Have you heard the news? Brain training is a scam: overhyped and understudied—a waste of time worse than Angry Birds.

That, at least, is the gist of news reports that ran in 2016, when the country's best-known brain-training company, Lumosity, was fined $2 million by the Federal Trade Commission (FTC) for making exaggerated claims in its advertisements. The FTC action came 15 months after the release of a statement from more than 70 neuroscientists who said they objected to unsubstantiated claims from the burgeoning brain-game industry. “Despite widespread claims, there is little evidence that brain-training games provide easy boosts to cognitive function,” concluded an article by Simon Makin published in these pages in July 2015. “Making brain training look effective is easy because performance on the games inevitably gets better with practice.”

But is that all there is to it? Does training on cognitive games result only in improvements on those games, as critics allege, with little evidence of meaningful real-world benefit? No young science is without controversy, and a number of studies by reputable scientists have indeed failed to find substantive cognitive benefits from such training. Yet shortly after the release of the “consensus” statement against brain training in 2014, more than 120 other scientists signed a response that cited 132 published studies showing that brain training does work.

No one is claiming that brain games will transform an average Joe into a Shakespeare or an Einstein. But there is plenty of evidence that computer-based cognitive training offers real benefits for certain populations. Most notably it can cut an older person's risk of having a car accident in half, mitigate the loss of basic cognitive abilities in people with schizophrenia, and improve the working memory of children who have attention deficits or are recovering from cancer. Quietly and persistently, outside the spotlight of news coverage, mainstream scientists at leading medical institutions around the world are building an ever stronger case for cognitive training targeted to such groups. Supported by the National Institutes of Health, the U.S. Naval Research Laboratory and other prominent funding organizations, they have published hundreds of randomized, placebo-controlled clinical trials in peer-reviewed medical journals.

These researchers worry that unjustified claims akin to those that drew the FTC's ire will tarnish the field and cause their own findings to be thrown out with the bathwater. “As an advocate and researcher for children who have, at this point, no other viable options, it's frustrating for me to hear the attacks on cognitive training,” says psychologist Kristina K. Hardy, who works with childhood cancer survivors at Children's National Health System near Washington, D.C. (and who reports no financial connection with any brain-training company). “The science is promising. There's every reason to be optimistic about this approach.”

Aging, Interventions and Independence

We should first get one thing straight: there is no evidence that cognitive training can prevent or slow the progression of Alzheimer's disease. Even for the condition known as mild cognitive impairment, often a precursor to Alzheimer's, the evidence is murky. But a new study, presented at a medical conference last summer, may change all those conclusions if its results pass muster in a scientific journal.

It should be noted that physical exercise—in particular, resistance training—has plenty of evidence showing it builds the brain as well as the body in older adults. And when you combine exercise with a healthy diet, ample social interaction and the use of brain games, as Finnish researchers did in a randomized controlled study published in 2015 in the Lancet, you get especially good results on cognitive function. But let's zoom in on the easiest intervention: computerized cognitive exercises that anyone can do and that may require as little as 10 total hours of practice.

Perhaps the most startling study in older adults, published in 2014, involved 2,832 volunteers in six cities across the U.S., whose average age was 73 when the research began. A team at Johns Hopkins University randomized the volunteers into one of three different types of cognitive training or into a no-contact control group. The training groups targeted memory, speed of processing or reasoning; each group tackled a series of tasks that essentially taxed those skills with progressively tougher challenges. Participants completed 10 one-hour sessions, followed in some cases by four booster sessions.

The people who did memory training saw no benefit. (Go figure.) But those who had undergone training for reasoning or speed of processing did gain from the experience. A full 10 years later people in this group continued to outperform the control group on those functions. And tentative results presented in summer 2016 at the Alzheimer's Association annual meeting found that those who underwent some of the booster sessions were up to 48 percent less likely to be diagnosed with dementia 10 years later. Confirmation in a peer-reviewed journal could be a game-changer for the field.

