF02245170.pdf

Greene, T. R., & Noice, H. (1988). Influence of positive affect upon creative thinking and

problem solving in children. Psychological Reports, 63(3), 895-898.

doi:10.2466/pr0.1988.63.3.895

Guilford, J. P. (1950). Creativity. American Psychologist, 5(9), 444–454.

doi:10.1037/h0053487 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Guilford, J. P. (1957). Creative abilities in the arts. Psychological Review, 64 (2), 110-118.

doi:10.1037/h0048280

Guilford, J. P. (1967). The nature of human intelligence. New York City, NY: McGraw-Hill.

Guilford, J. P., Christensen, P. R., Merrifield, P. R., & Wilson, R. C. (1978). Alternate Uses:

Manual of instructions and interpretations. Palo Alto, CA: Mind Garden.

Guilford, J. P., & Hoepfner, R. (1971). The analysis of intelligence. New York City, NY:

McGraw-Hill.

Harbour, W. (2015). The big bang theory: Mad geniuses and the freak show of higher

education. Review of Disability Studies: An International Journal, 11(2). Retrieved

from https://scholarspace.manoa.hawaii.edu/bitstream/

10125/58630/1/1543.pdf

Haro, G., Calabrese, J. R., Larsson, C., Shirley, E. R., Martín, E., Leal, C., & Delgado, P. L.

(2007). The relationship of personality traits to substance abuse in patients with bipolar

disorder. European Psychiatry, 22(5), 305-308. doi:10.1016/j.eurpsy.2007.03.009

Harrington, D. M. (1975). Effects of explicit instructions to “be creative” on the

psychological meaning of divergent thinking test scores. Journal of Personality, 43(3),

434-454. doi:10.1111/j.1467-6494.1975.tb00715.x

Hathaway, S. R., McKinley, J. C., & MMPI Restandardization Committee. (1989). MMPI-2:

Minnesota Multiphasic Personality Inventory-2: manual for administration and

scoring. Minneapolis, MN: University of Minnesota Press.

Henry, C., Sorbara, F., Lacoste, J., Gindre, C., & Leboyer, M. (2001). Antidepressant-

induced mania in bipolar patients: identification of risk factors. Journal of Clinical

Psychiatry, 62(4), 249-255. Retrieved from

https://pdfs.semanticscholar.org/a6f5/8bb06192bb5b23f7aa1c6c07d30864246fad.pdf BIPOLAR DISORDER AND CREATIVE POTENTIAL

Hirt, E. R., Melton, R. J., McDonald, H. E., & Harackiewicz, J. M. (1996). Processing goals,

task interest, and the mood–performance relationship: A mediational analysis. Journal

of Personality and Social Psychology, 71(2), 245. doi:10.1037/0022-3514.71.2.245

Hocevar, D. (1979). Ideational fluency as a confounding factor in the measurement of

originality. Journal of Educational Psychology, 71(2), 191-196. doi:10.1037/0022-

0663.71.2.191

Hubbard, E. M., & Ramachandran, V. S. (2005). Neurocognitive mechanisms of

synesthesia. Neuron, 48(3), 509-520. doi:10.1016/j.neuron.2005.10.012

Isen, A. M. (1999). On the relationship between affect and creative problem solving. Affect,

Creative Experience, and Psychological Adjustment, 3, 17. Retrieved from

https://books.google.de/books?

hl=en&lr=&id=FTbvCgAAQBAJ&oi=fnd&pg=PA3&ots=bfZy0-5G-

z&sig=PHqzHfyVxLKj_UqYYwwVTHjnKdo&redir_esc=y#v=onepage

&q&f=false

Isen, A. M., Daubman, K. A., & Nowicki, G. P. (1987). Positive affect facilitates creative

problem solving. Journal of Personality and Social Psychology, 52(6), 1122. Retrieved

from http://www.factorhappiness.at/downloads/quellen/S11_Isen.pdf

Isen, A. M., & Means, B. (1983). The influence of positive affect on decision-making

strategy. Social Cognition, 2(1), 18-31. doi:10.1521/soco.1983.2.1.18

Jäger, A. O., Süß, H. M., & Beauducel, A. (1997). Berliner Intelligenzstruktur-Test: BIS-Test.

Göttingen: Hogrefe.

Jamison, K. R. (1993). Touched with fire: Manic–depressive illness and the artistic

temperament. New York, NY: Free Press.

Jamison, K. R. (1996). Touched with fire. New York, NY: Simon and Schuster. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Johnson, S. L., Edge, M. D., Holmes, M. K., & Carver, C. S. (2012). The behavioral

activation system and mania. Annual Review of Clinical Psychology, 8, 243-267.

doi:10.1146/annurev-clinpsy-032511-143148

*Johnson, S. L., Tharp, J. A., & Holmes, M. K. (2015). Understanding creativity in bipolar I

disorder. Psychology of Aesthetics, Creativity, and the Arts, 9(3), 319.

doi:10.1037/a0038852

Kasprow, W. J., Catterson, D., Schachtman, T. R., & Miller, R. R. (1984). Attenuation of

latent inhibition by post-acquisition reminder. The Quarterly Journal of Experimental

Psychology, 36(1), 53-63. doi:10.1080/14640748408402194

Kaufman, J. C., & Beghetto, R. A. (2009). Beyond big and little: The four c model of

creativity. Review of General Psychology, 13(1), 1-12. doi:10.1037/a0013688

Kaufman, J. C., Bromley, M. L., & Cole, J. C. (2006). Insane, poetic, lovable: Creativity and

endorsement of the “Mad Genius” stereotype. Imagination, Cognition and

Personality, 26(1), 149-161. doi:10.2190/J207-3U30-R401-446J

Kaufman, G., & Vosburg, S. K. (1997). ‘‘Paradoxical” effects of mood on creative problem

solving. Cognition and Emotion, 11, 151-170. doi:10.1080/026999397379971

Kendler, K. S., McGuire, M., Gruenberg, A. M., O'hare, A., Spellman, M., & Walsh, D.

