Artists: how crazy and how smart are they?

And is being an artist heritable?

I’m on something like a lecture circuit this fall. The bad thing about it is that it leaves me less time for other projects, such as the East Hunter Substack. The good thing is that it gets me to interesting places and I get to look at topics I normally don’t do research on. As I was invited to talk at a school which trains musicians, I got to look into some interesting studies about the psychological characteristics of artists.

One stereotype about artists is that they suffer from mental problems more often than average. I don’t know what stereotype says about artists’ intelligence – I guess one would imagine that while you can’t become a successful artist without a reasonably high, probably above average level of intelligence, you don’t need to be a genius in the traditional sense to make it as an artist. Of course it is always an interesting question if it is nature or nurture which makes us who we are – for example, an artist. Let’s start with this last question. The first study we need to look at is this one:

Creativity is the tendency to generate or recognize ideas, alternatives, or possibilities. Following a study on the genetic contribution to working in a creative profession, based on polygenic score analysis, we report the total heritability of this trait in a large sample of adult twins and their siblings registered with the Netherlands Twin Register. Data from 6755 twins and 1817 siblings were analyzed using genetic structural equation modeling. Working in a creative profession is relatively rare in our sample (2.6% of twins and 3.2% of siblings). Twin correlations (identical 0.68 and fraternal 0.40) commended a model with additive genetic factors (full model estimate 0.56), shared (full model estimate 0.12), and unique environmental factors (full model estimate 0.32). Genetic model fitting resulted in a best-fitting model existing of additive genetic factors and unique environmental factors, resulting in a heritability of 0.70.

I think the abstracts says (almost) everything. “Creative professions” were defined as “having positions in the fields of dance, film, music, theatre, visual arts, or writing”. (This will be important later because other authors give themselves a much wider berth with this definition). Although there are somewhat over 8000 people in this sample, power is constrained by the fact that – unsurprisingly – not that many (236 to be precise) people work in these creative professions. I think it’s also not surprisingly that the most common creative profession is graphical design.

The usual pattern was observed: if you had a monozygotic twin who was a creative, your chance of also being a creative was higher than if you had a dizygotic twin who was a creative. This is what suggests heritability. The authors are on the fence if there is a shared environmental effect, that is, if growing up in certain families per se makes you more likely to be an artist beyond the genetic effect. MZ correlations are not quite twice as high as DZ correlations, but modeling them without a shared environmental effect works fine. Of course we should never forget that we would need thousands, if not tens of thousands of twins to reject a shared environmental effect of 0.1-0.2 for a trait which has 2.5% population prevalence as we see here (Fig. 2 is the nice illustration of this in the linked paper). I should note that of course there are earlier twin studies on similar phenotypes (with very similar results). So yes, there are certainly genetic effects that make you more likely to be an artist, and maybe the family culture also contributes somewhat – but lots of caveats about this latter thing, of course assortative mating can look like a shared environmental effect too, and artists probably do pair with other artists.

Our second question concerns if authors have particular patterns of mental illness and intelligence. For this, it is best to look at two amazing studies of the Swedish population registries, both works of Simon Kyaga. One of them is this:

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

As with many of these population registry studies, frankly the most amazing thing is that this database exists. Using Swedish population registers run by the government, the authors can match data of hundreds of thousands of people regarding 1) their job, 2) if they were ever treated in the medical system for psychiatric illness (in this case, only 3 illnesses: unipolar depression, bipolar disorder and schizophrenia), 3) if they were men, their age 18 IQ from their military conscription IQ test, and 4) similar data from their family members. The questions are:

-          What is the relationship between mental illness and working in a creative profession and mental health? I should note that “creative profession” here also includes scientists, but there are subgroup analyses so we can look at just the traditional definition.

-          Are there notable familiar patterns in this, such as that if your brother is schizophrenic but you aren’t, you are more likely to become a creative?

-          How does all this relate to IQ?

Let’s see what they found! The most important chart is this:

Here we see a odds ratio of working in a creative profession as a function of mental illness, either in the potential artists/mental health patients themselves, or in their families. Top is schizophrenia, middle is bipolar disorder, bottom is unipolar depression. You need to focus on the blue vertical line at 1: if the dots are to the left, it means reduced risk, if they are to the right, it means increased risk, if the bars on the side do not overlap with the vertical line the risk is significant. Let’s start with interpretations in the middle because that is the most straightforward!

For bipolar disorder, all the dots are on the right, meaning that if either you or somebody in your family has bipolar disorder, you are more likely to work as a professional – here the categories are pooled, but soon you will see that this applies to both artists and scientists.

Let’s go to the top: for schizophrenia, most dots are on the right (more so towards the top, for closer family members), but the uppermost isn’t. What this means is that your chances of being a creative is higher if members of your family are schizophrenic, but it isn’t if you personally are. What this means we will soon explore.