Credit: Neil Webb

The study, known as ACTIVE (Advanced Cognitive Training for Independent and Vital Elderly), found that the most effective computerized training was the one that targeted speed of processing, specifically a game challenging people's “useful field of view,” or UFOV. For a brief blip of a moment, the player sees one of two similar-looking objects in the center of a computer screen and, simultaneously, a third object on the far edge of the screen. The challenge is to correctly identify which of the two objects was in the center and where exactly the object on the periphery was. Easy at the beginning, it becomes harder as the images flash by ever more quickly. Most everyone, however, gets more accurate at faster speeds over a period of days and weeks.

Psychologist Karlene Ball, now at the University of Alabama at Birmingham, first developed the task in the 1970s. In 2010 a study by Ball and her colleagues involving 908 older drivers found that 10 hours of training could cut the participants' rate of motor vehicle crashes in half for up to six years later. The training has also been shown to help drivers respond four tenths of a second faster in response to an unexpected obstacle. “I refer to that as the difference between seeing a deer cross ahead of you on the highway and seeing it crash through your windshield,” says gerontologist Fred Wolinsky of the University of Iowa, who has collaborated on studies involving Ball's UFOV game.

The results have been impressive enough to gain the interest of the American Automobile Association and some car insurers, who are offering the training free or at a discount. “I've been working on this for the past 40 years of my life,” Ball says. “I was very surprised that, first, we could get as much improvement as we did and, second, that it lasted for so long.” Volunteers who participated in the same 10 hours of UFOV training, she found in a separate study, were also 30 percent less likely to exhibit symptoms of depression five years later.

Ball licensed the game in 2008 to Posit Science, a company co-founded by neuroscientist Michael Merzenich, formerly at the University of California, San Francisco. Customers can access versions of the game at www.drivesharp.com.

But UFOV games are not the only ones shown to improve, or at least preserve, cognitive abilities in older people. In November 2015 scientists at King's College London and their colleagues—using data from a six-month online experiment involving 2,912 adults older than 60—reported that participants randomized to training in reasoning or general cognitive skills showed enhanced performance on reasoning tasks that were different from the ones they had practiced. They also self-reported that they were better able to participate in essential activities of daily living. Most interestingly, there was a dose-response effect, so the more people trained, the better they did.

Working-Memory Workout

Another area of intense research is the use of computerized games for children and adults who have attention deficits. The games target working memory, which is the ability to maintain and juggle multiple items in your head—the cognitive equivalent of chewing gum while walking. You need working memory to recall the beginning of a paragraph even as you are reading the end or to do mental math. Working memory is essential for learning, reasoning and comprehension.

The strength of an individual's working memory was long believed to be a fixed trait until a small 2002 study by cognitive neuroscientist Torkel Klingberg of the Karolinska Institute in Sweden suggested it might be developed. He had children practice four tasks designed to stretch memory muscles, such as listening to a string of numbers or letters and then recalling them in reverse order. After a total of 10.5 hours of practicing games that put progressively tougher demands on their working memory, children diagnosed with attention-deficit/hyperactivity disorder (ADHD) showed improvements on other, untrained measures of working memory.

Credit: Neil Webb

Since then, more than 200 studies of working-memory training have been published in the scientific literature about adults and children, healthy or diagnosed with ADHD and other conditions. Not all the studies have demonstrated benefits; a 2013 meta-analysis concluded that “current findings cast doubt on both the clinical relevance of working memory training programs and their utility as methods of enhancing cognitive functioning in typically developing children and healthy adults.”

But a more recent, 2015 meta-analysis from a group of Dutch researchers found reliable evidence of gains for children and adolescents with learning disabilities. And a second 2015 meta-analysis of 12 previously published studies involving children or adults with ADHD or other working-memory impairments found “persisting training benefits for inattention in daily life.”

Klingberg's working-memory games are now offered by psychologists and other trained therapists through a program called Cogmed, owned by Pearson Education, an education and testing company based in London. (The cost varies depending on the therapist, but most charge fees between $1,500 and $2,000 for a full course of training.) Klingberg, who reports no current financial or other relationship with the company, says that although the effects on hyperactivity or impulsivity are nil, the program has been shown to reduce symptoms of inattention by roughly one third of a standard deviation. “That's pretty good,” he notes. “We shouldn't expect miracles.”

