October 23, 2014
The New York Times
Clive Thompson

“You just crashed a little bit,” Adam Gazzaley said.

It was true: I’d slammed my rocket-powered surfboard into an icy riverbank. This was at Gazzaley’s San Francisco lab, in a nook cluttered with multicolored skullcaps and wires that hooked up to an E.E.G. machine. The video game I was playing wasn’t the sort typically pitched at kids or even middle-aged, Gen X gamers. Indeed, its intended users include people over 60 — because the game might just help fend off the mental decline that accompanies aging.

It was awfully hard to play, even for my Call of Duty-toughened brain. Project: Evo, as the game is called, was designed to tax several mental abilities at once. As I maneuvered the surfboard down winding river pathways, I was supposed to avoid hitting the sides, which required what Gazzaley said was “visual-motor tracking.” But I also had to watch out for targets: I was tasked with tapping the screen whenever a red fish jumped out of the water. The game increased in difficulty as I improved, making the river twistier and obliging me to remember turns I’d taken. (These were “working-memory challenges.”) Soon the targets became more confusing — I was trying to tap blue birds and green fish, but the game faked me out by mixing in green birds and blue fish. This was testing my “selective attention,” or how quickly I could assess a situation and react to it.

After only two minutes of play, I was making all manner of mistakes, stabbing frantically at the wrong fish as the game sped up.

“It’s hard,” Gazzaley said, smiling broadly as he took back the iPad I was playing on. “It’s meant to really push it.”

“Brain training” games like Project: Evo have become big business, with Americans spending an estimated $1.3 billion a year on them. They are also a source of controversy. Industry observers warn that snake-oil salesmen abound, and nearly all neuroscientists agree there’s very little evidence yet that these games counter the mental deficits that come with getting older. Gazzaley, however, is something of an outlier. His work commands respect from even the harshest critics. He spent five years designing and testing the sort of game play I had just experienced, and he found that it does indeed appear to prompt older brains to perform like ones decades younger. (“Game changer,” the cover of Nature magazine declared when it published his findings last year.) Now Project: Evo is on its own twisty path — the Boston company that is developing it, Akili, which Gazzaley advises, is seeking approval from the Food and Drug Administration for the game. If it gets that government stamp, it might become a sort of cognitive Lipitor or Viagra, a game that your doctor can prescribe for your aging mind.

In recent years, neuroscientists have begun to map, in increasing detail, just what happens as the brain ages. The picture is bleak. Beginning in our late 40s and 50s, our working memory dims, and we lose the ability to juggle simultaneous tasks. It becomes harder to screen out distractions, to stay focused while reading or shopping. Processing speed — that is, the brain’s ability to react to stimuli — slows, which is one reason older people struggle to follow the speech of chattering children. Scientists have begun to trace the physical changes behind this decline. For example, the myelin sheathing that covers the brain’s white matter degrades, and the brain has a harder time coordinating its different regions engaged in a mental task. This dropoff has nothing to do with Alzheimer’s or dementia; this is normal aging in an otherwise healthy adult. “It’s a rough life, being a nervous system over 60 or 70 years,” says Jonathan King, who directs a cognitive-aging program at the National Institute on Aging.

Since Gazzaley began his career two decades ago, in his 20s, he has been fascinated by the puzzle of aging. Back then, neuroscience was in the midst of the “neuroplasticity” revolution, the discovery that the mature brain can change and evolve. Scientists used to believe that once you became an adult, your brain’s capabilities were fixed, like plaster. But in the 1990s and early 2000s, aided by new brain-scanning tools, they realized this wasn’t true. If you start doing something that taxes your brain in productive ways, forcing it to repeatedly engage declining skills — learning a new language, for instance — those skills get measurably sharper. The problem, of course, is that most of us are pretty lazy. We’re not often going to take up mentally difficult activities in our dotage.

Video games seemed like one possible shortcut. Researchers were discovering that playing them appeared to improve some cognitive abilities in children: Avid players were better at noticing visual stimuli and shifting the focus of their attention, the very tasks that old brains find difficult.

