The Rocky Horror Pixel Show

Studying creative thinking in the lab, under tightly controlled conditions, isn't the easiest way to make a living as a psychologist. Even for the wilderness of human thinking, creative ideas seem to be deliberately designed to defy empirical inquiry. They come and go as they please – visits from the Muse, our light bulb moments. Most brain scientists would rather try to nail jelly to the wall. But can this ever excuse ill-conceived efforts to identify the mechanisms of creativity in the brain? Are they anything more than modern-day phrenology?

Suppose an advanced alien lifeform visits Earth to investigate whether Homo sapiens is worth saving. They don't have a portable consciousness-detector, a small antenna-held gizmo they can conveniently hold to our heads to check for signs of inner musings. What would they identify as the defining characteristic of being human?

Taking a quick look around and seeing what we have done with the place, they'd be hard pressed to put any other item on top of their list than our creativity and inventiveness. We are an intensely creative species, and there isn't an element of the periodic table we haven't tinkered with to utterly transform the world we live in. All progress in the arts, sciences and engineering originates from the capacity to change existing thinking patterns, break with the present, and create something new. Creativity, and its derivative products – the knowledge and artefacts that make up human culture – is the quintessence of our humanity. Finding its cognitive and neural mechanisms couldn't be any more central to our enterprise.

There is also a very practical goal here. By uncovering the nuts and bolts of how a three-pound pile of electrified biochemistry conceives of mathematical theorems, invents kitesurfing, creates beautiful art, discovers the laws of nature, thinks of space rockets, and designs buildings that look like sea shells, we might be able to enhance this process. It would be an instant game-changer for any nation or company that gets a handle on how to enhance the invention machine. From Google to London's Royal College of Arts, from the world's medical associations to the military, they would be all over any paradigm edging us closer to that.

Given all this, one would think that the neurocognitive mechanisms of creativity are the subject of intense research efforts in the behavioural and brain sciences, with dozens of labs and big bucks behind them. But this is not the case. Only a handful of labs tackle this vexed problem empirically; funding is practically nonexistent. In fact, while there has been a veritable explosion of knowledge about the mind over the past 70 years or so, creativity has become the most striking exception to this success story. It's hard to think of a mental phenomenon so central to the human condition that we understand so little. We don't have a single sound mechanism to explain the extraordinary creative capacities of an Einstein or a Shakespeare.

Come over here! And be creative! Now!

So, what are cognitive neuroscientists to do if they want to catch creative thoughts in flagrante? They can't simply take volunteers, shove them into the nearest brain scanner and tell them: please be creative! That's a given. But wait. That's exactly what they do. Allow me to introduce the two paradigms that have probably generated well over 90 per cent of anything you might have ever read about the brain mechanisms of creativity.

The first is divergent thinking, proposed by Joy Paul Guilford in 1950. The concept is defined as the ability to generate multiple solutions to open-ended questions, and it is the basis for several standardised testing methods for 'creativity'. The one most often used in neuroscience studies is the infamous Alternative Uses Test, or AUT for short. It asks participants to generate alternative uses for common objects such as a brick or a car tyre.

Try it. Take one minute to write down all the alternative uses for a brick. Then score the test on three factors. First is ideational fluency, which is simply the total number of ideas you had. If you came up with eight items your score is 8. Second is flexibility, which is the number of different types or categories of ideas. If all you could think of were uses as a weapon (hit an enemy, throw at a blocked door, etc.) your score would be a meager 1. Third is originality. This is often assessed with the so-called consensual method, in which naive judges rate unusual answers with 1 point, unique answers with 2 points, and so on. Suppose you had one unusual and one unique item, your originality score would amount to 3. Sum up all points (8 + 1 + 3) and your overall creativity score would stand at 12.

