Is there anything more mysterious – or human – than the creative impulse? Whatever the field of endeavor: music, art, science, business… What accounts for the inspired burst of innovation? The spark that flits to flame and lights the way to something entirely new?
Roger Beaty became interested in this question as an undergraduate, in a class that explored the psychology of genius. An amateur jazz pianist, he was already well-versed in improvisation. But “this was the first time I realized that you could study creativity scientifically,” he remembers. Combing through case studies of Picasso and the Beatles, following in their gigantic footprints, was a way toward understanding why some people are more creative than others.
Roger Beaty, an amateur jazz pianist, plays a Miles Davis piece on the keyboard in his lab.
There are several cognitive processes involved. Memory is a crucial one, says Beaty, now an assistant professor of psychology at Penn State.
“Memory is what we already know. Creativity involves going beyond what we know – but if we don’t know anything, we can’t create anything new,” said Beaty.
What’s really relevant, he said, is the organization of memory, how a person’s brain catalogs disparate concepts and experiences in order to facilitate making connections. It’s an ability that varies between individuals.
Also important is the ability to focus, to narrow one’s attention to the task at hand. But focus needs to be balanced with spontaneity, Beaty said. “Creative people tend to be open to experience, to seeing things in new ways.”
These processes and others all have their roles to play. But how much does each contribute to an individual’s creativity? Is there something noticeably different going on inside the head of an innovator? How does creativity happen in the brain?
The neuroscience of creativity is an emerging field that has attracted researchers from several disciplines, but it can seem an odd combination. How do you fix a thing as ephemeral as creativity? How pin the butterfly of a new idea to the realm of neurons and physiology?
First, Beaty said, you have to agree on a definition. Researchers in the field generally accept that in order to be considered creative, an idea must be both new and useful.
“Pure novelty is not enough,” he said, “even if ‘usefulness’ in a domain like abstract art is not so clear cut.” Creativity is essentially the solving of a problem, even if it’s a problem that no one knew existed.
To measure creativity in individuals, researchers employ various tests. One requires giving a study participant pairs of randomly-selected words – “shoe” and “door,” say, or “rowboat” and “parrot” – and asking them to rate how closely these words are related to one another.
“People who are more creative are able to see connections between things that might seem unrelated,” Beaty explained.
Another test of “divergent thinking” asks participants to find new uses for common objects, like a sock or a brick. One creative person, Beaty reported, suggested using a sock as a water-filtration system.
Things start to get really interesting when people perform these tasks while researchers observe their brain activity via a functional MRI scan, which provides a real-time image of blood flow to various parts of the brain.
In a study published in the Proceedings of the National Academy of Sciences (PNAS) in 2018, Beaty and colleagues asked 163 people to complete an alternate uses task while in the scanner. Noting the areas that were lighting up in participants’ brains, indicating activity, they computed the linkages between these regions. Thus for each individual they were able to create a map of connectivity that could be related to performance of the task at hand – essentially a map of creative thinking.
This next step was even more revealing. The researchers put the brain-connectivity patterns of the people whose answers on the test were deemed most creative into a computer model, then brought in a fresh set of participants to take the test. Just from comparing a new person’s connectivity patterns with the model, they found, they could predict what that individual’s creativity score would be.
Detail from a figure in Roger Beaty’s PNAS paper shows functional brain networks associated with high-creative thinking ability. The red lines on the brain represent connections predictive of creativity scores and the dots represent brain regions: the larger the dot, the more predictive connections. Regions of the default, executive control, and salience networks are indicated within this larger network by the larger dots. The strength of connections between these three regions predicted people’s ability to think of creative uses for test objects.
Ultimately, Beaty said, the study pinpointed three primary networks in the brain that are involved in creative thinking. The first, called the default network, is the area that activates when a person is relaxing, daydreaming, thinking of nothing in particular.
“It’s the place for spontaneous ideas,” said Beaty. “It’s also strongly related to memory.”
The second network is default’s opposite, the executive control network.”It’s involved in focusing our attention to accomplish challenging tasks,” he said.
“The thing about these two networks is they typically don’t work together,” he added. “If your mind is wandering you don’t need focused attention, and when you’re focusing you don’t want spontaneous thoughts slipping in. It’s kind of an antagonistic relationship.”