In 2005, Nintendo released Brain Age, a slightly tongue-in-cheek game that purported to “keep your mind in shape” through a blitz of visual quizzes — like the famous Stroop Effect test, in which the word “blue” is printed in black, for example, and you have to correctly name the font’s color. (Not as easy as it sounds.) The brain-training industry was born, and soon ads from companies like Lumosity were promising to “challenge your brain with scientifically designed training.” Posit Science, a company founded by the neuroscientist Michael Merzenich, produced BrainHQ games meant to improve capacities like your “useful field of view,” the scientific term for the width of your peripheral vision. (Yes, it too shrinks with age.)

The big question about brain games is whether they sharpen everyday skills. If you regularly play a memory game — like Lumosity’s version of the old classic Concentration — you’ll get better at playing the game. But does it help you recall where you left your reading glasses? Does it improve your brain overall? Research has shown scant evidence of that. Even crossword puzzles — often touted as a pen-and-paper form of brain training — seem to suffer from this problem. All they do is make you better at doing crosswords.

Gazzaley surmised that if a game prodded several different mental abilities simultaneously, learning to resolve the “interference” produced by such multitasking would strengthen the brain generally. So he asked game designers from LucasArts, who made Star Wars video games, to do some freelance work for him. “They said, ‘Well, you know, we’ve been teaching teenagers how to kill aliens for almost 20 years now, and we’re ready to do something different, do something of impact,’ ” Gazzaley says.

By 2009, their collaboration yielded NeuroRacer. A prototype of what became Project: Evo, it required players to pilot a car down a winding path at a constant speed while trying to keep from running off the road. At the same time, users had to pay attention to a stream of flashing icons and press a button on the game controller whenever a circle appeared. As you got better at playing it, the game got harder, in order to keep you at the edge of your abilities.

To test how the game affected older minds, Gazzaley sorted 46 participants between the ages of 60 and 85 into three groups. One group played NeuroRacer three times a week for a month. Another played a simplified version without the multitasking: Players either drove the car or clicked on the circles but not both during the same game. The third group didn’t play at all.

The results were stark. Older adults who played the hardest version of NeuroRacer became very good at it — as good as 20-year-olds playing it for the first time. And crucially, there was “transfer.” Standard laboratory tests used to gauge a person’s working memory and ability to sustain attention showed that the NeuroRacer vets had “improved significantly.” And those skills weren’t the ones the game was specifically designed to focus on — their improvement was just a positive side effect. The players didn’t merely become better at NeuroRacer; they also became sharper at other things. The control groups — whose members didn’t play the game or who didn’t drive and identify objects at the same time — didn’t get the same boost. They just got older.

Gazzaley could also see his subjects’ brains change on E.E.G. readings. The electrical patterns of those who played the full NeuroRacer resembled those of 20-year-olds, just as their gaming performances had. Key measures of activity in the prefrontal cortex altered, suggesting improvement in what researchers call executive control. Measures of brain “coherence” were better, too, indicating that different parts of the brain were in better communication with one another. Perhaps most remarkably, these gains held up over time. When Gazzaley brought the participants back into the lab six months later — during which time they hadn’t played the game — the multitasking players still performed better on diagnostic tests than the control groups.

The results sent a jolt through the academic-neuroscientist community. A Boston health care firm approached Gazzaley about producing a commercial version to “move this technology out of the lab into industry,” he says, which led to the creation of Akili. Gazzaley embarked on fresh collaborations. One is with Zynga, the company responsible for the hit game FarmVille, which is helping design MediTrain, a game that promotes “mindfulness” meditation techniques.

The lab where Gazzaley works is like a child’s fantasy version of a research facility. There’s a mood-lit game-playing room with a lipstick red chair and an 85-inch plasma TV, “the biggest, highest-quality TV you can buy.” When I visited the place earlier this month, his 24-year-old research associate, Cammie Rolle, jumped and waved her arms in front of the screen, as if doing aerobics. An Xbox Kinect camera tracked her movements as she played Gazzaley’s new game, Body Brain Trainer. Each jerk and leap controlled a horse on the screen, as she tried to avoid obstacles and lunged toward targets (vegetables and colors, in this case). The goal, Gazzaley said, was to see how physical activity might influence mental training. He hopes that forcing subjects to coordinate physical movements with mental pattern-matching will generate the sort of productive interference that, as in the case of NeuroRacer, strengthens multitasking and executive control.

