Scientific American, October 31, 2008


In Robert Plomin’s line of work, patience is essential. Plomin, a behavioral geneticist at the Institute of Psychiatry in London, wants to understand the nature of intelligence. As part of his research, he has been watching thousands of children grow up. Plomin asks the children questions such as “What do water and milk have in common?” and “In what direction does the sun set?” At first he and his colleagues quizzed the children in person or over the telephone. Today many of those children are in their early teens, and they take their tests on the Internet.

In one sense, the research has been a rousing success. The children who take the tests are all twins, and throughout the study identical twins have tended to get scores closer to each other than those of nonidentical twins, who in turn have closer scores than unrelated children. These results–along with similar ones from other studies–make clear to the scientists that genes have an important influence on how children score on intelligence tests.

But Plomin wants to know more. He wants to find the specific genes that are doing the influencing. And now he has a tool for pinpointing genes that he could not have even dreamed of when he began quizzing children. Plomin and his colleagues have been scanning the genes of his subjects with a device called a microarray, a small chip that can recognize half a million distinctive snippets of DNA. The combination of this powerful tool with a huge number of children to study meant that he could detect genes that had only a tiny effect on the variation in scores.

Still, when Plomin and his co-workers unveiled the results of their microarray study–the biggest dragnet for intelligence-linked genes ever undertaken–they were underwhelming. The researchers found only six genetic markers that showed any sign of having an influence on the test scores. When they ran stringent statistical tests to see if the results were flukes, only one gene passed. It accounted for 0.4 percent of the variation in the scores. And to cap it all off, no one knows what the gene does in the body.

“It’s a real drag in some ways,” Plomin says.

Plomin’s experience is a typical one for scientists who study intelligence. Along with using microarrays, they are employing brain scans and other sophisticated technologies to document some of the intricate dance steps that genes and environment take together in the development of intelligence. They are beginning to see how differences in intelligence are reflected in the structure and function of the brain. Some scientists have even begun to build a new vision of intelligence as a reflection of the ways in which information flows through the brain. But for all these advances, intelligence remains a profound mystery. “It’s amazing the extent to which we know very little,” says Wendy Johnson, a psychologist at the University of Minnesota.

In some ways, intelligence is very simple. “It’s something that everybody observes in others,” says Eric Turkheimer of the University of Virginia. “Everybody knows that some people are smarter than others, whatever it means technically. It’s something you sense in people when you talk to them.”

Yet that kind of gut instinct does not translate easily into a scientific definition. In 1996 the American Psychological Association issued a report on intelligence, which stated only that “individuals differ from one another in their ability to understand complex ideas, to adapt effectively to the environment, to learn from experience, to engage in various forms of reasoning, to overcome obstacles by taking thought.”

To measure these differences, psychologists in the early 1900s invented tests of various kinds of thought, such as math, spatial reasoning and verbal skills. To compare scores on one type of test to those on another, some psychologists developed standard scales of intelligence. The most familiar of them is the intelligence quotient, which is produced by setting the average score at 100.

IQ scores are not arbitrary numbers, however. Psychologists can use them to make strong predictions about other features of people’s lives. It is possible to make reasonably good predictions, based on IQ scores in childhood, about how well people will fare in school and in the workplace. People with high IQs even tend to live longer than average.

“If you have an IQ score, does that tell you everything about a person’s cognitive strengths and weaknesses? No,” says Richard J. Haier of the University of California, Irvine. But even a simple number has the potential to say a lot about a person. “When you go see your doctor, what’s the first thing that happens? Somebody takes your blood pressure and temperature. So you get two numbers. No one would say blood pressure and temperature summarize everything about your health, but they are key numbers.”

Then what underlies an intelligence score? “It’s certainly tapping something,” says Philip Shaw, a psychiatrist at the National Institute of Mental Health (NIMH). The most influential theory of what the score reflects is more than a century old. In 1904 psychologist Charles Spearman observed that people who did well on one kind of test tended to do well on others. The link from one score to another was not very tight, but Spearman saw enough of a connection to declare that it was the result of something he called a g factor, short for general intelligence factor.

