The Invisible Gorilla: And Other Ways Our Intuitions Deceive Us (33 page)

Few studies have even investigated whether training on simple perception and memory tasks has any consequences for our daily mental chores. Although many studies have shown that people who are more cognitively active when they’re younger preserve their abilities better as they age, such studies are correlational.
⁵¹
Thinking about the illusion of cause reminds us that an association between two factors can occur even if neither one causes the other. The only way to study the effects of brain training on daily cognition is to conduct an experiment, randomly assigning some people to training conditions and others to control conditions, and then measuring the results of training. Over the past decade several clinical trials have done just that.

The largest experiment to date started in 1998 and randomly assigned 2,832 seniors to one of four groups: verbal memory training, problem solving, processing speed, or a control group that did no cognitive training.
⁵²
This massive clinical trial, funded by the National Institutes of Health and conducted by researchers from many universities, hospitals, and research institutes, was known as the ACTIVE trial, which stands for “Advanced Cognitive Training for Independent and Vital Elderiy.” In the experiment, each group practiced one particular task for ten sessions of one hour each, spread out over about six weeks, and after the training, their performance was tested both on a set of laboratory tasks and on some real-world tasks. The hope was that training on the cognitive tasks would help to keep the brain sharp, leading to improvements on other cognitive tasks and on real-world functioning.

Not surprisingly, if you practice doing a visual search task for ten hours, you get better at visual search. If you practice a verbal memory task for ten hours, you get better at verbal memory. Many of the participants, particularly for the speed-of-processing training, showed improvements immediately after training, and the improvements lasted for years. However, the improvements were limited to the specific tasks they learned and did not carry over to the non-trained laboratory tasks. Practicing verbal memory buys you almost nothing for your processing speed, and vice versa.

Later followup surveys of participants in the ACTIVE study did
show some evidence for transfer to real-world performance. Participants in the training groups reported fewer problems with daily activities than did people in the no-training control group. Of course, in this case, the participants knew they were in a training group and that they were expected to improve, so some of the self-reported benefits could be due to placebo effects.

Unfortunately, the results of the ACTIVE study are consistent with other studies. Training tends to be specific to the task that is trained. If you play Brain Age, you’ll get better at the specific tasks included in the software, but your new skills won’t transfer to other sorts of tasks. In fact, in the now vast cognitive-training literature, almost none of the studies document any transfer to tasks outside the laboratory, and most show only narrow transfer of skill between laboratory tasks—from the one practiced to those that are very similar.
⁵³
If you want to get better at Sudoku, and especially if you like doing Sudoku, by all means, do more Sudoku. If you think that doing Sudoku will keep your mind sharp and help you avoid misplacing your keys or forgetting to take your medicine, you’re likely succumbing to the illusion of potential. The same goes for solving crossword puzzles, a favorite recommendation of those who believe that mental exercise can keep the brain sharp and stave off dementia and the cognitive effects of aging: Unfortunately, people who do more crosswords decline mentally at the same rate as those who do fewer crosswords.
⁵⁴
Practice improves specific skills, not general abilities.

Please don’t get us wrong. We’re not trying to argue that there is literally no potential for increasing our mental abilities. Our intellectual capacities are never frozen in place. We all have tremendous potential to learn new skills and to improve our abilities. Indeed, neuroscience research is showing that the plasticity of the adult brain—its ability to change in structure in response to training, injury, and other events—is much greater than previously believed. The illusion is that it is
easy
to unlock this potential, that it can be discovered all at once, or that it can
be released with minimal effort. The potential is there, in everyone, to acquire extraordinary mental abilities. Most people, without any training, can remember a list of about seven numbers after hearing it only once. Yet one college student trained himself to be able to remember up to seventy-nine digits.
⁵⁵
His feat was extraordinary, revealing a latent potential for exceptional digit memory, but it took hundreds of hours of practice and training. In principle, most people have the same potential ability, and could do the same thing with enough practice.

Genius is not born fully formed—it takes years to develop, and it follows a predictable trajectory. Mozart’s early compositions were not masterpieces, and Bobby Fischer made plenty of mistakes when he was learning the game of chess. Both likely possessed exceptional talent to develop, but they did not become great without training and practice. And their greatness was limited to the domains they trained in. Training your memory for digits will not help you remember names. However, expertise in a domain does improve many other abilities
within that domain
that were not specifically trained.

A series of classic experiments conducted by the pioneering cognitive psychologists Adriaan de Groot, William Chase, and Herbert Simon demonstrated that chess masters can remember far more than seven items when the items tap into their expertise.
⁵⁶
We repeated their studies ourselves by testing Chris’s friend Patrick Wolff, a grandmaster who had won the U.S. championship twice. We brought Patrick into the lab and showed him a diagram of a chess position from an obscure master game for just five seconds. We then gave him an empty chessboard and a set of pieces and asked him to re-create the position from memory. Remarkably, he could reconstruct the position with nearly 100 percent accuracy even when it contained twenty-five or thirty pieces, far more than the typical seven-item limit for short-term memory.

After watching him perform this feat a few times, we asked him to explain how he did it. He first pointed out that the training of a chess grandmaster doesn’t include practice in setting up chess positions after seeing them for just a few seconds. He said that he was able to quickly make sense of the positions and to combine pieces into groups based on the relationships among them. In essence, by recognizing familiar patterns,
he stuffed not one but several pieces into each of his memory slots. As he became an expert in chess, he developed other skills that help in playing chess well—mental imagery, spatial reasoning, visual memory—all of which contributed to his ability to do this memory task better than other people. However, being an expert in chess did not make him an imagery, reasoning, or memory expert in general. In fact, when the chess positions we showed him had the same number of pieces arranged on the board randomly, his memory was no better than that of a beginner, because his chess expertise and database of patterns were of little help. The same principle applies to the student who stretched his memory span to seventy-nine digits—his new memory capacity was specific to combinations of numbers, so even after several months of training with numbers, he still had a span of only six items when tested with letters.
⁵⁷
In other words, he trained his potential ability to remember numbers, but that training did not transfer to any other skills.