“We think the embodied cognition will create a faster learning curve,” he said. It was certainly a workout. After two minutes of jumping and dodging, Rolle’s heart rate was over 140. They won’t know if it improved her brain, though, for another year.

As eager as Gazzaley is to promote his research, he is also reflexively cautious. He warned me several times that his findings will need more testing and pointed out that the data is tentative when it comes to how deeply it affects everyday mental performance. The truth is that despite 15 years of research, we don’t actually know how — or if, really — brain-training games work. “It’s a big, muddled mess,” says Thomas Redick, a cognitive psychologist at Purdue University. The published research is a grab bag of contradictory findings: Some experiments find minimal improvements, others none. Often studies test games with such different emphases — working memory, field of view or mental speed — that the results can’t be compared meaningfully. Findings from experiments with subjects who have conditions like A.D.D. don’t necessarily apply to the elderly. And while there is plenty of work on children, including claims made by the educational firm Pearson that its product CogMed increases their working memory, the evidence on this front is also uncertain.

In the commercial world, though, hyperbole reigns. App stores are littered with brazen claims — Elevate-Brain Training, for example, is “based on extensive research.” Ulman Lindenberger, a director at the Max Planck Institute, recently published a study that found that 100 days of cognitive training yielded a “relatively minor” improvement in working memory. Soon afterward, a German brain-training firm cited his paper on its website, despite the lack of any connection between his research and its product. The company even appropriated the Max Planck logo.

This month, an international group of 30 scientists — including Lindenberger — became so fed up that they issued “The Consensus on the Brain Training Industry From the Scientific Community,” a withering statement denouncing the hype by both companies and media. “Claims promoting brain games,” they wrote, “are frequently exaggerated and at times outright misleading.” One of the group’s organizers, Laura Carstensen, the director of the Stanford Center on Longevity, says, “I started just feeling like we were obligated — we the scientific community in the aging world — to say something and to not just sit by and have this go on.” She is particularly enraged by claims that games can forestall Alzheimer’s. “I find that unconscionable, because there’s zero evidence for that, and it is on top of the list of aging people’s fears.” Carstensen has heard from older people who assumed the games’ benefits were proven, as well as from poor patients who were forgoing other daily items to pay monthly subscription fees for brain games.

Carstensen and her colleagues recommend that people adopt more healthful habits, including regular physical activity, because a weak cardiovascular system limits blood flow to the brain. Active efforts to master new skills — a new sport, another language — may help, as does an active social life. None of these techniques, of course, have million-dollar industries intent on proving they’re good for our brains.

That said, Carstensen and other skeptics do not dismiss brain training entirely. If you’re playing these games in moderation, they say, there isn’t any harm. And maybe they help. “It’s perfectly possible that if you do this for some time that you will see some changes in your brain,” Lindenberger says. “I mean, why not?” They all want to see more rigorous research done, and many cite Gazzaley’s work as a model.

Gazzaley himself signed the letter, though he pushed the group to use less pessimistic rhetoric. “I really was a pain,” he says. One concern was that excessively negative statements might scare off research-funding agencies. “Do we really want to sabotage this?” he asked them. “Does anyone actually think there’s nothing here? My view is that from the work that we’ve done, there’s a signal. I’m a cautious optimist. If I didn’t think there was a signal, I’d be out of here — I’m not going to waste my entire career.”

Even Michael Merzenich, the founder of Posit Science — one of the oldest makers of brain-training products — acknowledges that skepticism is warranted. “There is a lot of hocus-pocus in the marketplace and a lot of exaggeration,” he told me. Yet when it comes to his own BrainHQ games, Merzenich confidently argues that they are proven to work, noting that they have been used in many more studies than his competitors’ products. “For a lot of people, in the state they’re in, sitting down in front of your iPad or computer can save their bacon,” he says. He too recommends leading an active life but says that “for time spent, there’s nothing more efficient for driving neurological change” than his games. His views make Merzenich a polarizing character in the field. When I asked Lindenberger about Merzenich, he said, “I don’t see how he can base his business on science in the sense of having evidence for these very big claims.”