Unipolar depression on the bottom is the also quite straightforward. Most dots cluster on the left, including the one for ‘patient’, meaning that despite stereotypes if you or family members suffer from depression, you are less likely to become a creative professional. (But there are some interesting exceptions which I’m not sure how to interpret.)

Now, let’s look at IQs! The authors tabulate these on the Stanine scale used by virtually all of these Swedish military IQ studies. This scale goes from 1 to 9, with the mean at 5.

There are two things to note here. First, all kinds of creative professionals (look at the column Patient in the top half) have IQs higher than not only the population average (5), but also their own family members. The gap is – unsurprisingly – the biggest for scientists, which is obviously the most cognitively demanding job, but artists are also smarter than average. Family members are also smarter than the population average for all professions, and the closer they are genetically to the artists/scientists, the truer this is, in line with the genetic basis of intelligence.

Second, the opposite is true or sort of true for psychiatric patients (bottom half). Patients all have lower IQs than the population average and their own family members, but the family members also have middling or slightly below average scores, and for the worst illness – both in terms of just how severe it objectively is and how it impacts IQ - , schizophrenia, there is even a gradient, with more distant family members having higher IQs. So, all in all, being a creative is correlated with higher IQ, but being a psychiatric patient is correlated with having a lower IQ!

There is one more analysis we need to look at, contained in this table:

The first nice thing about this table is that we see the psychiatric illness to creative profession risk spelled out by subcategories of creative jobs (bottom two thirds), at least for two people: the patients themselves and their siblings. Looking at the crude estimates, we see that, interestingly, the trends replicate for both scientists and actual artists. If you or your sibling is bipolar, you are more likely to work in these professions, if you or your sibling is depressed, this becomes less likely, while a mixed pattern is seen for schizophrenia: illness in the family increases the chance that you become a creative, but your own illness reduces it. Negative effects are stronger for scientific occupations, suggesting that while it is sort of possible to be a depressed or schizophrenic artist, it is really hard to do science with these illnesses burdening you.

The second interesting thing happens in the second column, after IQ adjustment. Basically, what is the chance of being a creative professional among people who have a psychiatric illness but do not have the IQ disadvantage these illnesses usually convey? The risks all flip positive (with one exception, and estimates are not always significant). Without the IQ hit, psychiatric illness predisposes you to creative profession!

How do you interpret this? The authors try, but they are too careful. I also try my own speculations. I think the predisposition for mental illness, especially schizophrenia and bipolar disorder, does also predispose people to creative professions, maybe because of the unusual thinking patterns that come with it and maybe (especially for artists), the strong emotions they produce, which can be inspire you to create, maybe especially so if they are negative. The familiar gradient suggests that these overlapping predispositions are genetic predispositions, in line with Cheverud’s conjecture. However, you need above-average IQ to be a successful creative. This masks the predisposing effect because mental illness runs in families of below-average intellect, and at least schizophrenia also takes a direct hit on your IQ, so while schizophrenia in the family is a predisposing factor for creative professions, your own schizophrenia normally isn’t. By controlling for IQ we take the blinders off and see what’s truly going on: among people with similar intellect, having mental illness is a predisposing factor for creative professions, even if you yourself are ill.

The authors did a follow-up study soon after, with now over 1 million people and more diagnoses. This is the one:

Mental illness, suicide and creativity: 40-Year prospective total population study

The methods are similar, the main results are depicted in the looooong figure:

For most of the newly included diagnoses, the usual pattern was present, in that having these disorders in the family is more predisposing towards creative professions than having it yourself. Alcohol and drug abuse made creative professions less likely across the board – don’t forget, this is not causal drinking and weed smoking, but being hospitalized for alcohol- or drug-related problems! – and anorexia nervosa made creative professions more likely, in line with observations that this illness is the most related to high family SES and high IQ. An interesting observation in this study was that authors are not like other artists. They really do suffer from all kinds of mental illnesses!

The previous findings – high IQ in creatives & families, low IQ in psychiatric patients & families – was replicated (supplement, Table S4). They interestingly also saw low IQ in anorexics! But these were of course anorexic men, who may not be very typical patients. The authors say that “differences in IQ (available for men only) does not account for any of the associations”, but I’m not sure if this is technically true. Yes, unlike for the disorders in the previous study, lots of disorders in the family or in you – especially substance abuse – still make it less likely that you become a creative professional even if you control for IQ, but this control shifts the risk towards positive for almost all disorders, suggesting that low IQ in the mentally ill and their families is indeed an important reason why these people become creative professionals less likely than we would expect.

TL;DR: creative professions run in the family for genetic reasons and artists are more bipolar than average but not mentally ill overall. However, their families – but still not for all disorders! – usually are.