For creativity to happen, however, the two have to learn to get along. It’s the interplay between them, in fact, that makes the magic: an iterative process between idea generation and evaluation. That’s where a third player, the salience network, comes in, acting as a kind of toggle between them.
All three of these networks, Beaty said, become active during a creative task. The degree of a person’s creativity depends on the strength of connections between them.
Can creativity be taught?
It’s tempting to conclude that creative people’s brains are simply wired differently. The question then becomes: Is that wiring fixed forever? Might it be changeable? Can a person’s creativity be improved? Once the PNAS study was published, Beaty said, “That was the first thing people wanted to know.”
The popularity of creativity workshops for business leaders and aspiring artists would seem to suggest that creative potential can be developed, or at least unlocked. But can those brain connections actually be strengthened? It’s an open question, and one that Beaty and his colleagues now have NSF funding to try to answer.
Is scientific creativity different from the artistic kind? Are there separate flavors of creativity?
The context for their new study is STEM education. In particular, the researchers will look at whether scientific creativity can be fostered in college students. Their plan is to scan the brains of incoming first-year students, then scan again at intervals after participants have had training in STEM fields, checking to see if connections are strengthened over time. “But first we have to come up with a good test of scientific creativity,” Beatty said.
Which raises another question: Is scientific creativity different from the artistic kind? Is either of these distinct from garden-variety problem-solving? Are there separate flavors of creativity?
“There are some general traits that seem to be shared,” Beaty said. “Flexibility of thinking, the ability to make connections,” including the ability to draw analogies, a subject he investigated as a graduate student. “Then you have to have the domain-specific training, the 10,000 hours people talk about.” Alas, mastery of a subject, while a prerequisite for creativity, is no guarantee.
A test of specifically scientific creativity, he suggests, might include things like hypothesis generation.
“The ability to come up with good research questions, experimental design,” said Beaty. “The sorts of thing I’d like to be better at myself, honestly.”
Once he and his colleagues have a test they are happy with, and if they do see connectivity changes over time, the researchers plan to turn their focus to ways of enhancing scientific creativity in K-12 classrooms.
Humans and machines
It’s an ambitious program, given the challenges built into studying something so elusive.
One problem is the sheer awkwardness of trying to capture the creative process: Lying in a thrumming MRI machine is not very conducive to writing poetry or composing music. To get around this, Beaty and other researchers have rigged up keyboards and drawing pads with non-metal components to make them MRI-compatible. Even with these enhancements, he acknowledges, creativity doesn’t always happen on demand.
An MRI-compatible drawing pad and stylus allow participants to create pictures while their brains are being scanned to reveal regions of activity.
There’s also the difficulty of proving causality. “The connectivity patterns we’re studying are correlated but we don’t know if there is causality,” Beaty explained. Combining fMRI with a technique called transcranial electrical stimulation may help with this: By stimulating the brain with very mild currents through the scalp, he and his team may be able to directly induce increased neuron activity in specific brain regions, then measure the effects on performance.
Yet another challenge is the subjectivity of human testers assigning scores for creativity. “It’s not like determining whether someone got a correct answer on a memory test,” Beaty said. “People vary quite a bit in what they think is creative.”
Training helps to standardize scoring. But Beaty also relies on machine-learning methods like latent semantic analysis, an algorithm that finds word-use patterns in blocks of text and assigns probability to the occurrence of unusual pairings. “It turns out fortunately that the values you get from these text analyses correlate fairly well with human readings,” he said.
In fact, machine learning and artificial intelligence are making rapid inroads in the science of creativity, he added. Some of his colleagues are working on co-creative agents, computer programs that work with human partners to come up with new ideas. Others study generative adversarial networks, or GANs, that pit powerful clusters of computers known as neural networks against one another, mimicking the creative back-and-forth of idea generation and evaluation. Just last year, a pair of GANs produced a painting that so successfully imitated the work of an Old Master it was deemed indistinguishable, and sold at auction for half a million dollars.
Is this, then, the future for this quintessentially human quality? Will we one day be outsourcing the mystery of creativity to machines?
“That’s a question that’s not really settled yet,” Beaty said with the hint of a smile. “But I think there’s more time for us before we completely throw in the towel.”