A mop of light brown hair shakes as a slender nine-year-old boy named Jack bangs furiously at his keyboard. Jack's eyes are fixed on a clock with six hands, which denote the month, day, hour, minute, second and 60th of a second. As soon as he types 10:28:2:14:56:32, a new clock appears, and he hammers out another set of numbers. An affable 14-year-old student named Marti had just taught me the exercise, and I guessed I could have solved one of these clocks in a few minutes. Jack was finishing one every seven seconds.

Jack's incessant clacking is virtually the only sound in this small classroom of eight- and nine-year-olds. The others work silently. One or two wear an eye patch, copying symbols onto grids. A dark-haired girl listens through headphones to a list of words she must memorize and repeat to a teacher. One boy stares at a Norman Rockwell painting; his job is to extract its main idea and write it down.

Jack, his long red sleeves poking out from under a blue school T-shirt with the initials “EAS,” cracks clock codes for 35 minutes almost nonstop. As with others at Eaton Arrowsmith School, a private facility for the learning disabled on the second floor of a building on the campus of the University of British Columbia in Vancouver, Jack attends six periods of brain exercises, among them reading clocks, copying symbols, tracing complex designs, memorizing patterns and performing mental arithmetic. Jack has one period of English and one period of mathematics, but none of his other classes resemble those in an ordinary school.

Arrowsmith students do not learn about the branches of government or genres of literature. Their time is devoted instead to fortifying mental processes such as attention, memory and reasoning. For decades, psychologists have thought that such fundamental thinking capacities were fixed. In research circles, evidence is mounting that they may not be.

The Arrowsmith program is intensive, requiring students to dedicate 80 percent of their day for three to four years to brain remediation before returning to regular school. Although the school boasts a number of success stories, support for its regimen's effectiveness is mainly anecdotal. Nevertheless, a commitment to its style of brain fitness for children is gathering steam in the scientific, clinical and even mainstream educational communities.

A small collection of brain-training workouts has emerged from neuroscience and psychology laboratories in recent years, and several are now being marketed and sold. Some build working (short-term) memory, a kind of mental whiteboard that is linked to intelligence. Others target basic number sense, for math, or sound perception, for reading. Another trains reasoning.

In many cases, the tools are aimed at learning problems such as dyslexia, dyscalculia or attention-deficit/hyperactivity disorder (ADHD). But some educators are starting to offer brain training to the general school population. “I see the technology as making it possible to individualize training and learning for everybody,” says psychologist Allyson P. Mackey of the Massachusetts Institute of Technology. “Even kids performing fairly well might have a weakness, and if you patch that up, they would perform much better.”

The extent to which children can overcome intellectual deficits or raise their IQ through mental calisthenics is largely unknown. Although data suggest that the training can be useful, it does not always work. In addition, researchers are only now beginning to explore whether the measured gains in children's thinking skills translate into academic achievement.

Still, many scientists and educators believe that with the proper tools, students can increase their intellectual capacity—an idea that could transform lives. “The question is: What are we capable of as human beings?” asks Howard Eaton, a learning disabilities specialist who founded Eaton Arrowsmith. The notion that people can fundamentally alter their brain, he says, “changes your whole perspective on human possibility.”

Levers for Learning

Of course, school has long been based on the premise that the brain is flexible: learning new information and skills involves changes at the neural level. Yet people have assumed that a person's capacity to learn is fairly stable. Part of this capacity lies in executive functions, a set of faculties governed by a structure called the prefrontal cortex that sits just behind the forehead. These faculties are working memory, cognitive flexibility—the ability to find alternative solutions to problems and shift from one idea or action to another—and self-regulation, the ability to inhibit competing or inappropriate actions.

School is not traditionally designed to alter executive function, nor is it typically structured to tweak basic math ability or facility for listening to language. It does not do these things, in part because people assumed those basics of the brain were immovable.

But it is no secret that environment can have a powerful effect on intellectual capacity. Stressful circumstances such as those that accompany poverty, for example, can virtually shut down executive functions. A child's socioeconomic class can also strongly influence language skills that are fundamental to reading. “For a four- or five-year-old, the difference in language exposure between a child from a low socioeconomic class and one from a high socioeconomic class can be as much as 30 million words,” says neuroscientist William M. Jenkins, chief scientific officer at Scientific Learning in Oakland, Calif.