(1993). The Roscommon family study: IV. Affective illness, anxiety disorders, and

alcoholism in relatives. Archives of General Psychiatry, 50(12), 952-960.

doi:10.1001/archpsyc.1993.01820240036005

Kim, B. N. (2016). Testing the Behavioral Activation System Model of Bipolar Disorder in

the High-Risk Group (Unpublished doctoral dissertation). Seoul National University,

Seoul, Korea. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Kim, K. H. (2008). Meta‐ analyses of the relationship of creative achievement to both IQ and

divergent thinking test scores. The Journal of Creative Behavior, 42(2), 106-130.

doi:10.1002/j.2162-6057.2008.tb01290.x

Kim, B. N., & Kwon, S. M. (2017). The link between hypomania risk and creativity: The role

of heightened behavioral activation system (BAS) sensitivity. Journal of Affective

Disorders, 215, 9-14. doi:10.1016/j.jad.2017.02.033

Kleinkorres, R., Forthmann, B., & Holling, H. (2021). An Experimental Approach to

Investigate the Involvement of Cognitive Load in Divergent Thinking. Journal of

Intelligence, 9(1), 3. doi:10.3390/jintelligence9010003

Konstantopoulos, S. (2011). Fixed effects and variance components estimation in three‐ level

meta‐ analysis. Research Synthesis Methods, 2(1), 61-76. coi:10.1002/jrsm.35

Koukopoulos, A., & Sani, G. (2014). DSM‐ 5 criteria for depression with mixed features: a

farewell to mixed depression. Acta Psychiatrica Scandinavica, 129(1), 4-16.

doi:10.1111/acps.12140

Krishnan, K. R. R. (2005). Psychiatric and medical comorbidities of bipolar

disorder. Psychosomatic Medicine, 67(1), 1-8.

doi:10.1097/01.psy.0000151489.36347.18

Kupka, R. W., Luckenbaugh, D. A., Post, R. M., Leverich, G. S., & Nolen, W. A. (2003).

Rapid and non-rapid cycling bipolar disorder: a meta-analysis of clinical studies. The

Journal of Clinical Psychiatry, 64(12), 1483-1494. doi:10.4088/jcp.v64n1213

Kyaga, S., Lichtenstein, P., Boman, M., Hultman, C., Långström, N., & Landen, M. (2011).

Creativity and mental disorder: family study of 300 000 people with severe mental

disorder. The British Journal of Psychiatry, 199(5), 373-379.

doi:10.1192/bjp.bp.110.085316 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Lauronen, E., Veijola, J., Isohanni, I., Jones, P. B., Nieminen, P., & Isohanni, M. (2004).

Links between creativity and mental disorder. Psychiatry: Interpersonal and Biological

Processes, 67(1), 81-98. doi:10.1521/psyc.67.1.81.31245

Lichtenstein, P., Yip, B. H., Björk, C., Pawitan, Y., Cannon, T. D., Sullivan, P. F., &

Hultman, C. M. (2009). Common genetic determinants of schizophrenia and bipolar

disorder in Swedish families: a population-based study. The Lancet, 373(9659), 234-

239. doi:10.1016/S0140-6736(09)60072-6

Lipsey, M. W. (2003). Those confounded moderators in meta-analysis: Good, bad, and

ugly. The Annals of the American Academy of Political and Social Science, 587(1), 69-

81. doi:10.1177/0002716202250791

Locke, B. A., & Putman, P. (1971). Center for epidemiological studies depression scale.

Lubow, R. E., Ingberg-Sachs, Y., Zalstein-Orda, N., & Gewirtz, J. C. (1992). Latent

inhibition in low and high “psychotic-prone” normal subjects. Personality and

Individual Differences, 13(5), 563-572. doi:10.1016/0191-8869(92)90197-W

Lubow, R. E., & Moore, A. U. (1959). Latent inhibition: the effect of nonreinforced pre-

exposure to the conditional stimulus. Journal of Comparative and Physiological

Psychology, 52(4), 415. doi:10.1037/h0046700

Ludwig, A. M. (1995). The price of greatness: Resolving the creativity and madness

controversy. New York City, NY:Guilford Press.

Lyubomirsky, S., King, L., & Diener, E. (2005). The benefits of frequent positive affect:

Does happiness lead to success? Psychological Bulletin, 131(6), 803.

doi:10.1037/0033-2909.131.6.803

MacGill, M. (2019, February 22). The difference between bipolar and depression. Medical

News Today. Retrieved from https://www.medicalnewstoday.com/articles/314582.php BIPOLAR DISORDER AND CREATIVE POTENTIAL

Mansell, W., Morrison, A. P., Reid, G., Lowens, I., & Tai, S. (2007). The interpretation of,

and responses to, changes in internal states: an integrative cognitive model of mood

swings and bipolar disorders. Behavioural and Cognitive Psychotherapy, 35(5), 515-

539. doi:10.1017/S1352465807003827

Martin, L. L., Ward, D. W., Achee, J. W., & Wyer, R. S. (1993). Mood as input: People have

to interpret the motivational implications of their moods. Journal of Personality and

Social Psychology, 64(3), 317. doi:10.1037/0022-3514.64.3.317

Martindale, C. (1999). Biological bases of creativity. In R. J. Sternberg (Ed.), Handbook of

Creativity (pp. 137-152). Cambridge: Cambridge University Express.

McCraw, S., Parker, G., Fletcher, K., & Friend, P. (2013). Self-reported creativity in bipolar

disorder: prevalence, types and associated outcomes in mania versus

hypomania. Journal of Affective Disorders, 151(3), 831-836.

doi:10.1010/j.jad.2013.07.016

McGuffin, P., Rijsdijk, F., Andrew, M., Sham, P., Katz, R., & Cardno, A. (2003). The

heritability of bipolar affective disorder and the genetic relationship to unipolar

depression. Archives of General Psychiatry, 60(5), 497-502.

doi:10.1001/archpsyc.60.5.497

Mednick, S. A. (1968). The remote associates test. The Journal of Creative Behavior, 2(3),

213-214. doi:10.1002/j.2162-6057.1968.tb00104.x

Merikangas, K. R., Jin, R., He, J. P., Kessler, R. C., Lee, S., Sampson, N. A., ... & Ladea, M.

(2011). Prevalence and correlates of bipolar spectrum disorder in the world mental

health survey initiative. Archives of General Psychiatry, 68(3), 241-251.

doi:10.1001/archgenpsychiatry.2011.12 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Millon, T., Davis, R. D., & Millon, C. (1997). Millon Clinical Multiaxial Inventory-III

[Manual Second Edition]. Retrieved from https://www.dottoressabonasera.it/wp-

content/uploads/2018/03/MCMI-III-tra-diagnosi-di-personalità-e-novità.pdf

Mitchell, P. B., Wilhelm, K., Parker, G., Austin, M. P., Rutgers, P., & Malhi, G. S. (2001).