How general intelligence arose from the brain, Spearman could not say. In recent decades, scientists have searched for an answer by finding patterns in the test scores of large groups of people. Roughly speaking, there are two possible sources for these variations. Environmental influences–anything from the way children are raised by their parents to the diseases they may suffer as they develop–are one source. Genes are another. Genes may shape the brain in ways that make individuals better or worse at answering questions on intelligence tests.

Starting in the 1960s, scientists have gotten clues about the roles of genes and environment by studying twins. To see why twins are so important to intelligence researchers, imagine that a pair of identical twins are separated as babies and adopted by different parents. They have the same genes but experience different environments. If their genes have no influence at all on their intelligence test scores, then you would expect that the scores would be no more similar to each other than those of two unrelated people. Yet if genes do play a critical part in intelligence, identical twins should be more similar.

“Two people with the same genes correlate as much as a person does with himself a year later,” Plomin says. “Identical twins reared apart are almost as similar as identical twins reared together.” But these similarities also take time to emerge. “By the age of 16 these adopted-away children resemble their biological parents’ IQ just as much as kids do who are reared by their biological parents,” Plomin adds.

Results such as these persuaded Plomin that genes have a crucial role in intelligence, although they clearly do not act alone. “That’s what led me to say, ‘What we need to do is begin to find some genes,’” he says.

In the early 1990s, when Plomin started his search for genes, he had little company. “I knew nobody else would be crazy enough to do it,” he remarks.

Plomin could not simply scan the human genome, because it had not been mapped yet. But geneticists had identified a number of genes that, if mutated in certain ways, were associated with mental retardation. Other variations in those genes, Plomin reasoned, might produce subtler differences in intelligence. He and his colleagues compared children who scored well or poorly on intelligence tests. They looked for variants of the 100 genes that showed up unusually often in one group or the other. “We didn’t really find any there,” he says.

So Plomin expanded the search. Rather than looking at a predefined set of genes, he mapped thousands of genetic markers sprinkled across the chromosomes of his subjects. If a marker turned up frequently in high- or low-scoring students, there might be an intelligence-linked gene not far away. He and his colleagues added more children to their study so that they could detect genes with weaker effects. At one point in the research, Plomin thought he had found an authentic link between intelligence and a gene known as IGF2R that encodes a growth factor receptor which is active in the brain. But when he and others tried to replicate the result, they failed. “It doesn’t look like that has panned out,” he says.

Plomin suspected that he needed more genetic markers to find intelligence genes. When eggs and sperm develop, their chromosomes swap segments of DNA. The closer two segments of DNA are to each other, the more likely they are to be passed down together. But in Plomin’s early studies, millions of DNA nucleotides separated each pair of markers. It was possible that intelligence genes were so far from a genetic marker that they were sometimes getting passed down together and sometimes not. He needed a much denser set of genetic markers to reduce the chance of this happening.

It was with great delight that Plomin got his hands on microarrays that could detect 500,000 genetic markers–hundreds of times more than he had previously used. He and his colleagues got cheek swabs from 7,000 children, isolated their DNA, and ran it through the microarrays. And once more the results were disappointing.

“I’m not willing to say that we have found genes for intelligence,” Plomin declares, “because there have been so many false positives. They’re such small effects that you’re going to have to replicate them in many studies to feel very confident about them.”

Failing to find genes for intelligence has, in itself, been very instructive for Plomin. Twin studies continue to persuade him that the genes exist. “There is ultimately DNA variation responsible for it,” he says. But each of the variations detected so far only makes a tiny contribution to differences in intelligence. “I think nobody thought that the biggest effects would account for less than 1 percent,” Plomin points out.

That means that there must be hundreds–perhaps thousands–of genes that together produce the full range of gene-based variation in intelligence. Plomin doubts that some genes are specialized just for verbal skills and that other genes are just for spatial understanding. In twin studies, individuals tend to have similar scores on tests for all of those different kinds of intelligence. If genes belonged to specialized sets, a person could inherit one kind of aptitude and not the others.

Plomin also surmises that his results offer some hints as to how genes influence intelligence in the brain. “If there are many genes of small effect, it’s highly unlikely that they are all going to focus in one area of the brain,” he argues. Instead the genes may be influencing a large network of brain regions. And each of those intelligence-associated genes may produce many different effects in different parts of the brain.

