Researchers are finding new links between specific genes and intelligence. Can we use this knowledge to make people smarter?

I n the opening moments of the 2011 thriller Limitless, Bradley Cooper stands on the edge of a high-rise balcony, lamenting his “four-digit IQ,” preparing to plunge to his doom.

Cooper is playing Eddie, a layabout writer transformed by brain-boosting pills into a Wall Street wunderkind. At the start of the movie, Eddie’s working on a novel, but he seems blissfully ignorant of the literary tradition that will ultimately leave him teetering on the side of a skyscraper: Taking shortcuts to get smarter never ends well.

There’s Flowers for Algernon, in which poor Charlie Gordon undergoes experimental surgery to raise his IQ and then watches all that he gains fade away. In Understand, a 1991 novelette by Ted Chiang, a near-drowning victim is made super-intelligent by an experimental drug treatment, but his desire to know all is eventually his downfall. The movie Phenomenon gives us a man, played by John Travolta, whose sudden brilliance turns fatal.

Hearing these stories, Icarus, Pandora, and Prometheus would nod in recognition. To tinker with intelligence is the height of hubris — and it always comes with a steep price.

And yet we’re going there. New technologies are now making it possible to look for each and every gene that influences whether a person will become smart or dull. This spring, researchers released the largest study of this kind on intelligence and identified 40 genes that had never before been linked to IQ. Another study, published this month, uncovered three genetic variants associated with extremely high IQ scores.

To some scientists, these new studies are promising steps toward a true understanding of the biological basis of intelligence. If indeed that is possible — and the genetics underlying intelligence are not too complex to ever unravel — it’s likely that people will use that knowledge to try to enhance human minds, whether by engineering brighter babies or devising ways to biologically boost our brains. But — and here’s the literary comeuppance, the inevitable meeting of Bradley Cooper and balcony edge — what if upgrading intelligence is too fraught a project, riddled with too many undesirable side effects, to ever come to fruition?

People have been interested in improving intelligence since before there were reliable ways to measure the concept. The coiner of the term “eugenics,” Francis Galton, got the idea of breeding “better” humans after reading his half-cousin Charles Darwin’s 1859 book On the Origin of Species. He was also the author of the first attempt to study exceptional intelligence: an 1869 tome called Hereditary Genius.

“[A]s it is easy … to obtain by careful selection a permanent breed of dogs or horses gifted with peculiar powers of running, or of doing anything else, so it would be quite practicable to produce a highly gifted race of men by judicious marriages during several consecutive generations,” Galton wrote.

Throughout the early 20th century, eugenics would become mainstream thinking, leading to abuses like forcible sterilization and the entire Nazi movement. Modern behavioral geneticists are thus quick to point out the differences between their interest in genes and that of people like Galton. Modern geneticists realize, for example, that genes aren’t destiny, and that the environment and public policy play a huge role in determining individual fates, says Daniel Benjamin, a University of Southern California professor who studies “genoeconomics,” which uses genetic data to explore economic behavior.

“It’s really important when doing this work on behavior and genetics to always be aware and have in the front of our minds the really negative history of misuse and misinterpretation of research in this area,” Benjamin says.

But there is evidence that when selection for intelligence is poised as a matter of choice or prevention, the public becomes rather accepting. In surveys of the clients of in-vitro fertilization clinics in the United States, about half say they would pay for a genetic screening that would reveal whether an embryo would likely become a person with well-below-average intelligence, says Stephen Hsu, a physicist at Michigan State University who develops algorithms for predicting the characteristics that a genotype will yield. “In Asia, it’s like 90 percent,” Hsu says. Such predictive genetic tests don’t yet exist yet, but “it’s really just a matter of time.” When you throw in the rapid advances in the CRISPR gene-editing technique, it’s not hard to imagine things getting very Gattaca, very quickly.

This potential future is coming into view largely because of the plummeting cost of genome sequencing. That has allowed researchers to examine the genetic code of larger and larger groups of people in genome-wide association studies, or GWAS. This method involves gathering huge numbers of people with a certain trait — say, tens of thousands of people who have taken IQ tests and scored within the normal range — and scanning their entire genomes.

The target of these investigations are genetic variants called single nucleotide polymorphisms, or SNPs, that occur in at least 1 percent of the population. SNPs are akin to typos in single base pairs of DNA. They’re a handy tool for looking at the relationship between genes and traits because they appear throughout the genome and are relatively common in the population. The role of rarer mutations, such as the cause of genetic conditions like cystic fibrosis, is well understood, but trying to figure out why someone has an IQ of 110 rather than the average (100) requires a more subtle approach.

Researchers have long known that intelligence is highly heritable, meaning much of the difference between your IQ and someone else’s — about 40 to 80 percent, depending on your age — is explained by genetic differences. But it’s still unclear which particular genes are responsible. If you remember reading headlines over the years about this gene or that gene being linked to genius, forget them. Chances are, those findings failed to replicate.

The reason? Intelligence, like other complex human traits, is caused by tiny contributions by thousands of genes, most of which have yet to be identified. No gene’s contribution is large enough to pop out in a small sample of people.