The clinical features of bipolar depression: a comparison with matched major

depressive disorder patients. The Journal of Clinical Psychiatry, 62, 212-216.

doi:10.4088/JCP.v62n0314a

Moher, D., Liberati, A., Tetzlaff, J., & Altman, D. G. (2009). Preferred reporting items for

systematic reviews and meta-analyses: the PRISMA statement. Annals of Internal

Medicine, 151(4), 264-269. doi:10.1371/journal.pmed-1000100

Mouchiroud, C., & Lubart, T. (2001). Children's original thinking: An empirical examination

of alternative measures derived from divergent thinking tasks. The Journal of Genetic

Psychology, 162(4), 382-401. doi:10.1080/00221320109597491

Mullen, B., Muellerleile, P., & Bryant, B. (2001). Cumulative meta-analysis: a consideration

of indicators of sufficiency and stability. Personality and Social Psychology

Bulletin, 27(11), 1450-1462. doi:10.1177/01461672012711006

Murray, G., & Johnson, S. L. (2010). The clinical significance of creativity in bipolar

disorder. Clinical Psychology Review, 30, 721-732. doi:10.1016/j.cpr.2010.05.006

*Nadasi, C. (1997). Relationship between creativity self-actualization and hypomania.

Graduate Student Theses, Dissertations, & Professional Papers, 4939. Retrieved from

https://scholarworks.umt.edu/cgi/viewcontent.cgi?article=5974&

context=etd

Nijstad, B. A., De Dreu, C. K., Rietzschel, E. F., & Baas, M. (2010). The dual pathway to

creativity model: Creative ideation as a function of flexibility and persistence. BIPOLAR DISORDER AND CREATIVE POTENTIAL

European Review of Social Psychology, 21(1), 34-77.

doi:10.1080/10463281003765323

Paek, S. H., Abdulla, A. M., & Cramond, B. (2016). A meta-analysis of the relationship

between three common psychopathologies—ADHD, anxiety, and depression—and

indicators of little-c creativity. Gifted Child Quarterly, 60(2), 117-133.

doi:10.1177/0016986216630600

Papworth, M. A., & James, I. A. (2003). Createativity and Mood: Towards a Model of

Cognitive Mediation. The Journal of Creative Behavior, 37(1), 1-16.

doi:10.1002/j.2162-6057.2003.tb00823.x

Pavuluri, M. N., Henry, D. B., Nadimpalli, S. S., O’Connor, M. M., & Sweeney, J. A. (2006).

Biological risk factors in pediatric bipolar disorder. Biological Psychiatry, 60(9), 936-

941. doi:10.1016/j.biopsych.2006.04.002

Peterson, J. B., & Carson, S. (2000). Latent inhibition and openness to experience in a high-

achieving student population. Personality and individual Differences, 28(2), 323-332.

doi:10.1016/S0191-8869(99)00101-4

Peterson, J. B., Smith, K. W., & Carson, S. (2002). Openness and extraversion are associated

with reduced latent inhibition: Replication and commentary. Personality and Individual

Differences, 33(7), 1137-1147. doi:10.1016/S0191-8869(02)00004-1

Plucker, J. A., Qian, M., & Wang, S. (2011). Is originality in the eye of the beholder?

Comparison of scoring techniques in the assessment of divergent thinking. The Journal

of Creative Behavior, 45(1), 1-22. doi:10.1002/j.2162-6057.2001.tb01081.x

Ponterotto, J. G., & Reynolds, J. D. (2013). The “genius” and “madness” of Bobby Fischer:

His life from three psychobiographical lenses. Review of General Psychology, 17(4),

384-398. doi:10.1037/a0033246 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Power, R. A., Steinberg, S., Bjornsdottir, G., Rietveld, C. A., Abdellaoui, A., Nivard, M.

M., ... & Stefansson, K. (2015). Polygenic risk scores for schizophrenia and bipolar

disorder predict creativity. Nature Neuroscience, 18(7), 953-955. doi:10.1038/nn.4040

Preckel, F., Wermer, C., & Spinath, F. M. (2011). The interrelationship between speeded and

unspeeded divergent thinking and reasoning, and the role of mental speed. Intelligence,

39(5), 378-388. doi:10.1016/j.intell.2011.06.007

Ramachandran, V. S., & Hubbard, E. M. (2001). Synaesthesia--a window into perception,

thought and language. Journal of consciousness studies, 8(12), 3-34. Retrieved from

https://sites.oxy.edu/clint/physio/article/synaesthesiaawindowintoperception.pdf

R Core Team. (2018). R: A language and environment for statistical computing. Vienna,

Austria: R Foundation for Statistical Computing. Retrieved from http://www.r-

project.org/

Reiter-Palmon, R., Forthmann, B., & Barbot, B. (2019). Scoring divergent thinking tests: A

review and systematic framework. Psychology of Aesthetics, Creativity, and the Arts,

13(2), 144-152. doi:10.1037/aca0000227

Reuter, M., Panksepp, J., Schnabel, N., Kellerhoff, N., Kempel, P., & Hennig, J. (2005).

Personality and biological markers of creativity. European Journal of Personality:

Published for the European Association of Personality Psychology, 19(2), 83-95.

doi:10.1002/per.534

Rhodes, M. (1961). An analysis of creativity. The Phi Delta Kappan, 42(7), 305-310.

Richards, R., Kinney, D. K., Benet, M., & Merzel, A. P. (1988). Assessing everyday

creativity: Characteristics of the Lifetime Creativity Scales and validation with three

large samples. Journal of Personality and Social Psychology, 54(3), 476.

doi:10.1037/0022-3514.54.3.476 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Rihmer, Z., & Kiss, K. (2002). Bipolar disorders and suicidal behaviour. Bipolar

Disorders, 4(1), 21-25. doi:10.1034/j.1399-5618.4.s1.3.x

*Rodrigue, A. L., & Perkins, D. R. (2012). Divergent thinking abilities across the

schizophrenic spectrum and other psychological correlates. Creativity Research

Journal, 24(2-3), 163-168. doi:10.1080/10400419.2012.677315

Ruiter, M., & Johnson, S. L. (2015). Mania risk and creativity: A multi-method study of the

role of motivation. Journal of Affective Disorders, 170, 52-58.

doi:10.1016/j.jad.2014.08.049

Runco, M. A. (2007). A hierarchical framework for the study of creativity. New Horizons in

Education, 55(3), 1-9.