In that light, the notion that tailored coaching could boost a child's potential to learn is less difficult to fathom. In recent years researchers have designed curricula to promote self-regulation, a skill essential to both academic success and social and emotional maturity. Meanwhile other scientists have set out to design interventions aimed more squarely at an individual child's intellectual capacity.

One critical lever on intellect is working memory. Various cognitive skills depend on this mental scratch pad. Attention, in particular, entails mentally taking note of important information. “If you can't hold a plan in mind, you'll get distracted,” says cognitive neuroscientist Torkel Klingberg of the Karolinska Institute in Sweden.

In 1997 Klingberg was studying the neural basis of working memory when he came across a paper showing that kids with ADHD very often had limited working memory capacity. Although working memory was widely believed to be a static trait, Klingberg was radically optimistic about its pliability. “I thought of it as a muscle that could be trained,” he recalls.

With attention deficits in mind, Klingberg and his colleagues created workouts for recalling locations—such as directions to a shopping mall—as well as verbal information. In some exercises, users try to reproduce the order in which an array of red bulbs or asteroids light up. As with all good training programs, these adapt to the child: as his or her performance improves, the game gets harder. At higher levels, the asteroids move, or the grid of lights rotates before the player has to recall the order. A verbal task requires remembering a series of digits and repeating it in reverse.

In a study published in 2005 by Klingberg's team, 22 kids aged seven to 12 who had severe attention problems played these games for 35 to 40 minutes a day for 25 days. The children improved significantly more on a standard assessment of working memory than did 22 kids with ADHD who used much easier versions of the same games. In addition, parents of the trained youths said their children became more attentive. Based on such results, Klingberg founded a company called Cogmed, now owned by Pearson, the education firm, to market the software. (Klingberg is no longer associated with Cogmed.)

Scientists have since garnered additional support for these games as remedies for ADHD. In 2010 psychologist William B. Benninger of Ohio State University and his colleagues found that children and adolescents with ADHD who did the drills at home became more attentive and better organized and had fewer symptoms, according to their parents, than those who did not exercise their recall. In 2012 psychologist Julie B. Schweitzer of the University of California, Davis, School of Medicine and her associates reported that the training significantly reduced “off-task” behavior in 12 children with ADHD while doing schoolwork—that is, looking away from a work sheet, a more real-world measure of focus.

Working memory can have a profound effect on learning in general. Among children who score in the lowest 10 percent of the population on working-memory tests, more than four fifths have considerable problems in reading or math, or both. From a test of 345 children between the ages of eight and 11, psychologist Darren L. Dunning, now at the University of East Anglia in England, and his colleagues identified 42 children who fell in the lowest 15 percent in working-memory ability. They assigned 22 of them to intensive, in-school Cogmed training for five to seven weeks; the others received a less taxing version of the program. By the end of the instruction, the children who did the more intensive exercises showed a big boost in all aspects of working memory, whereas the other kids reaped only minimal gains. Moreover, six months later the kids who got the real training scored significantly higher on a standardized test of mathematical reasoning than they had at the beginning of the trial, indicating that they used their trained brains to learn more math.

Getting Out of a Jam

Another critical component of academic success is reasoning, the capacity to think logically, connect ideas and solve problems in novel situations. Reasoning is a higher-level skill that depends on executive functions such as working memory and attention. The clock puzzles Jack solves so expertly are supposed to train reasoning by building a child's understanding of relations such as those between the different hands on a clock.

Several years ago M.I.T.'s Mackey, then a psychology graduate student at the University of California, Berkeley, wanted to see if she could sharpen reasoning in disadvantaged children. Collaborating with her colleague, psychologist Silvia A. Bunge, Mackey selected computer and commercial board games that rely heavily on reasoning. In one board game, called Rush Hour, players need to figure out how to get a car to escape a traffic jam while still obeying the laws of the road. Other games depended on logic or on integrating different pieces of information.

They asked 17 students, aged seven to 10, from an elementary school in a high-poverty neighborhood in Oakland, Calif., to play the games for an hour a day, two days a week, for eight weeks. Another 11 students played games that taxed processing speed—how quickly they could make sense of information—instead. The kids who played the reasoning games saw their scores jump by more than 30 percent on a standard test of that skill—and their IQ scores rose 10 points on average. The students who played games that exercised processing speed upped their ranking on a test of that ability by 30 percent. “We were really surprised at how big the gains were,” Mackey says.