Pearson says it has no plans to seek approval from the U.S. Food and Drug Administration for its game as a treatment for ADHD, but both Posit Science and a Boston-based enterprise called Akili Interactive Labs have said they are working toward that goal. Akili's games are based on research by another U.C.S.F. neuroscientist, Adam Gazzaley. A pilot study presented in October 2015 at the American Academy of Child and Adolescent Psychiatry showed that their games were not only safe but engaging and helpful in improving attention, working memory and impulse control in a group of 80 kids with ADHD.

Cancer and Cognition

After chemotherapy for breast cancer, many women report suffering the mental fog of “chemo brain,” characterized by thinking and memory problems. A meta-analysis, published in October 2015 by researchers at Northwestern University's Feinberg School of Medicine, found that compared with other previously tested approaches such as drugs and physical exercise, “cognitive therapy protocols delivered after chemotherapy ... hold the most promise.” These protocols include the same kinds of approaches—aimed at improving verbal memory, attention and processing speed—that are offered by Cogmed and Posit Science.

Credit: Neil Webb

Three other randomized trials, all using Cogmed, have found significant benefits for survivors of childhood cancer. The largest of the studies, led by neuropsychologist Heather Conklin of St. Jude Children's Research Hospital (who has no financial relationship with Cogmed), involved 68 survivors of either childhood acute lymphoblastic leukemia or brain tumor who had been diagnosed with cognitive deficits following treatment. The researchers assigned half the children, whose average age was 11, to do 25 training sessions on Cogmed at their homes, with weekly telephone coaching sessions. They put the other half on a waiting list to be trained after the study was completed. Compared with children on the waiting list, the kids who did the training significantly improved their performance on a host of cognitive tests.

Moreover, before-and-after exams showed that while subjects answered working-memory problems inside an MRI, two areas of their frontal lobes used less blood to solve problems after training than they had needed before—possibly because the engine on top of their shoulders had become more efficient. “These are kids who were developing more or less normally until they got cancer and received a treatment that changed the trajectory of their brain development,” says Hardy, who collaborated on all three of the childhood cancer studies.

About 20 to 40 percent of children who survive leukemia experience long-term cognitive changes, she says, as do 80 to 100 percent of childhood survivors of brain tumors treated with radiation. “Working memory is one of the key abilities that changes in these children,” Hardy says. “It changes early and leads to reductions in IQ and academic functioning over time. That's one of the reasons we are so excited about Cogmed. It specifically targets a neurocognitive domain that's among the first to be impacted in these children.” She is now collaborating on two randomized trials aimed at preventing cognitive decline in childhood cancer survivors rather than merely trying to correct it after the fact.

Social Training and Schizophrenia

Hallucinations and delusions may be the most obvious symptoms of schizophrenia, but serious cognitive dysfunction is also commonly part of the picture, and antipsychotic medications do little to address it. Pioneering research by psychiatrist Sophia Vinogradov, now at the University of Minnesota, has shown that computerized cognitive training can make a major difference.

Partnering with Posit Science, Vinogradov and her colleagues have published more than two dozen randomized studies finding significant improvements in verbal memory, learning and daily functioning after training. Most of the studies have employed auditory training that might seem, at first glance, more fitting for people with hearing loss than for people with schizophrenia. “Sound sweeps,” for instance, present a series of tones either rising or falling, like the sound of a siren approaching or receding. Laughably easy at a slow speed, they quickly become seemingly impossible to distinguish when played at, say, 12 milliseconds. Nonsense syllables are likewise presented at progressively faster speeds, with ever increasing levels of distracting background noise.

Vinogradov settled on such auditory challenges because people with schizophrenia are known to have basic defects in sensory processing, which are believed to underlie, at least in part, the defects in higher-level thinking skills long known to accompany the disease. Her studies have now demonstrated benefits in adults with established schizophrenia, in those with recent onset, and in adolescents and young adults at high risk for psychosis. “We definitely are seeing improvements in cognition, specifically with auditory training, where we see large gains in verbal cognitive operations,” says psychologist Melissa Fisher, who collaborated with Vinogradov and is also now at Minnesota and has served as a consultant to Posit Science. “I don't think we have a definitive answer yet as to how much it is helping. But it's definitely promising.”