Runco, M. A., & Acar, S. (2012). Divergent thinking as an indicator of creative

potential. Creativity Research Journal, 24(1), 66-75.

doi:10.1080/10400419.2012.652929

Runco, M. A., Illies, J. J., & Eisenman, R. (2005). Creativity, originality, and

appropriateness: What do explicit instructions tell us about their relationships?. The

Journal of Creative Behavior, 39(2), 137-148. doi:10.1002/j.2162-6057.2005.tb01255.x

Runco, M. A., & Jaeger, G. J. (2012). The standard definition of creativity. Creativity

Research Journal, 24(1), 92-96. doi:10.10400419.2012.650092

Runco, M. A., & Mraz, W. (1992). Scoring divergent thinking tests using total ideational

output and a creativity index. Educational and Psychological Measurement, 52(1), 213-

221. doi:10.1177/001316449205200126

Runco, M. A., & Okuda, S. M. (1991). The instructional enhancement of the flexibility and

originality scores of divergent thinking tests. Applied Cognitive Psychology, 5(5), 435-

441. doi:10.1002/acp.2350050505 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Russ, S.W. (1999). Emotion/affect. In M. A. Runco & S. R. Pritzker (Eds)., Encyclopedia of

Creativity (Vol. 1, pp. 659-668). San Diego, CA: Academic Press

*Rybakowski, J. K., & Klonowska, P. (2011). Bipolar mood disorder, creativity and

schizotypy: an experimental study. Psychopathology, 44(5), 296-302.

doi:10.1159/000322814

Sachs, G. S., Printz, D. J., Kahn, D. A., Carpenter, D., & Docherty, J. P. (2000). The expert

consensus guideline series: medication treatment of bipolar disorder. Postgrad Med, 1,

1-104. Retrieved from

http://files.getaufc.webnode.com.br/200000231-e511ee60c3/Bipolar_2000.pdf

Said-Metwaly, S., Fernández-Castilla, B., Kyndt, E., & Van den Noortgate, W. (2020).

Testing conditions and creative performance: Meta-analyses of the impact of time

limits and instructions. Psychology of Aesthetics, Creativity, and the Arts, 14(1), 15–38.

doi:10.1037/aca0000244

Sánchez, Ó., Martín-Brufau, R., Méndez, F. X., Corbalán, F. J., & Limiñana, R. M. (2010).

The relationship between optimism, creativity and psychopathological symptoms in

university students. Electronic Journal of Research in Educational Psychology, 8,

1151-1178. Retrieved from ERIC database. (EJ911609)

*Santosa, C. M., Strong, C. M., Nowakowska, C., Wang, P. W., Rennicke, C. M., & Ketter,

T. A. (2007). Enhanced creativity in bipolar disorder patients: A controlled

study. Journal of Affective Disorders, 100(1-3), 31-39. doi:10.1016/j.jad.2006.10.013

Schildkraut, J. J., Hirshfeld, A. J., & Murphy, J. M. (1994). Mind and mood in modern art: II.

Depressive disorders, spirituality, and early deaths in the abstract expressionist artists of

the New York School. The American Journal of Psychiatry, 151, 482-488. doi:10.1176/

ajp.151.4.482 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Schlesinger, J. (2009). Creative mythconceptions: A closer look at the evidence for the" mad

genius" hypothesis. Psychology of Aesthetics, Creativity, and the Arts, 3(2), 62.

doi:10.1037/a0013975

*Schuldberg, D. (1990). Schizotypal and hypomanic traits, creativity, and psychological

health. Creativity Research Journal, 3(3), 218-230. doi:10.1080/10400419009534354

Schuldberg, D. (2001). Six subclinical spectrum traits in normal creativity. Creativity

Research Journal, 13(1), 5-16. doi.1207/S15326934CRJ1301_2

Scott, J., Colom, F., & Vieta, E. (2007). A meta-analysis of relapse rates with adjunctive

psychological therapies compared to usual psychiatric treatment for bipolar

disorders. International Journal of Neuropsychopharmacology, 10(1), 123-129.

doi:10.1017/S1461145706006900

*Siegel, J., & Bugg, J. M. (2016). Dissociating divergent thinking and creative achievement

by examining attentional flexibility and hypomania. Psychology of Aesthetics,

Creativity, and the Arts, 10(4), 416. doi:10.1037/aca0000071

Silvia, P. J., & Kimbrel, N. A. (2010). A dimensional analysis of creativity and mental

illness: Do anxiety and depression symptoms predict creative cognition, creative

accomplishments, and creative self-concepts?. Psychology of Aesthetics, Creativity,

and the Arts, 4(1), 2. doi:10.1037/a0016494

Silvia, P. J., Martin, C., & Nusbaum, E. C. (2009). A snapshot of creativity: Evaluating a

quick and simple method for assessing divergent thinking. Thinking Skills and

Creativity, 4(2), 79-85. doi:10.1016/j.tsc.2009.06.005

Silvia, P. J., Winterstein, B. P., Willse, J. T., Barona, C. M., Cram, J. T., Hess, K. I., ... &

Richard, C. A. (2008). Assessing creativity with divergent thinking tasks: Exploring the

reliability and validity of new subjective scoring methods. Psychology of Aesthetics,

Creativity, and the Arts, 2(2), 68. doi:10.1037/1931-3896.2.2.68 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Simeonova, D. I., Chang, K. D., Strong, C., & Ketter, T. A. (2005). Creativity in familial

bipolar disorder. Journal of Psychiatric Research, 39(6), 623-631.

doi:10.1016/j.jpsychires.2005.01.005

Simonton, D. K. (2010). Creative thought as blind-variation and selective-retention:

Combinatorial models of exceptional creativity. Physics of Life Reviews, 7(2), 156-179.

doi:10.1016/j.plrev.2010.02.002

Simonton, D. K. (2019). Creativity and psychopathology: The tenacious mad-genius

controversy updated. Current Opinion in Behavioral Sciences, 27, 17-21.

doi:10.1016/j.cobeha.2018.07.006

Smith, R. E., & Yang, X. (2004). Toward a general theory of creativity in advertising:

Examining the role of divergence. Marketing Theory, 4(1-2), 31-58.

doi:10.1177/147059310404486

Soeiro-de-Souza, M. G., Dias, V. V., Bio, D. S., Post, R. M., & Moreno, R. A. (2011).