In large sample sizes, though, subtle effects of individual genes can become more apparent. In May, an international group of intelligence researchers published a massive GWAS paper on 78,308 individuals, larger than any GWAS on intelligence ever conducted before. The sample was a pool of many individuals (all of European descent) who’d previously been enrolled in behavioral genetics studies and had taken intelligence tests. The study revealed 52 genes that correlated with IQ scores, including 40 that had never been linked to smarts.

“For over a century, we have known intelligence is heritable, but we were only able to pinpoint the actual genes involved thanks to GWAS,” says Danielle Posthuma, the lead author on the paper and a statistical geneticist at Vrije Universiteit in Amsterdam.

In another new study, this one published July 4 in the journal Molecular Psychiatry, scientists honed in on the exceptionally brilliant, using GWAS techniques to compare the genomes of 1,238 people whose IQ scores are in the top 0.03 percent of the population and the genomes of 8,172 regular schmoes. The researchers found three SNPs that were significantly more common among the geniuses.

For researchers like Hsu, these results are exhilarating. We’re steps away, he says, from identifying genetic predictors for smarts. Already a study on the genetics of educational attainment (a concept closely correlated with IQ) that surveyed the genomes of 750,000 people found 600 SNPs that correlated with how many years of school someone completes. Something similar on IQ scores, with a million or more study participants, could easily yield the kind of information ambitious parents could use to avoid or alter less-promising genetic profiles, Hsu says.

Seven years ago, when he and research collaborators at the Beijing Genomics Institute proposed looking for genetic variations among people at different intelligence levels, “people laughed,” he says. They said the variations wouldn’t exist. Now, it’s clear that they do — and the technology to do this kind of work isn’t slowing down. “Everybody needs to stop and consider, what is this going to be like five years from now or 10 years from now?” he says. “It’s going to have huge societal impacts and people are not ready for it.”

Not everyone thinks the future will look so different, though.

“I would have stopped 10 years ago myself, and that’s almost to say I wouldn’t have started,” says Wendy Johnson, an intelligence researcher at the University of Edinburgh who has been a critic of genome-based searches for the key to intelligence. “I think we’re wasting a lot of money.”

What makes Johnson so skeptical is the utter complexity of intelligence. Obviously, the environment influences intelligence: A kid deprived of basic care and stimulation is not likely to thrive. But there’s no clear dividing line between the influence of genes and the influence of environment. For example, high-IQ parents tend to take their children to museums, do science experiments with them, and read them a dozen picture books a day. That simulating environment is created, in part, because of the same “smart” genes that mom and dad pass down to their kids. Disentangling the contributions of genes versus environment isn’t easy.

And it gets weirder. The environment might also influence how heritable IQ is. When kids are well off, with their basic needs met and plenty of cognitive stimulation, genes often explain much of the variation between them. Against a backdrop of deprivation, however, genes sometimes recede into the background, at least according to some research. It’s not only that both nature and nurture matter, but that they influence each other in different ways in different people.

Let’s say it turns out that there’s a genetic variant that helps make a person extremely verbal. You inherit this variant, get a graduate degree in English and marry a fellow academic in your department. Your first child gets a nice set of verbal genes, but he also gets all the benefits of a couple of English professors at home. You and your spouse talk all the time, and not just about The Odyssey. When you drive in the car, you chatter to your child about whether you’re turning right or left, what the “W” or “E” on your rearview mirror compass means, about the distance to your destination, and about maps. Maybe those conversations spark his curiosity, so you’ll set him on your lap and tinker around on Google Earth together. Before long, your kid is getting pretty good at spatial reasoning.

Now multiply your imaginary map-loving kid times a few hundred thousand and imagine they all get their genomes sequenced. A geneticist trying to understand intelligence might see that people with those verbal-boosting genes seem to be very good at spatial reasoning. “Aha,” she’ll think, “these genes must be driving general intelligence, the ability to work through multiple types of cognitive tasks.” But she’ll be wrong.

In other words, sweeping genome-wide studies can turn up correlations between genetic variants and a complex trait, but they can’t fill in the blanks that explain why that correlation exists. “It’s very fascinating, but it’s also muddy as all get-out,” says Johnson.

It’s possible that intelligence simply isn’t something understandable on the level of genes. Philosopher Rom Harré has compared the idea of studying intelligence on a genetic basis with researching what a carpet is using only a mass spectrometer. Sure, you might learn a bit about polyester blends, but it’s not going to tell you much about how to pull a room together.

“If a geologist wanted to explain plate tectonics and the way the continents have moved over millions of years, they wouldn’t want to be doing chemical analyses of little rocks that they pick up in their backyard,” says Eric Turkheimer, a psychologist at the University of Virginia. “You have to think about scale when you’re studying things, and not all things that human beings do are very explicable on the scale of neurons and genes.”