Do you think this captures your creativity? Do you think that a score of 12 as compared to, say, 9 would tell you anything about a person's creative abilities? How do you think bona-fide, creative giants like Ludwig van Beethoven or Marie Curie would have scored on the AUT? Whenever I give the AUT to my students in class and ask them the same questions, they laugh. I have given the AUT to groups of artists and witnessed their anger upon realising that scientists reduce their most prized possession to this.

To be fair, the AUT is actually part of a larger test called the Torrance Test of Creative Thinking (TTCT: Torrance, 1974), which does have some decent, real-world predictive properties. But as the cognitive psychologist Mark Runco (2004), one of the 'custodians' of the TTCT, points out, the full TTCT is an hours-long test and, most importantly, not a test of 'creativity' but a test of divergent thinking. Unfortunately for us, and for progress, neuroscientists have ignored such disclaimers. In neuroscience studies, the mini AUT is used (for scanning purposes), and the results are routinely proclaimed as discoveries about creativity. And as soon as the media gets involved, all inhibition is lost, and the overselling of the findings deteriorates into outright pseudoscience.

The second neuroimaging paradigm uses music – jazz improv mostly. Here a melodic string is completed in one of two conditions, a set pattern from memory (control condition), and an improv string ('creativity' condition), with variables like length, or cadence all held constant. In case you have doubts about the ecological validity of this paradigm, especially in a sterile brain-scanning lab, allow me to tell you a little anecdote. One of the participants – a famous jazz musician – in perhaps the single most famous experiment of this type (Limb & Braun, 2008) took part once in a panel discussion at a conference I was at. He was asked if he considered what he did in the improv condition to be creative. Without a whiff of hesitation, he said no. It was an emperor-has-no-clothes moment.

Weird mental test X gives pretty brain image Y

Neuroimaging creativity sounded like a good idea 20 years ago; a low-hanging fruit if ever there was one. There were ready-to-go 'creativity tests', and the university's brand-new neuroimaging centre was just next door. How can you fail? Some brain region was indeed firing away with extra oomph, and as for the topic itself: does it get any sexier? Alas, a bit of level-headed thinking would have saved a lot of grant money and us from this pixelated Potemkin village.

It's true that the great neuro show of recent decades has revealed a good deal about the human brain and how it functions. Yet at the same time, such a frenzied and hyped atmosphere invites excesses, and creativity research is perhaps the best example. Even if we bracket test validity and the artificial lab conditions (see Abraham, 2018), there are two additional problems that render divergent thinking theoretically incoherent for neuroscience.

First, divergent thinking is a false category formation (Dietrich & Kanso, 2010). It only takes a moment's reflection to see that we can also be creative with convergent thinking, a fact everyone in the field acknowledges. For all the uplifting stories, the Einsteins riding on beams of light, the Newtons watching falling apples (a myth likely originating from Voltaire) or Archimedes displacing bathwater, creative ideas can just as easily be the result of laborious trial and error. What would we otherwise make of Edison's 'empirical dragnet' method that yielded a total of 1093 patents; Watson and Crick's algorithmic approach to testing the stability of DNA base pairs; Bach's assembly-line tactic to composing hundreds of cantatas; or the imaginative ways in which NASA engineers solved the problems of the otherwise doomed Apollo 13 mission?

Since the exact opposite – convergent thinking – can also produce creative ideas, the obvious question arises of what, exactly, is creative about divergent thinking? No one has been willing, or able, to explain this to me. If both divergent and convergent thinking can lead to both creative and non-creative thinking, the concept of divergent thinking as a proxy for creative thinking makes no sense.

The false category problem also applies to all other conceptions of creativity that have been tried over the years, such as defocused attention, remote associations, flow, madness, lateral thinking, low arousal, daydreaming, REM sleep, right brains, mindfulness, unconscious thinking, prefrontal cortex, or the default mode network. Given their opposites also lead to creativity, they all fail to carve nature at the right joints.