Creativity and executive function across manic, mixed and depressive episodes in

bipolar I disorder. Journal of Affective Disorders, 135(1-3), 292-297.

doi:10.1016/j.jad.2011.06.024

Solano, C. H. (1987). Stereotypes of social isolation and early burnout in the gifted: Do they

still exist?. Journal of Youth and Adolescence, 16(6), 527-539. Retrieved from

https://link.springer.com/content/pdf/10.1007/BF02138819.pdf

*Srivastava, S., Childers, M. E., Baek, J. H., Strong, C. M., Hill, S. J., Warsett, K. S., ... &

Ketter, T. A. (2010). Toward interaction of affective and cognitive contributors to

creativity in bipolar disorders: A controlled study. Journal of Affective

Disorders, 125(1-3), 27-34. doi:10.1016/j.jad.2009.12.018

Sternberg, R. J. (2012). The assessment of creativity: An investment-based

approach. Creativity research journal, 24(1), 3-12. doi:10.1080/10400419.2012.652925 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Takeuchi, H., Taki, Y., Hashizume, H., Sassa, Y., Nagase, T., Nouchi, R., & Kawashima, R.

(2011). Failing to deactivate: the association between brain activity during a working

memory task and creativity. Neuroimage, 55(2), 681-687.

doi:10.1016/j.neuroimage.2010.11.052

Taylor, C. L. (2017). Creativity and mood disorder: A systematic review and meta-

analysis. Perspectives on Psychological Science, 12(6), 1040-1076.

doi:10.1177/17456917699653

Thys, E., Sabbe, B., & De Hert, M. (2014). The assessment of creativity in

creativity/psychopathology research–a systematic review. Cognitive

Neuropsychiatry, 19(4), 359-377. doi:10.1080/13546805.2013.877384

Torrance, E. P. (1974). The Torrance Tests of Creative Thinking: Norms-Technical Manual.

Princeton, NJ: Personal Press.

Torrance, E. P. (1990). The Torrance Tests of Creative Thinking: Norms - Technical Manual

Figural (Streamlined) Forms A & B. Bensenville, IL: Scholastic Testing Service.

Torrance, E. P. (1998). The Torrance Tests of Creative Thinking: Norms - Technical Manual

Figural (Streamlined) Forms A & B. Bensenville, IL: Scholastic Testing Service.

Torrance, E. P. (2008). The Torrance Tests of Creative Thinking: Norms - Technical Manual

Figural (Streamlined) Forms A & B. Bensenville, IL: Scholastic Testing Service.

Verhaeghen, P., Joorman, J., & Khan, R. (2005). Why we sing the blues: the relation between

self-reflective rumination, mood, and creativity. Emotion, 5(2), 226.

doi:10.1037/1528-3542.5.2.226

Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal

of Statistical Software, 36(3), 1-48. Retrieved from ftp://ctan.uib.no/pub/cran/web/

packages/metafor/vignettes/metafor.pdf BIPOLAR DISORDER AND CREATIVE POTENTIAL

*Von Stumm, S., Chung, A., & Furnham, A. (2011). Creative ability, creative ideation and

latent classes of creative achievement: What is the role of personality? Psychology of

Aesthetics, Creativity, and the Arts, 5(2), 107. doi:10.1037/a0020499

Vosburg, S. K. (1998). Mood and the quantity and quality of ideas. Creativity Research

Journal, 11(4), 315-324. doi:10.1207/s15326934crj1104_5

Wechsler, D. (2008). Wechsler adult intelligence scale–Fourth Edition (WAIS–IV). San

Antonio, TX: NCS Pearson, 22, 498. doi:10.2298/psi171001001l

Weisberg, R. W. (1994). Genius and madness?: A Quasi-experimental test of the hypothesis

that manic-depression increases creativity. Psychological Science, 5(6), 361-367.

doi:10.1111/j.1467-98280.1994.tb00286.x

Welsh, G. S., & Barron, F. (1959). Barron-Welsh Art Scale: a Portion of the Welsh Figure

Preference Test. Psychologists Press.

Whitfield, J. B., Nightingale, B. N., OBrien, M. E., Heath, A. C., Birley, A. J., & Martin, N.

G. (1998). Molecular biology of alcohol dependence, a complex polygenic

disorder. Clinical Chemistry and Laboratory Medicine, 36(8), 633-636.

doi:10.1515/CCLM.1998.111

Whitfield-Gabrieli, S., Thermenos, H. W., Milanovic, S., Tsuang, M. T., Faraone, S. V.,

McCarley, R. W., ... & Wojcik, J. (2009). Hyperactivity and hyperconnectivity of the

default network in schizophrenia and in first-degree relatives of persons with

schizophrenia. Proceedings of the National Academy of Sciences, 106(4), 1279-1284.

doi:10.1073/pnas.0809141106

Wills, G. I. (2003). Forty lives in the bebop business: Mental health in a group of eminent

jazz musicians. The British Journal of Psychiatry, 183(3), 255-259.

doi:10.1192/bjp.183.3.255 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Yamada, Y., & Nagai, M. (2015). Positive mood enhances divergent but not convergent

thinking. Japanese Psychological Research, 57(4), 281-287. doi:10.1111/jpr.12093

Yamamoto, K. (1964). A further analysis of the role of creative thinking in high-school

achievement. The Journal of Psychology, 58(2), 277-283.

doi:10.1080/00223980.1964.9916745

Young, R. C., Biggs, J. T., Ziegler, V. E., & Meyer, D. A. (1978). A rating scale for mania:

reliability, validity and sensitivity. The British Journal of Psychiatry, 133(5), 429-435.

doi:10.1192/bjp.133.5.429

Young, L. N., Winner, E., & Cordes, S. (2013). Heightened incidence of depressive

symptoms in adolescents involved in the arts. Psychology of Aesthetics, Creativity, and

the Arts, 7(2), 197. doi:10.1037/a0030468

*Zabelina, D. L., Condon, D., & Beeman, M. (2014). Do dimensional psychopathology

measures relate to creative achievement or divergent thinking? Frontiers in

psychology, 5, 1029. doi:10.3389/fpsyg.2014.01029

Zenasni, F., & Lubart, T. (2002). Effects of mood states on creativity. Current Psychology