This is a jarring opinion in a world where new technologies have allowed us to look at the brain in smaller and smaller chunks. But Turkheimer’s argument recently got a boost from a paper published in the journal Cell. In the article, Stanford University geneticist Jonathan Pritchard and his colleagues argue that complex traits aren’t polygenic, or influenced by multiple genes, as geneticists have long assumed. No, Pritchard argues: They’re omnigenic, or influenced by every gene.

In essence, the omnigenic hypothesis posits that the networks regulating genes are so interconnected that any gene expressed in a given tissue is going to have some impact, no matter how infinitesimal, on the function of that tissue. What’s more, the genes likely aren’t neatly arranged in discrete clusters, as behavioral geneticists have hoped.

Indeed, many of the genetic variants linked to intelligence that have discovered so far are involved in expansive tasks in the very structure of the brain. For example, Posthuma and her colleagues found associations between intelligence and genes involved in the formation of synapses, in the development of different types of neurons, and in guiding the growth of the axons that neurons use to transmit messages. Logically, those processes of brain development would have a relationship to intelligence, says Terrance Deacon, a biological anthropologist at the University of California, Berkeley. But it’s also not exactly groundbreaking to say that being smart has something to do with your brain having, at some point, developed.

Some other genes associated with intelligence appear to be busy all over the body. One discovered by Posthuma and her team had previously been linked to bone formation and hypertension. The SNPs of note that were found in people with ultra-high IQs were located in a gene called ADAM12, which encodes an enzyme that binds to a protein that binds to insulin-like growth factors. Suffice it to say, the protein and its enzyme are found in basically every tissue, doing lots of different stuff.

It gets even messier. Posthuma and her team also found that genetic variants associated with intelligence were also overrepresented in people who are tall, in people who have autism, and in people who had successfully kicked a cigarette habit. They were less frequently found in people with depression, schizophrenia, and Alzheimer’s. Whatever these genes are doing for intelligence, they’re mixed up in a lot of other stuff, too. Start to tinker with them, and who knows what strings you’ll unravel.

And yet, even amid all the noise and confusion, genetics may yet tell us something about intelligence and give us tools to boost it that are a lot less ethically fraught than embryo selection or genome editing.

For many behavioral geneticists, the biggest promise of the latest research is that even if the number of genetic variants is too big to ever think of manipulating, those variants could help explain what intelligence actually is. If, for example, SNPs correlated with intelligence do turn out to cluster in and around genes and regions involved in synapse formation, perhaps we’ll find that the way synapses develop and are maintained explains why some brains are better at reasoning than others. Research hasn’t yet illuminated that black box. Studying the differences between people in synapse structure is extremely difficult because the infinitesimally small spaces between neurons don’t preserve well in dead brain tissue. A few studies on surgically removed brain tissue have found gender differences in synapse structure, but no one has done that kind of work with an eye on human intelligence.

Genome-wide investigations can also help researchers build what are called polygenic scores, which are essentially estimates of how much genetic variance influences a trait (in this case, intelligence). The more genomes that are sequenced, the more accurate the polygenic score. These scores can then feed into research on how the environment and genes interact, says Christopher Chabris, a cognitive psychologist at Union College in New York and Geisinger Health System in Pennsylvania. That kind of research, in turn, can answer questions like, “Who needs early intervention?”

Take the example of dyslexia, says USC’s Daniel Benjamin. Teachers don’t know which kids are going to struggle to read until the kids are old enough to learn to read. If you could catch them earlier with a genetic screening test, you could start an intervention program before the frustration and anguish of dyslexia even set in. “There are lots of examples like that,” Benjamin says.

And there’s the rub: You don’t actually need to fiddle with the genome to biologically boost intelligence. In fact, humans have already been raising their IQ scores, and all without knowing the genetics or the biology of it all. In a phenomenon dubbed the Flynn Effect, intelligence quotients have been steadily creeping upward in much of the world since the invention of intelligence testing about a century ago. A 2014 study found that people have improved their scores by an average of 2.3 standard points per decade. While it’s possible that people have simply become better test-takers, many researchers believe that the Flynn effect really reflects improved intelligence, and that it’s been driven by such unsexy interventions as preventing infectious diseases and improving nutrition. People also have become much taller in the same timeframe, bolstering the idea that nutrition is improving both our bodies and our brains. (In some countries, adult men are, on average, nearly eight inches taller than their counterparts were a century ago.)

Maybe someday we’ll max out on the improvements that can be gained from nutrition, healthy living, and targeted educational programs, and there will be no choice but to get under the hood and adjust the genome if we want to get smarter. In a world where the United Nations estimates that 795 million people don’t get enough food on a daily basis and the National Resource Defense Council figures that 18 million Americans could have been exposed to lead (a known saboteur of cognitive abilities) in their drinking water in the last year, it’s probably safe to say we aren’t there yet.

Maybe we’ll dive into the genome anyway, chasing every last advantage we can wring from the science. Or maybe as we apply our cognitive skills to the painstaking project of understanding those very skills, we’ll find other ways to sharpen minds. Instead of mucking around with base pairs, perhaps we’ll find our toolbox in schools, in homes, in communities. Maybe those sci-fi writers are right, and there is no shortcut to smarts.