Second, divergent thinking is, like creativity itself, a compound construct consisting of many different separate and distributed mental processes with no one having the slightest clue what they are and in what mix. Although this problem is also widely acknowledged, there is no effort under way to break divergent thinking down further so as to link it to the kinds of processes we use to operationalise all other psychological phenomena, such as working memory, cognitive control, perceptual processes, semantic memory, inhibition or executive attention. Remember that neuroscientists hunt for underlying mechanisms; that is, individual mental processes that can actually be identified by neuroimaging. Divergent thinking, therefore, does nothing to address the initial problem, the fallacy of taking creativity as a monolithic entity. In fact, the concept is used today the way it was initially conceived by J.P. Guilford in 1950!

For neuroimaging, all three confounds – validity, false category formation, compound construct – combine to make defeat certain. Simply put, if you fail to isolate the subject matter of interest in your study, you cannot use neuroimaging to search for mechanisms. You just don't know what the brain image shows!

The evolution of phrenology

Phrenology is to neuroscience as astrology is to psychology, the quintessential pseudoscience of the discipline. Mention the P-word to a group of neuroscientists and you won't be able to finish your sentence, so eager would they be to interrupt you and put ideological distance between themselves and phrenology. Trouble is, however, the basic error that fuelled phrenology some 200 years ago is nearly impossible to shake.

Franz Joseph Gall, the father of phrenology, associated a total of 27 regions with specific mental faculties, all without so much as a shred of evidence. The list contained such doozies as a centre for mirthfulness, combativeness, marvellousness, secretiveness and, my personal favourite, the organ of philoprogenitiveness – which he located, if you must know, just above the middle of the cerebellum. It's easy to laugh at this bumps-on-the-skull idea today, but few people appreciate the enormous popularity phrenology enjoyed at the time. In Victorian Britain, it ranked with Darwin's theory of evolution.

For creativity, the best-known example is, of course, the right-brain theory. It emerged in the 1970s and has proven to be a particularly vicious mutation of phrenology, probably because it was the first to metastasise to a global audience. You can still find a business seminar on how to think with both sides of your brain and an endless supply of books and magazines promising an easy step-by-step programme on how to tap into your creative right-brain potential. You might as well ask someone to make better use of the thalamus.

The unlikely story of the right brain developed from split-brain patients who had their corpus callosum cut to manage life-threatening epilepsy. Subsequent research showed that there are indeed several cognitive functions that are lateralised – language most prominently. The generalisation commonly extracted from this is that the left hemisphere (LH) is more skilled at analytical tasks, such as sequential reasoning, while the right hemisphere (RH) is more skilled at tasks requiring synthesis, such as seeing the whole of a picture (see Gazzaniga et al., 2014).

Anyone with two neurons to rub against one another can readily imagine what happened next. Generalisations that feed into our phrenological thinking habits inevitably set the stage for a broadside of flag-waving oversimplifications. And creativity was such an easy target, given our overly romantic view of it. Isn't regular old thinking – conscious, analytical and systematic – perfectly suited, in other words, to the logicality of LH? And isn't creative thinking – intuitive, primal, holistic and delightfully irrational – custom-made, as it were, to the mysterious ways of RH? Sadly, LH has since been the go-to brain half of methodical and unimaginative engineers, while RH has become the creative playground of inspired artists. Today, everyone recognises the right-brain theory as a dud, of course, but its persistence in the general public underscores the need, in clear and vivid form, of how important it is to systematically demolish ideas gone bad.

Phrenological thinking is so pervasive and the opportunities for flummoxing so abundant that the field is full of such misbegotten proposals. The next candidate in this wild-goose chase was the prefrontal cortex, which owed its brief moment in the limelight to its general position at the apex of human mental faculties. It didn't stay there long. It was quickly followed, in about the mid-2000s, by an idea that was also just too irresistible for the myopic theorising that has come to characterise this field – the default mode network (DMN). Proposed by Marcus Raichle and colleagues, this network of brain areas shows heightened activity at 'rest' and is thought to support mindwandering, daydreaming or moments of introspective thought. As was the case for the right-brain theory, this struck all the right chords in some quarters. Researchers jumped on it and, for about 10 years or so the DMN was the bull to ride.