Letters, (8), 33-50. Retrieved from

https://www.researchgate.net/profile/Franck_Zenasni

/publication/30428186_Effects_of_Mood_States_on_Creativity/links/02e7e52aa01a5a8

345000000.pdf BIPOLAR DISORDER AND CREATIVE POTENTIAL

Appendices

Appendix A

Table A1

Coding scheme. level variable explanation study authors publication year age of study years as of 2019 country country of conduction published yes, no type of study mean difference, correlation sample mean age mean age in years age group adults ( < 25), young adults (18 - 25), school children ( > 18), mixed gender % female matched CG yes, no IQ testa IQ test used IQa mean of IQ BD BD presentation bipolar I, bipolar mixed, subclinical BD status subclinical (healthy), euthymic, manic, depressed BD assessment tool name of instrument BD assessment category based on DSM, HPS or other clinical diagnosis yes, no, partly number of episodesa number of affective episodes onset agea age of first affective episode hospitalizationa mean duration of hospitalization (in months) durationa mean duration of illness (in years) severitya mean of the diagnosis instrument used comorbidity yes, no medication all, mixed DT DT measure name of the instrument DT test based on AUT, TTCT, or other DT scoring fluency, flexibility, originality, elaboration, total DT domain verbal, figural, numeric, mixed DT time given time restriction to the task (i.e., time in minutes per item) DT instruction quantity, quality DT calculation summed, averaged Notes. IQ = Intelligence Quotient. DSM = Diagnostic and Statistical Manual of Mental

Disorders. HPS = Hypomanic Personality Scale. AUT = Alternate Uses Test. TTCT = BIPOLAR DISORDER AND CREATIVE POTENTIAL

Torrance Test of Creative Thinking a Not enough information was reported in the studies to further analyze the variable. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Table A2

Overview over BD assessment tools used in the included studies. level author name of BD assessment tool professional APA (2000) Diagnostic and Statistical Manual of Mental Disorders (DSM-IV; 4th ed., text rev.) self Angst et al. (2005) Hypomania Checklist-32 (HCL-32)

Eckblad & Chapman, 1986 Hypomanic personality scale (HPS)

Altman, Hedeker, Peterson, Altman Self-Rating Mania Scale (ASRM) & Davis (1997)

Young, Biggs, Ziegler, & Young Mania Rating Scale (YMRS) Meyer (1978)

Hathaway, McKinley, & Minnesota Multiphasic Inventory 2nd Edition MMPI Restandardization (MMPI-2) Committee (1989)

Millon, Davis, & Million Millon Clinical Multitaxial Inventory-III (1997) (MCMI-III) BIPOLAR DISORDER AND CREATIVE POTENTIAL

Table A3

Overview over DT assessment tools used in the included studies.

domain name of DT test paradigm verbal Alternate Uses Task (AUT; list as many uses as possible for various common Guilford, 1967) objects (e.g. a paper box, a piece of coal, a half glass of water)

Unusual Uses Task (UUT; list as many unusual uses as possible for various Guilford, 1967) common objects (e.g. a paperclip, a blanket, a barrel)

Unusual Uses Task (UUT; think of unusual uses for various common objects Guilford, Christensen, Merrifield, (e.g. shoes, a button, a key) & Wilson, 1978)

Alternate Uses Task (AUT; think of creative uses for various common objects Guilford, 1978 with Harrington’s modification; Harrington, 1975)

Torrance Test of Creative Test battery with various verbal DT tasks (e.g. just Thinking Verbal (TTCT-V; suppose task: list as many problems as possible if a Torrance, 1990) hypothetical situation was possible)

figural Torrance Test of Creative Test battery with various figural DT tasks (e.g. Thinking Figural (TTCT-F; construct and complete the drawings from existing Torrance, 1990) lines and structures as for example triangles)

mixed Torrance Test of Creative Test battery with various DT tasks, comprised of Thinking (TTCT; Torrance, 1990) verbal and figural ones (see TTCT-V and TTCT-F for more detailed description)

Abbreviated Torrance Test for Shorter version of the TTCT comprised of three Adults (ATTA; Goff & Torrance, tasks (one verbal, two figural). 2002)

Berlin Intelligence Structure Test Test battery with several DT tasks, verbal, figural, (BIS; Jäger, Süss, & Beauducel, and numeric (e.g. masselon test, characteristics and 1997); inventiveness scale abilities test, insight test, possible applications test) BIPOLAR DISORDER AND CREATIVE POTENTIAL

Appendix B

Table B1

Descriptive information on variables including number of studies, effect sizes, and participants per variable as well as mean, standard deviation, and Q-statistic for moderators.

level variable m k N M SD Q df p study age of study 13 42 1,857 9.77 7.83 1.02 1 .31 country 12.13 4 .02* England 2 3 313 Turkey 1 1 58 Poland 1 16 88 USA 9 22 1,398 study type 5.33 2 .07 mean difference 5 19 1,441 correlational 9 23 528 sample mean age 11 35 1,575 28.01 7.54 2.88 1 .09 age group 5.47 4 .24 adults 5 25 528 young adults 6 15 1,119 school children 1 1 128 mixed 1 1 82 gender (% f) 12 40 1,672 60.11 7.01 0.78 1 .38 matched CG 0.42 2 .81 yes 2 17 146 no 3 8 382 DT DT test 4.09 3 .25 AUT 7 14 1,220 TTCT 5 12 549 other 1 16 88 DT scoring 4.66 5 .46 fluency 5 16 610 flexibility 1 1 97 originality 6 11 752 elaboration 1 1 97 total 7 13 1,074 DT domain 9.46 4 .05 verbal 10 20 1,578 figural 3 6 358 numeric 1 4 88 mixed 4 12 367 DT time 10 33 1,575 3.12 0.85 0.10 1 .75 DT instruction 3.96 2 .14 quantity 6 26 576 quality 8 16 1,381 DT calculation 4.62 3 .20 summed 8 32 962 averaged 4 7 458 other 2 2 285 BD BD presentation 5.00 3 < .001*** bipolar I 2 5 170 BIPOLAR DISORDER AND CREATIVE POTENTIAL

bipolar mixed 4 21 440 subclinical 7 16 1,247 BD status 25.02 4 < .001*** healthy 8 17 1,329 euthymic 5 17 528 manic 1 4 88 depressed 1 4 88 BD assessment 5.74 3 .13 DSM 5 23 528 HPS 6 12 1,157 other 4 7 494 clinical diagnosis 5.87 3 .12 yes 4 23 411 no 3 7 1,247 partly 2 16 199 comorbidity 1.06 2 .59 yes 2 7 202 no 3 18 227 medication 1.26 1 .53 all 1 16 88 mixed 4 9 440 Notes. m = number of studies. k = number of effect sizes. N = sample size. Q = Cochran’s Q test for heterogeneity. df = degrees of freedom.