She is now trying to reproduce her results in a larger sample of kids at risk for school failure and determine whether the training translates into gains in academic achievement. “If we can show these kinds of games lead to better test scores, we've taken a huge step forward,” Mackey notes.

In a 2014 study, Andrea Paula Goldin of the University of Buenos Aires and her colleagues took that step with computer training. They asked half of more than 100 six-year-olds from disadvantaged backgrounds to play three challenging computer games designed to train working memory, planning and inhibitory control. They found that the 10-week intervention improved grades in math and language arts among those who received the cognitive training and whose grades were low at the start of the study.

Five Dots or Six?

Computerized training programs can bolster basic math ability, too. Doing mental math depends heavily on working memory, which we use to hold and manipulate numbers. One Arrowsmith exercise involves adding one small number to the next as the digits appear sequentially, keeping a running total, and reporting the sum at the end.

Other programs train number sense, a basic sense of quantity that enables us to immediately compare, say, the number of dots in two different arrays or subtract or add dots. Scientists have tied this sense to a location in the brain: a narrow indentation on its surface called the intraparietal sulcus. Without a well-developed sense of number, children will have trouble with math and may develop dyscalculia, a mathematical learning disability that afflicts up to 7 percent of the school-aged population. A nine-year-old with dyscalculia might, for example, confuse five dots with six or be unable to say whether 50 is greater or less than 100.

About 10 years ago neuroscientist Stanislas Dehaene of the French National Institute of Health and Medical Research and his colleagues created a Web-based game called Number Race, in which players compare quantities of dots and associate them with number symbols and learn some basic addition and subtraction facts. In 2006 the researchers reported that 15 seven- to nine-year-olds with dyscalculia who played Number Race got somewhat better at comparing numbers, making quick visual assessments of quantity and subtracting one-digit numbers [see “How to Build a Better Learner,” by Gary Stix].

Three years later the researchers tested the software on younger children at risk for math difficulties: 53 French kindergartners from low socioeconomic backgrounds. Some of the children played Number Race for six 20-minute sessions, and others used a commercially available reading program, after which the kids switched tasks. Number Race, but not the reading software, improved the kids' ability to compare numbers represented as symbols, suggesting that the program honed the ability to connect number symbols to quantity. The team has since developed a more advanced game, called Number Catcher, that exercises basic calculation skills and represents numbers in different ways.

Lumosity, a company based in San Francisco, offers a suite of brain-training games aimed at various capacities, including several related to math. One involves arithmetic problems that appear in falling raindrops that must be solved before the drops splash into water at the bottom of the screen. In another, users compare the value of mathematical expressions presented in pairs.

Such drills have shown promise in boosting math proficiency in children with inborn impediments to learning the subject. In a 2011 pilot study, psychiatrist Shelli R. Kesler of Stanford University and her colleagues found that playing these games 20 minutes a day for six weeks ameliorated characteristic math-related deficiencies in 16 girls who had a genetic condition called Turner syndrome. After the training, the girls scored significantly higher on tests of number sense, processing speed and cognitive flexibility.

Sounding It Out

As with math, reading involves a complex set of intellectual capacities, including reasoning and executive skills. Dyslexia, a reading disability in which the brain has trouble recognizing and processing words, afflicts 5 to 17 percent of children. In the early 1990s neuroscientist Paula Tallal of Rutgers University hypothesized that a root cause in many cases was auditory, specifically a deficit in detecting rapid changes in similar sounds, such as “da” and “ba.” Such difficulties, she argued, prevented children from acquiring good language skills, leading to dyslexia.

In 1996 Tallal and neuroscientist Michael Merzenich, now an emeritus professor at the University of California, San Francisco, founded Scientific Learning to develop computer software to correct auditory-processing problems in children with reading difficulties. The program, called Fast ForWord Language, helps children hear and discriminate phonemes by first slowing them down and emphasizing certain rapidly changing parts of speech, Jenkins says. Then it gradually speeds up and softens the emphasis. The software also morphs the sounds, making them more or less similar, depending on a child's proficiency.