Sharp-eyed readers will have recognised this as simply another false category formation. There is of course no reason to think that the other large-scale brain network – the central-executive network (CEN), which controls executive functions and shows heightened activity when we focus our attention on a specific task – couldn't also generate creative thoughts. As this became clear, the phrenological explanation evolved yet again. The latest twist is that creativity is purported to be associated with enhanced connectivity and a back-and-forth interplay between these networks (see e.g. Beaty et al., 2016).

The all-important thing to remember here is this. All of these claims are extracted from – wait for it – the one-minute AUT! What's more, the findings are not packaged and sold as studies of divergent thinking. Aside from the occasional qualifying remark (as if this turns the water into wine), the papers routinely make grand proclamations about the brain mechanisms of creativity!

Clearly, we need a sanity check. Creativity is a topic where respectable people, even those of the highest scholarly standing, regularly rise to levels of speculation that can safely be called imprudent. Creative people are supposed to use more of their brain (somehow, for no one can tell you exactly how, let alone link this to creativity); use less brain more efficiently (which is, come to think of it, the opposite claim); have more dopamine receptors (or was it serotonin?); have more densely packed neurons, or more synaptic connections, or a thicker corpus callosum… Indeed, one can find claims in the literature for the whole funhouse of brain structures in the telencephalon – hippocampus, visual cortex, temporal lobe and, why not, the basal ganglia. The next thing in tow, given the drift of things, is surely the mirror neuron system or neurogenesis in the hippocampus. And let's also not forget the legions of creativity coaches, leadership consultants and motivational speakers who circle the globe in a tireless effort to meme-launder all of this and emit it to audiences suffering from uncritical idolatry.

This shouldn't surprise us, I guess, given the time-tested ability of pseudoscience to expand in a vacuum. Will these phrenological thinking habits about creativity go extinct? I expect not. It's too seductive.

Now what?

Clearly, we need a restart, preferably one with ideas that are theoretically and conceptually sound. Here are five suggestions.

First is the Vaudeville conception (Dietrich, 2015). This view stops the deadly error of thinking about creativity as a monolithic entity and takes seriously the idea that creativity is a highly complex and multidimensional phenomenon. We might think of creativity as a single and cohesive entity in psychological terms, a personality trait – as in, Steve Jobs had it (notice the singular) and my grandfather didn't – but creativity, as such, might not exist as a distinct and separate entity at the neural level. Think philoprogenitiveness! But the whole rationale of neuroimaging studies rests on the assumption that creativity, or divergent thinking, is a discrete thing in the brain and that that thing is detectable by neuroimaging tools.

The underlying tacit assumption here is that there is such a thing as 'normal' thinking to which a separate and extra something – the creative bit – is specifically added to make the sparkling difference. Few people would probably subscribe to this position once it is laid bare, but without it neuroimaging makes no sense. And it wouldn't matter if the 'creative bit' is a place or a network, the tacit assumption is the same. By way of comparison, it's easy to point out England on a map, but you can't find all people in the world who speak English that way. Consider how scientists, entrepreneurs, designers, or ballet dancers can be creative in their respective domains. Creativity, then, is fully embedded and distributed in the brain, or in a word, everywhere. Asking neuroscientists for the location of creative thinking is like asking them for the location of thinking.

Second are types of creativity. This hits the same monolithic-entity spot but from a different angle. Given the highly complex and multifaceted nature of creativity, an obvious way to make it more tractable in the brain is to parse it into different types.