* p < .05, *** p < .001.

Drapeau & DeBrule, 2013 -0.02 [-0.41, 0.37] Drapeau & DeBrule, 2013 -0.08 [-0.52, 0.36] Drapeau & DeBrule, 2013 0.01 [-0.38, 0.40] Furnham, Batey, Anand, & Manifield, 2008 0.41 [ 0.07, 0.75] Johnson, Tharp, & Holmes, 2015 -0.22 [-0.61, 0.17] Johnson, Tharp, & Holmes, 2015 0.54 [ 0.15, 0.93] Rodrigue & Perkins, 2012 0.08 [-0.36, 0.52] Siegel & Bugg, 2016 0.01 [-0.38, 0.40] Siegel & Bugg, 2016 -0.05 [-0.44, 0.34] Siegel & Bugg, 2016 0.02 [-0.37, 0.41] Siegel & Bugg, 2016 0.03 [-0.36, 0.42] Siegel & Bugg, 2016 0.05 [-0.34, 0.44] Zabelina, Condon, & Beeman, 2014 0.54 [ 0.15, 0.93] Zabelina, Condon, & Beeman, 2014 0.72 [ 0.28, 1.16] Schuldberg, 1990 0.35 [ 0.15, 0.55] Nadasi, 1997 0.18 [-0.09, 0.45] Nadasi, 1997 0.28 [ 0.01, 0.56] Stumm, Chung, & Furnham, 2010 0.14 [-0.15, 0.43] Stumm, Chung, & Furnham, 2010 0.10 [-0.19, 0.39]

RE Model 0.18 [ 0.08, 0.27]

-1 -0.5 0 0.5 1 1.5 Observed Outcome BIPOLAR DISORDER AND CREATIVE POTENTIAL

Figure B1. Forest plot for subset of studies using a correlational design.

Alici, Ozg..ven, Kale, Yenihayat, & Baskak, 2018 -0.97 [-1.56, -0.38] Johnson, Tharp, & Holmes, 2015 0.26 [-0.13, 0.65] Johnson, Tharp, & Holmes, 2015 0.35 [-0.04, 0.74] Rybakowski & Klonowska, 2011 -0.07 [-0.51, 0.37] Rybakowski & Klonowska, 2011 0.41 [-0.03, 0.85] Rybakowski & Klonowska, 2011 0.30 [-0.14, 0.74] Rybakowski & Klonowska, 2011 0.36 [-0.08, 0.80] Rybakowski & Klonowska, 2011 0.33 [-0.11, 0.77] Rybakowski & Klonowska, 2011 0.45 [ 0.01, 0.89] Rybakowski & Klonowska, 2011 0.62 [ 0.18, 1.06] Rybakowski & Klonowska, 2011 0.69 [ 0.25, 1.13] Rybakowski & Klonowska, 2011 -0.22 [-0.66, 0.22] Rybakowski & Klonowska, 2011 -0.77 [-1.21, -0.33] Rybakowski & Klonowska, 2011 -0.41 [-0.85, 0.03] Rybakowski & Klonowska, 2011 -0.64 [-1.08, -0.20] Rybakowski & Klonowska, 2011 0.07 [-0.37, 0.51] Rybakowski & Klonowska, 2011 0.14 [-0.30, 0.58] Rybakowski & Klonowska, 2011 0.38 [-0.06, 0.82] Rybakowski & Klonowska, 2011 0.22 [-0.22, 0.66] Santosa et al., 2006 -0.28 [-0.62, 0.06] Santosa et al., 2006 0.08 [-0.26, 0.42] Srivastava et al., 2010 -0.47 [-0.86, -0.08] Srivastava et al., 2010 0.09 [-0.25, 0.43]

RE Model 0.05 [-0.13, 0.22]

-2 -1.5 -1 -0.5 0 0.5 1 1.5 Observed Outcome

Figure B2. Forest plot for subset of studies using mean differences. BIPOLAR DISORDER AND CREATIVE POTENTIAL

Table B2

Results of moderator analysis of the categorial moderators. level variable k d 99% CI p LL UL study country England 3 0.21 -0.13 0.55 .23 Turkey 1 -0.97 -1.75 -0.19 .02* Poland 16 0.12 -0.05 0.28 .28 USA 22 0.12 -0.02 0.25 .25 study type mean difference 23 0.05 -0.09 0.19 .50 correlational 19 0.17 0.02 0.32 .03* sample age group adults 25 0.06 -0.08 0.20 0.42 young adults 15 0.16 -0.01 0.33 0.07 school children 1 0.41 -0.24 1.06 0.22 mixed 1 0.08 -0.63 0.77 0.82 matched CG yes 17 0.06 -0.15 0.27 .56 no 8 0.04 -0.25 0.33 .78 DT DT test AUT 14 0.13 -0.05 0.31 .15 TTCT 12 0.06 -0.13 0.26 .54 other 16 0.12 -0.06 0.29 .20 DT scoring fluency 16 0.09 -0.08 0.27 .30 flexibility 1 -0.05 -0.75 0.65 .89 originality 11 0.05 -0.16 0.26 .63 elaboration 1 0.03 -0.67 0.73 .93 total 13 0.18 -0.01 0.37 .07