So far, only one study has followed a large group of older subjects for years and also looked for real-world effects. Advanced Cognitive Training for Independent and Vital Elderly, or Active, began in 1999 with nearly 3,000 healthy older adults. One group of the subjects played a crude speed-of-processing computer game, originally developed by Karlene Ball, the chairwoman of the psychology department at the University of Alabama at Birmingham, and later acquired by Posit Science. It flashed pictures for shorter and shorter durations and challenged players to identify them. These subjects received a total of only about 10 hours’ worth of training over five weeks, with a few hours of follow-up “booster” training nearly a year later and then again three years after that.

Yet the benefits seemed to hold up. When tested fully 10 years later, the group of video-game players still performed better on a speed-of-processing test than members of a control group that received no training. More startling were some results found outside the lab: Those who had played the computer game had been involved in 50 percent fewer car accidents than the control group. The training seemingly made them more perceptive behind the wheel.

The next step toward legitimacy — beyond long and patient studies like this one — might come from the F.D.A. Several games are now going through the federal approval process. Akili is currently arranging trials of Project: Evo to treat conditions that accompany routine aging and as a tool to detect the early onset of Alzheimer’s. Eddie Martucci, Akili’s vice president, says he wants to do enough tests to prove the game’s power not just to the F.D.A. but also to doctors, so they’ll prescribe it: “It’s not our business model to squeak through,” Martucci says. “We want doctors and regulatory agencies to have no doubt.”

Merzenich’s company, Posit Science, is pursuing F.D.A. approval for a speed-of-processing game that has shown success in treating hemispatial neglect in stroke patients, a condition that causes patients to lose one side of their vision and sometimes neglect a limb. “We saw some real-world effects,” says Tom Van Vleet, a neuropsychologist who worked on the game before he was hired by Posit Science. After playing the game, he claims, “one guy used his arm for the first time in six years.”

Ann Stewart, who is 66, participated in Gazzaley’s NeuroRacer study. Recently I visited her cozy, art-filled home, where she has lived alone since her husband died several years ago. Witty and white-haired, Stewart hardly seemed cognitively challenged. Before retiring, she was a partner in a commercial real estate firm. (“I had a huge Rolodex of brokers and people,” she said.) She also reads avidly and sings; not long ago, she attended a Big Sur musicians’ camp. But in her 60s, she found herself becoming forgetful, leaving her purse in the car overnight. “That was scary,” she told me. When she heard about the trial, she jumped at the opportunity.

She became quite addicted to NeuroRacer — and good at it. “I was really kind of sad when they took it away from me!” she said. Her experience illustrates both the potential and the limits of brain training. Technically, she’s part of the cohort who improved and says she feels she became better at multitasking, “more conscious of what I was doing every day.” But she struggled when I asked her for examples of how this had improved her daily life. In fact, to keep from forgetting her purse, she hit upon a more prosaic, low-tech solution: She bought a bigger one that’s harder to leave behind accidentally.

The elderly have long been masters of devising clever tricks to compensate for mental failings, turning objects all around them into cognitive props. Medicine might be left on the kitchen table, its presence there a daily reminder that pills need to be taken. To-do lists on Post-it notes serve as scaffolds for their memory. If you’ve already lost cognitive function, and brain training can only go so far, you find other ways to cope.

It might very well be that equally promising technology for our brains will augment rather than improve them. Already technology firms are developing methods to help the elderly by offloading memory and cognition, creating digital tools more sophisticated than oversize purses. The company Vitality, for example, has created GlowCaps, pill bottles that track when they’ve been opened. If users forget to take their pills on time, LEDs in the bottle cap might light up as a reminder; if it goes unopened for hours, the bottle sends an alert by email or text. A 2010 trial found that users of these smart bottles had a 98 percent rate of taking meds on time, compared with 71 percent in a control group. (Full disclosure: I have given an unpaid talk about my book “Smarter Than You Think” for Partners Healthcare, the nonprofit agency that financed this experiment.) A small study involving a similar device in China also found positive results.

You can imagine a world where tools like this offer compensating help in even more active ways. Self-driving cars could significantly help seniors cope with the mental challenges associated with driving. Wearable computers like the Apple Watch or Google Glass could combine GPS and location sensors to remind our aging brains of tasks we planned to do: If you’re at the grocery store, don’t forget the dog food.

In younger people, digital tools have prompted social fears — at what point have we outsourced so much of our brain that it’s no longer our brain? The elderly, already dealing with cognitive decline, may not have as much use for such metaphysical questions. Having the world help you think, through whatever means, may be philosophy enough.