The ultimate test of Plomin’s hypothesis will have to wait until scientists finally put together a list of genes that have an indisputable effect on how the brain works and that are associated with intelligence scores. That list may take a long time to come together, but Plomin is encouraged by new results from an entirely different line of research: a burst of new neuroimaging studies that attempt to find the mark of intelligence in the brain itself.

Shaw and his colleagues at the NIMH have been analyzing brain scans of schoolchildren. Researchers have made images of their developing brains once a year, and Shaw has focused much of his attention on what the pictures reveal about the growth of the cortex, the outer rind of the brain where the most sophisticated information processing takes place. The cortex continues to change shape and structure until people reach their early 20s. And Shaw has found that differences in intelligence test scores are reflected in how brains develop.

In all children the cortex gets thicker as new neurons grow and produce new branches. Then the cortex thins out as branches are pruned. But in some parts of the cortex, Shaw found, development took a different course in children with different levels of intelligence. “The superclever kids started off very thin,” Shaw says. “They got really relatively thicker, but in adolescence they got thinner again very quickly.”

A similar pattern has emerged from studies on adult brains. Researchers have found that people with high intelligence scores tend to have certain regions of the cortex that are larger than average. Shaw expects that some of those patterns will turn out to be the result of the environment. But these regions of the cortex tend to be the same size in twins, indicating that genes are responsible for some of the difference as well.

In recent years, scientists have also published a number of studies in which they claim to have found distinctive patterns of brain functioning in people who score high on intelligence tests. Recently Haier and Rex Eugene Jung of the University of New Mexico surveyed 37 studies examining regional brain size or activity to look for an overall pattern to their results. As Plomin would have predicted, Haier and Jung found no one “intelligence spot” in the brain. Instead they identified a number of significant regions scattered around the cortex. Other studies have implicated each of these regions in different kinds of cognition. “It looks like intelligence is built on these fundamental cognitive processes, like attention and memory, and maybe language ability,” Haier says.

Along with describing the gray matter tissue that makes up the cortex, these studies also find the signature of intelligence in the white matter that links distant parts of the cortex to one another. People with high intelligence tend to have tracts of white matter that are more organized than other people. “The white matter is like the wiring,” Haier says. “If you think about it, you know, intelligence really requires processing power and speed; the white matter would give it the speed; the gray matter would give it the processing power.”

Haier suggests that these parts of the “intelligence network” may work differently in different people. “You can think about being very intelligent because you have a lot of speed and a lot of processing–you have both,” he says. “Or you can think about a lot of one and less of the other. All these combinations may produce the same ultimate result, so you may have two equally intelligent people, but their brains are fundamentally arriving at that behavior, however you’re measuring it, in different ways.”

Haier acknowledges that these ideas are little more than speculation. Still, he argues that neuroimaging has already given scientists a far more solid understanding of intelligence. “I can predict full-scale IQ with the amount of gray matter in a small number of areas,” he says. Haier suspects that in the near future, 10 minutes in a magnetic resonance imaging scanner may reveal as much about high school students as four hours taking an SAT exam.

Some psychologists are not quite ready to take that step. They do not think IQ and g should be endowed with a deeper significance than they deserve. For one thing, there is much beyond the life of the mind than mentally rotating cubes and completing analogies. “I think human intelligence is multifaceted and very complex,” the University of Virginia’s Turkheimer says. Unfortunately, he adds, barely any work has been done on other facets of intelligence.

“We can use g for a lot of useful things, but I don’t believe it follows from that that human intelligence is a unitary thing called g that we can find in a literal way in the brain,” he says. Longitude and latitude are useful, too, for navigation, he notes, but that does not mean there is actually a grid carved into the earth.

Johnson of the University of Minnesota defends the g factor as tapping into something important in the brain, but, like Haier, she does not think it is a one-size-fits-all general intelligence. “While there is something general about intelligence, what makes my intelligence general is not the same thing that makes your intelligence general or any other person’s,” she says. “Our brains are plastic enough that we put together, each of us, a different kind of general intelligence.”

Pinpointing the role genes have in producing different kinds of intelligence will no doubt be very difficult. And it is entirely possible that a list of intelligence-linked genes may include many that do not actually have brain-specific functions. Turkheimer offers the following thought experiment: Imagine a gene that is related to the width of a woman’s birth canal. Women who carry a gene for a narrow birth canal tend to have more trouble in labor, and their babies run a higher risk of being oxygen-deprived. As a result, their babies, on average, have IQ scores a couple points below those who have a different version of the gene. And some of those children will also carry the narrow-canal gene.