I have proposed to initially divide creativity into three distinct types, a deliberate mode, a spontaneous mode and a flow mode (Dietrich, 2015). To avoid the pitfalls of previous such attempts – false category formation and compound construct – the three types are explicitly defined and delineated from one another based on established concepts in cognitive psychology and neuroscience. They are thus valid subtypes in the sense that they can be theoretically defended. Only by biting off smaller pieces of the larger pie, will we have a realistic line of attack that will eventually lead to meaningful data about how creativity happens in the brain. And it might also finally stop the bad habit of making grandiose statements about creativity per se. Since different types of creativity contain opposing cognitive and neural mechanisms – focused vs. defocused attention or DMN vs. implicit system, for instance – any global claim about creativity as a whole will almost certainly qualify as phrenology.

Third is evolutionary theory. Broadly speaking, when we think creatively and break new ground we can be said to explore an unknown problem space. We try out several different options or solutions in this space (variation) and pursue one but not the others (selection). Mathematically, this can be described as a set of evolutionary generate-and-test algorithms. Despite the broad agreement that the basic grammar and logic of evolutionary thinking applies to human creativity, the two-step evolutionary rationale has been nearly universally ignored in setting up empirical protocols. Given that both, variation and selection, likely engage different cognitive processes and brain regions, it's hard to imagine useful neuroimaging data from psychometric instruments that blend them.

Fourth is the brain's prediction system. A key difference from the Darwinian algorithms that do all the creating and designing in nature is that the brain's evolutionary algorithms operate with degrees of sightedness (Kronfeldner, 2010). That is, our exploratory walks through unknown problem spaces are not blind but (partially) informed. But how can we have any degree of sightedness of a problem space that is unknown? Unless we are prepared to consider creators a special class of prophets, we need a sound mechanism to explain what otherwise looks like clairvoyance.

I have recently proposed that predictive representations might be the neural mechanism for the partial sightedness in human evolutionary algorithms (Dietrich, 2015). Theorists have been converging from quite different quarters on the idea of prediction as a central purpose of brain function. It's a new and powerful paradigm in all of neuroscience. For creativity research, the brain's predictive computations represent a proposal for one specific component of the creative process, something that is much more readily subjected to empirical testing than the whole of creativity.

Fifth is the brain's dual architecture. It's well known that the brain has two anatomically and functionally distinct information-processing systems, one implicit and one explicit. And both can be creative. Needless to say, creative output for either the implicit or the explicit system would involve different mental processes and brain areas. For the neuroscientific study of creativity, this knowledge must become part of looking for possible mechanisms.

- Arne Dietrich is Professor and Chair in the Department of Psychology at the American University of Beirut, Lebanon

[email protected]

Key sources

Abraham, A. (2018). The neuroscience of creativity. Cambridge: Cambridge University Press.
Beaty, R.E., Benedek, M., Silvia, P.J. & Schacter, D.L. (2016). Creative cognition and brain network dynamics. Trends in Cognitive Science, 20, 87–95. doi:10.1371/journal.pone.0001679
Dietrich, A. (2015). How Creativity Happens in the Brain. London: Palgrave Macmillan.
Dietrich, A. & Kanso, R. (2010). A review of EEG, ERP and neuroimaging studies of creativity and insight. Psychological Bulletin, 136, 822–848.
Gazzaniga, S.M., Ivry, R.B. & Mangun, G.R. (2014). Cognitive neuroscience (4th edn). New York: W.W. Norton.
Guilford, J.P. (1950). Creativity. American Psychologist, 5, 444–454.
Kronfeldner, M.E. (2010). Darwinian 'blind' hypothesis formation revisited. Synthese, 175, 193–218.
Limb, C.J. & Braun, A.R. (2008). Neural substrates of spontaneous musical performance: An fMRI study of jazz improvisation. PLoS ONE 3(2): e1679. Raichle, M.E., MacLeod, A.M., Snyder, A.Z. et al. (2001). A default mode of brain function. Proceedings of the National Academy of Sciences, 98, 676–682.
Runco, M. (2004). Creativity. Annual Review of Psychololgy, 55, 657–687.
Torrance, E.P. (1974). Torrance Test of Creative Thinking. Lexington, MA: Personal Press.