DT domain verbal 20 0.14 -0.00 0.28 .06 figural 6 -0.12 -0.39 0.15 .39 numeric 4 -0.07 -0.40 0.27 .71 mixed 12 0.22 0.03 0.41 .03* DT instruction quantity 26 0.12 -0.02 0.25 .09 quality 16 0.09 -0.07 0.25 .29 DT calculation summed 32 0.09 -0.03 0.21 .15 averaged 7 0.05 -0.21 0.30 .72 other 2 0.37 -0.10 0.84 .12 BD BD presentation bipolar I 5 0.05 -0.26 0.36 .74 bipolar mixed 21 0.06 -0.09 0.21 .44 BIPOLAR DISORDER AND CREATIVE POTENTIAL

subclinical 16 0.17 0.01 0.34 .04* BD status healthy 17 0.17 0.04 0.30 .009** euthymic 17 0.14 0.00 0.28 .04* manic 4 0.25 -0.05 0.55 .10 depressed 4 -0.51 -0.81 -0.21 .001*** BD assessment DSM 23 0.05 -0.10 0.19 .51 HPS 12 0.21 0.02 0.39 .03* other 7 0.09 -0.16 0.34 .47 clinical diagnosis yes 23 0.08 -0.06 0.22 .27 no 16 0.17 0.01 0.33 .04* partly 3 -0.10 -0.48 0.29 .61 comorbidity yes 7 0.12 -0.18 0.42 .43 no 18 0.06 -0.13 0.26 .51 medication all 16 0.12 -0.10 0.33 .28 mixed 9 -0.05 -0.32 0.23 .74 Note. Small (d = 0.20), medium (d = 0.50), and large (d = 0.80) effect sizes regarding conventions for Cohen’s d (Cohen, 1988) are written in bold numbers.

* p < .05, ** p < .01, *** p < 0.001

Table B3

Results of moderator analysis of the continuous moderators.

95% KI LL UL p level variable k b0 b b0 b b0 b b0 b study study age 42 0.03 0.01 -0.16 -0.01 0.21 0.03 .78 .31 sample mean age 35 0.62 -0.02 0.03 -0.03 1.22 0.00 .04* .09 gender 40 0.63 -0.01 -0.54 -0.03 1.80 0.01 .29 .38 DT time 33 0.17 -0.02 -0.23 -0.15 0.56 0.11 .41 .75 Note. Small (d = 0.20), medium (d = 0.50), and large (d = 0.80) effect sizes (Cohen, 1988) are written in bold numbers.

* p < .05 BIPOLAR DISORDER AND CREATIVE POTENTIAL

Table B4

association

variable 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 1. age of study (c.) - 2. country (p.) .23 - 3. study type (d.) .05 .76*** - 4. mean age (c.) -.41* .79*** .87*** - 5. age group (p.) .15 .51** .91*** .97*** - 6. gender (c.) .12 .44* .12 -.35* .49* - 7. DT test (p.) .26 .76*** .72*** .69*** .50** .19 - 8. DT scoring (p.) .27 .43* .56*** .42* .34 .45 .68*** - 9. DT domain (p.) .35 .34 .52*** .41 .32 .18 .62*** .48** - 10. DT time (c.) .21 .44* .64*** -.70*** .76*** .73*** .57** .58* .38 - 11. DT instruction (d.) .30 .65*** .42** .44** .60** .03 .76*** .51* .33 .29 - 12. DT calculation (p.) .23 .45** .48** .70*** .44* .82*** .41** .47 .28 .79*** .30 - 13. BD presentation (p.) .33 .64*** .89*** .93*** .71*** .34 .65*** .52** .46** .74*** .52** .43** - 14. BD status (p.) .14 .47*** .91*** .96*** .58*** .25 .54*** .35 .39* .75*** .58** .39 .71*** - 15. BD assessment (p.) .08 .58*** 1.0*** .92*** .70*** .34 .57*** .52** .44* .69*** .59*** .45** .68*** .66*** - 16. clinical diagnosis (p.) .15 .55*** .87*** .96*** .81*** .31 .57*** .44* .35 .76*** .68*** .41** .72*** .69*** .65*** - 17. comorbidity (d.) .64*** NaN .59** .64*** NaN .19 .91*** NaN .65* .46* .00 NaN .92*** .55 NaN .60* - 18. medication (d.) .23 NaN .24 .83*** NaN .04 1.0*** NaN .57* .05 .47* NaN NaN NaN NaN NaN .73*** -

Notes. (d.) = dichotomous, (p.) = polytomous, (c.) = continuous. Pearson’s correlations were used between continuous moderator; pointbiseral correlations were used between continuous and dichotomous moderators;  were used between continuous and polytomous moderators; and

Cramer’s V were used between polytomous moderators. p-values for Pearson’s correlations were adjusted for multiple testing. Cells with NaN show relationship for which according statistics could not be computed because there was not enough information obtainable.

* p < .05, ** p < .01, *** p < .001. BIPOLAR DISORDER AND CREATIVE POTENTIAL 0 6 5 0 . 0 r o r r E

2 d 1 r 1 a . d 0 n a t S 8 6 1 . 0 4 2 2 . 0

-0.5 0 0.5

Observed Outcome

Figure B3. Funnel plot for analyzing publication bias for the subset using fluency as scoring of DT. BIPOLAR DISORDER AND CREATIVE POTENTIAL 0 5 7 0 . 0 r o r r E

5 d r 1 . a 0 d n a t S 5 2 2 . 0 3 . 0

-1 -0.5 0 0.5

Observed Outcome

Figure B4. Funnel plot for analyzing publication bias for the subset using originality as scoring of DT. BIPOLAR DISORDER AND CREATIVE POTENTIAL 0 5 7 0 . 0 r o r r E

5 d r 1 . a 0 d n a t S 5 2 2 . 0 3 . 0

-1 -0.5 0 0.5

Observed Outcome

Figure B5. Funnel plot for analyzing publication bias for the subset using mean differences. BIPOLAR DISORDER AND CREATIVE POTENTIAL 0 6 5 0 . 0 r o r r E

2 d 1 r 1 a . d 0 n a t S 8 6 1 . 0 4 2 2 . 0

-0.4 -0.2 0 0.2 0.4 0.6 0.8

Observed Outcome

Figure B6. Funnel plot for analyzing publication bias for the subset using correlations. BIPOLAR DISORDER AND CREATIVE POTENTIAL 0 6 5 0 . 0 r o r r E

2 d 1 r 1 a . d 0 n a t S 8 6 1 . 0 4 2 2 . 0

-0.2 0 0.2 0.4 0.6 0.8

Observed Outcome

Figure B7. Funnel plot for analyzing publication bias for the subset using mean differences after using the trim-fill method. White circles indicate unpublished studies that are missing to obtain funnel symmetry.