Several small studies indicate that the program is helpful. In 2007 neuroscientist Nadine Gaab of Harvard Medical School, along with Tallal, among other colleagues, reported significant improvement in language and reading skills in 22 children with dyslexia who used the program 20 minutes a day five days a week for eight weeks. They also saw in these youngsters increased activation in brain circuitry responsible for processing rapidly changing sounds. In a 2015 study, a team led by neuroscientist Sabine Heim, also at Rutgers, revealed small changes in brain activity in 21 school-aged children with language impairments after a month of intensive training using Fast ForWord. The changes hinted at improvements in sound, language and cognitive processing in these children.

Not all studies back up Fast ForWord's efficacy, however. In a 2011 meta-analysis (quantitative review) of six large studies, psychologist Charles Hulme of the University of York in England, and his colleagues, concluded that the program had little effect on children's language or reading difficulties. The mixed results may reflect differences in how the software was tested, including the degree to which adults monitored its use and motivated children to engage with it. After all, kids who quit using the program out of frustration or boredom are unlikely to benefit. In addition, not all children with dyslexia suffer from auditory problems. Indeed, the latest versions of Fast ForWord are designed to serve a wider swath of the student population by drilling reading skills, from phonological awareness to comprehension strategies, along with executive functions.

Games in the Mainstream

Most of the current student brain-training programs are aimed primarily at those with diagnosable deficits. The many U.S. schools that have adopted the Cogmed software, for example, typically offer it to students with attention deficits and other learning disabilities. More than 500 clinics in the U.S. and Canada also use Cogmed, primarily with clients who have ADHD.

Yet the programs are also trickling into the educational mainstream. Although Fast ForWord is geared toward kids with reading problems, many of the more than 4,000 schools that have the software include it as part of regular instruction. So far 14,000 students in more than 500 classrooms worldwide have played a suite of brain games offered through the Lumosity Education Access Program (LEAP), and more than a quarter of Lumosity.com online users are younger than 21. At least one private parochial school in New York City has made Cogmed software available to the kids in its fifth and sixth grades. “This isn't limited to students with learning difficulties,” says Nicole F. Ng, a former teacher who now manages LEAP. “It applies to a healthy student who wants to improve cognitive capabilities.”

Some data hint that brain training could benefit the typical learner. In an unpublished study that was conducted by Ng and her colleagues, 949 students aged six to 18 years in 43 schools played 28 Lumosity games for an average of six hours total during a semester. These youths raised their scores on a battery of neuropsychological tests significantly more than did 443 students who did not perform the exercises. The more hours a student trained, the more he or she improved on skills such as memory, processing speed and reasoning. Students who spent at least 10 hours of training saw measurable benefits, Ng says.

Yet doubts have emerged about the significance of those benefits. In October 2014 more than 70 psychologists and neuroscientists signed a statement released by the Stanford University Center on Longevity and the Max Planck Institute for Human Development in Berlin stating that there is scant scientific backing for claims that “brain games” can sharpen everyday mental functioning. And in January 2016 the Federal Trade Commission charged that Lumosity's ads were misleading in claiming that playing their games three to four times a week could boost performance at work or in school. Lumos Labs, the developer of the Lumosity games, paid $2 million to settle the charges.

Before arriving at Eaton Arrowsmith, Marti had struggled in regular school. “I'd be staying up so late trying to reread and understand,” she told me in 2012. “I ended up crying because I was overwhelmed with homework.” Marti has since graduated from the program, and her ninth grade report card in December 2012 displayed straight As. School officials have data on numerous children who, like Marti, have been helped by the curriculum.

Outside scientists often find the Arrowsmith approach intriguing but say it lacks rigorous scientific support. “I saw one kid doing math on a computer faster than I could do it,” says Adele Diamond, a developmental cognitive neuroscientist at the University of British Columbia. But, she warns, “I'd like to see data that it helps.” Indeed, one small, eight-month investigation of the school's curriculum led by educational psychologist Linda Siegel, also at the University of British Columbia, failed to show that it significantly improved students' scores on a battery of cognitive and achievement tests. New data are likely forthcoming, however. Microsoft CEO Satya Nadella and his wife, Anu, donated $100,000 for brain imaging and other research to evaluate the Arrowsmith program.

Science does increasingly suggest that the brain is far more supple than we once assumed. Eventually the educational community may decide that the data support adding at least a dollop of brain fitness to children's usual scholastic fare. Although no one knows exactly the form such training will take, putting children's mental muscles through the paces on a regular basis could lead to lasting benefits. “I envision improvement of cognitive skills as part of education much more than it is right now,” Mackey says.