Chess grandmasters can apply their expertise to perform a wide variety of chess tasks extremely well, even if they have never carried out those tasks before. One of the most dramatic examples is blindfold chess. Top players can play an entire game “blindfolded,” without ever looking at the board—they are told (in chess notation) what moves their opponents have made, and they announce the moves they would like to make in reply. Grandmaster-level players can play two or more blindfold games simultaneously, at a high level of skill, even if they’ve never tried this before. The exceptional chess memory and imagery abilities needed to perform this feat accrue more or less automatically as players become experts.

Working with Eliot Hearst (another psychology professor who is also a chess master), Chris conducted a study to measure how much worse chess grandmasters play when they can’t see the board and pieces.
⁵⁸
You might think that they’d make more errors because of the additional memory load of remembering where every piece is. To find out whether this supposition is true, Chris took advantage of a unique chess tournament that has taken place in Monaco every year since 1992. In the tournament, twelve of the world’s top players, including many world championship contenders, play each other twice: once under normal
conditions, and once under blindfold conditions. Since the same players are involved in the normal and blindfold games, any difference in the number of errors must be due to the conditions, not the competitors.

In total, from 1993 to 1998 there were about four hundred regular games and four hundred blindfold games played in the tournament, with each lasting an average of forty-five moves by each player. Chris used a chess-playing program called Fritz, which was recognized as one of the best software chess players in the world, to find all the serious mistakes the humans made. Fritz undoubtedly missed some of the most subtle errors, but larger blunders and significant mistakes were easy for it to catch.

Under normal playing conditions, the grandmasters made an average of two mistakes for every three games. These were major blunders, ones that could have—and often did—cost them a game against top-level opposition. The surprise, though, was that the rate of errors in blindfold chess was virtually the same. The grandmasters had trained their potential so well that they could perform their art without even looking at its elements (look, Ma, no board or pieces!). For those interested in unlocking their potential, that’s good news, of course. The bad news is that they didn’t become chess grandmasters by just listening to the right music or reading the right self-help books. They did it by concentrated study and practice over a period of at least ten years. The brain’s potential is vast, and you can indeed tap into it, but it takes time and effort.

Practicing games like chess will enhance your ability to do chess-related tasks, but the transfer is relatively limited. Advocates for adding chess to school curricula argue that “chess makes you smarter,” but there is no solid evidence for this claim from large, properly controlled experiments.
⁵⁹
Is there any evidence for broad transfer of skill to tasks and domains other than the one you practice?

Cognitive psychologists were jarred into rethinking the limits of transfer by a striking set of experiments published in 2003 by Shawn
Green and Daphne Bavelier of the University of Rochester.
⁶⁰
The central conclusion of these studies was that playing video games can improve your ability on a variety of basic cognitive tasks that are, at least on their surface, unrelated to the video games you play. Their first four experiments showed that expert video-game players, defined as people who had played at least four hours per week for the past six months, outperformed video-game novices on tests of some attention and perception abilities. Although this sort of comparison is interesting and provocative, as we discussed in
Chapter 5
, an association alone does not support a causal inference. It is quite possible that only people with superior abilities in attention and perception become video-game addicts, or that other differences between the experts and novices might contribute to the differences in cognitive performance. Dan’s colleague Walter Boot, a psychology professor at Florida State University, suggests one such factor: “People who are able to handle college while also spending a lot of time playing video games are different from people who need to spend more of their time studying.”
⁶¹
The only way to avoid such confounding factors and determine for sure whether playing video games improves attention and perception is to give novice players video-game training and then see whether their cognitive abilities have improved.

Green and Bavelier did exactly that in their final experiment. They recruited novice video-game players, defined as people who had spent little or no time playing video games in the past six months, and randomly assigned these subjects to one of two groups. One group spent one hour a day for ten days playing Medal of Honor, a fast-paced “first-person shooter” game in which players view and monitor their surroundings as if they were looking through the eyes of their character in the game’s world. A second group played the two-dimensional puzzle game Tetris for the same amount of time. Before this practice, each completed a battery of basic cognition, perception, and attention tasks, and after training, they repeated the same battery of tasks. For example, in one of the tasks, known as
Useful Field of View
, a simple object appeared for just a fraction of a second right where the subject was looking, and
subjects made a judgment about it (such as whether it was a car or a truck). At the same moment, another object appeared at some distance from where they were looking, and they had to determine where the peripheral object had appeared. The task measures how well people can focus attention on a central object while still devoting some attention to their periphery.

Green and Bavelier hypothesized that action video games would lead to better performance on this task because people have to focus on a wide field of view to do well in the games. In contrast, Tetris should not be of as much benefit because it doesn’t require players to distribute their attention as broadly. Their results confirmed their prediction: Subjects who practiced Medal of Honor showed dramatic improvement on a number of attention and perception tasks, but the Tetris group showed no improvement at all. Following training on Medal of Honor, subjects were more than twice as accurate in the field-of-view task as they had been before training. Before training, they correctly reported the location of about 25 percent of the peripheral targets, but after training they got more than 50 percent right.