“These babies are going to have a gene that’s correlated with low IQ,” Turkheimer says. “So do you conclude that that’s an IQ gene? Well, not really; it’s a birth-canal gene. The ways that genes could correlate with IQ are so variable that it’s almost impossible to know.”

Turkheimer’s own research illustrates another kind of complexity in the link from genes to intelligence: genes do not act in isolation from the environment. In fact, the same gene can have different effects in different environments. Turk­heimer was led to this realization when he noticed that the large twin studies on intelligence contained few poor children. “Very poor people don’t have the time or the resources or the interest to do volunteer studies,” he says.

Other databases, Turkheimer discovered, have more poor children in their ranks. He was able to analyze the test scores of hundreds of twins, taking into consideration their socioeconomic status–a category based on factors that include a family’s income and the education level of the parents. He found that the strength of genes’ effect depended on the socioeconomic status of the children. In children from affluent families, about 60 percent of the variance in IQ scores could be accounted for by genes. For children from impoverished families, on the other hand, genes accounted for almost none.

Turkheimer and his colleagues published these results in 2003; in May 2007 they replicated the pattern with another database. Instead of comparing IQ scores, the researchers looked at how 839 pairs of twins fared on the National Merit Scholarship Qualifying Test in 1962. Once more genes played little role in the variance of scores among poor children and played a far stronger one in more affluent children. Turkheimer posits that poverty brings with it powerful environmental forces that may shape intelligence from the womb through school and onward. But when children grow up in the relative stability of an affluent home, gene-based differences can begin to make themselves felt.

As if this complexity was not enough, scientists are also finding that genes, in turn, can alter the effect that the environment has on our intelligence. Last year British scientists found an association between breast-feeding and a boost in IQ test scores–but only if children carried a particular variant of a particular gene. If they carried another variant, breast-fed children scored no differently than children who drank formula.

Genes may also influence behavior in ways that influence how intelligence develops. “People create their own environments,” Johnson says. “If you see a kid who’s really interested in art or math, you’re just more likely to go out and get him a math book or some crayons. So they’ll practice it and become more different from the kid who doesn’t have the math book. Parents respond to what the kid does. Our models don’t measure that well at all.”

This effect might explain one of the most puzzling patterns in twin studies on intelligence: how the influence of genes becomes stronger on test scores as people get older. Genes may affect how people mold their intellectual environment. Choosing to seek out new experiences, reading books and engaging in conversations may alter the brain. And as children grow up and take over control of their own lives, this effect may get stronger.

“Intelligence is kind of an emergent property of the brain,” Shaw says. “The idea that you’re born with 15 genes, and they set in stone how intelligent you’re going to be and how your brain is going to develop, is almost certainly wrong.”

Intelligence may be enormously complex, and scientists may have made frustratingly little progress in understanding it. Yet many experts on intelligence still see some practical values in continuing the quest. Haier, for example, hopes that a brain-based understanding of intelligence will help teachers design strategies for educating children most effectively.

“It’s very important as we enter the 21st century to maximize and optimize education for people,” he argues. “That’s what’s at stake.”

By understanding people’s genetic profiles, Plomin suggests, it may be possible to find the best ways to foster learning. If, as he anticipates, microarray studies finally do reveal intelligence genes, it may then be possible to test children for which versions they carry.

“You could get an index of genetic risk,” Plomin says. “You could see which kids are at genetic risk for reading disability, and you could then intervene. The hope is that you could predict and intervene with programs to prevent those problems rather than waiting until they occur in school.”

And for some psychologists, it is enough that intelligence is such an intriguing part of human nature. “Intelligence and intelligence test scores are in many ways the best predictor in all of psychology,” Turkheimer says. “That’s what makes it fascinating. If you know my SAT scores and you want to know how I’m going to perform at practically anything, those SAT scores are far from a perfect predictor, but they’re far better than knowing about my personality. Intelligence really works. There’s this palpable psychological quality that allows you to make predictions about humans but gets very slippery when you try to tie it down in concrete numerical ways. So it’s just a very interesting scientific problem.”

Copyright 2008 Scientific American. Reprinted with permission.