Credit: Neil Webb

Equally promising and perhaps even more surprising are the benefits seen in early studies of computerized social training for people with schizophrenia, who often struggle with interpersonal relationships. A 2013 paper by Vinogradov and Fisher found that combining auditory training with computerized social training significantly improved scores on tests of social perception.

In 2014 they collaborated with psychologist Mor Nahum, director of research and development at Posit Science, on a pilot study involving SocialVille, a program designed to strengthen social cognition. “One exercise shows a person with facial expressions, and then you have to match it to the same expression afterward,” Nahum explains. “A healthy adult might need to see this face for only 15 milliseconds to be able to correctly identify it afterward, whereas a [person with schizophrenia] needs much longer—orders of magnitude longer. The differences are striking. So they train to get better.” After 24 hours of online game play from home or a clinic, people with schizophrenia significantly improved their performance on SocialVille and on standardized psychological measures of social cognition, social functioning and motivation.

Now Nahum is working with Vinogradov, Fisher, and others on a four-site, randomized trial of SocialVille with 128 patients aimed at gaining FDA approval as a treatment for the social deficits associated with schizophrenia—a development that, if successful, would be a milestone in the field of computerized cognitive training. “Our goal is to move this out of the lab and into the community,” Fisher says.

To Incrementality and Beyond

Researchers such as Fisher insist that the cloud hanging over the field of cognitive training in the wake of the action against Lumosity has a silver lining. “In the short term,” she says, “there has been concern that all cognitive training will be seen as fraudulent. But eventually I think it will be good for our field. Lumosity really did not have the research backing up their programs. But our group and many others do have strong evidence.”

Scientists are also assessing several additional populations who could benefit from these training exercises. The SocialVille program, for instance, is now being studied as a means of strengthening the interpersonal skills of people with autism. A handful of studies have shown that computerized training can better the mental functioning of those with Parkinson's disease. A small-scale study involving 21 children with Down syndrome found that Cogmed boosted their short-term memory. Another study published in November 2015 concluded it could enhance working memory in children with epilepsy. And a third, published in February of the same year, indicated that such training has a long-term benefit in preschoolers who had been very low birth-weight babies. All these results need to be confirmed by additional and larger studies.

Should positive results continue to accumulate, the next step is to find a way to make computerized cognitive training affordable for those who could benefit from it. “Most people can't afford to pay $1,500 to $2,000 out of pocket for this intervention,” Hardy says. “It already has as much evidence behind it as some of the drug-based interventions that are covered by insurance. There isn't an fda approval for stimulant medications given to survivors of childhood cancers. Yet that certainly is covered by most insurance plans for these kids, and it is prescribed all the time.”

And what about the great white whale of cognitive training: increasing intelligence in people with average or even above-average intelligence? Some academics who specialize in the study of IQ insist that improving it is a mission impossible. But that has not deterred the U.S. Office of the Director of National Intelligence from funding research into making intelligence analysts (read: spies) more, well, intelligent. Under the direction of the Intelligence Advanced Research Projects Activity, a program with the apt acronym “SHARP” (Strengthening Human Adaptive Reasoning and Problem-Solving) is now in its third year of sponsoring studies of computerized training and other methods, including physical exercise, mindfulness meditation and mild electrical stimulation of the brain.

Results from the SHARP program will likely be published in scientific journals later this year. “I don't want to steal the thunder of the research teams, and it's important to say that not everything worked,” says neuroscientist Alexis Jeannotte, the program's manager, “but I will say that they have seen some small but significant gains in IQ. We really feel there's strong evidence for what they're seeing. They have tested interventions in literally hundreds of subjects in multiple sites.”

Jeannotte's phrase “small but significant” is key. There are no shortcuts in the realm of human intelligence, no “limitless” brain boosters. We should be skeptical of anyone who promises them—particularly if they stand to make a profit. But there is good evidence to convince any skeptic that incremental and compounding improvement in cognitive function is achievable.

After all, our brain is clearly receptive to persistence and practice in education (just as negative events such as trauma or abuse hurt our cognitive abilities). Malleability is the defining attribute of the astonishing human brain. New studies of cognitive exercises and brain games—especially in vulnerable populations—underscore that. And that means, as Samuel Clemens might have put it, the reports of brain training's death have been greatly exaggerated.