Discuss as:

Big brains for video games

Erickson et al. / Cerebral Cortex
A cutaway image shows four brain structures studied in the video-game research.
the caudate nucleus (blue), putamen (red), nucleus accumbens (orange spot) and
hippocampus (green). Researchers found linkages between game performance
and the first three structures, but no linkage involving the hippocampus.

Does playing video games improve your brain? Or do bigger brains make it easier to learn video games?

Psychologists say they can predict how well you'll do on a video game by looking at the size of just three little structures inside your brain. If those structures are bigger, you'll probably catch on more quickly and do better.

But don't start bragging about how gamers are naturally brainier just yet. The psychologists have more puzzles to solve before they level up.

"We're really at the tip of the iceberg in understanding how all this gets put together," said the University of Pittsburgh's Kirk Erickson, the study's principal author.

The study, conducted by Erickson and 10 other researchers, appears in the journal Cerebral Cortex. "This is the first time that we've been able to take a real-world task like a video game and show that the size of specific brain regions is predictive of performance and learning rates on this video game," Erickson said in a news release issued today.

The team's findings come amid an increasing reliance on game-based "brain training" as a form of mental calisthenics. "Video games are now being used frequently through educational disciplines to train children and adults, in remedial situations, in business practices to train employees, and they're even being used by the military," Erickson told me.

Past research has shown that expert gamers tend to outperform novices on basic measures of attention and perception. Some studies have suggested that video-game training can help novices bridge the gap - while others indicated that the novices couldn't catch up after more than 20 hours of training.

Inside the learning machine
Erickson and his colleagues wondered whether physical characteristics in the brain were behind the variability in learning rates. The researchers focused on a region deep within the cerebral cortex, known as the striatum.

"Our animal work has shown that the striatum is a kind of learning machine - it becomes active during habit formation and skill acquisition," one of the study's co-principal investigators, Ann Graybiel of the Massachusetts Institute of Technology, said in the news release. "So it made a lot of sense to explore whether the striatum might also be related to the ability to learn in humans."

Thirty-nine experimental subjects, ages 18 to 28, were recruited at the University of Illinois. Ten of them were male, 29 of them were female, and none of them had played video games for more than three hours a week over the preceding two years. Each subject's brain was measured using high-resolution MRI scans to map the relative size of structures in the striatum.


  The video game Space Fortress can be manipulated to test various aspects of cognition.

Then each subject received 20 hours of training to play a video game specifically created for research purposes, called Space Fortress. It's basically an Asteroids-type arcade game, in which the object is to knock down and destroy an enemy fortress while dodging space mines. However, the game has lots of extra twists that require close attention. Some of the players were told to focus exclusively on running up a high score, while others were told to shift their priorities between several goals.

The result? The subjects who had more volume in an area called the nucleus accumbens did significantly better in the early stages of training. Meanwhile, those who were well-endowed in different areas of the striatum, known as the caudate nucleus and putamen, handled the shifting strategies better.

"These are people who had healthy brains," Erickson said. "These aren't learning-disabled people. But we were still able to distinguish essentially who would be more affected by the training in this video game."

What the brain areas do
The linkages between the areas of the brain and their effects made sense, based on what neuroscientists have previously learned. For example, the nucleus accumbens has been linked to the brain's emotional response to reward and punishment. That would explain the pleasures and frustrations that gamers feel during the early stages of learning a game.

"The putamen and the caudate have been implicated in learning procedures, learning new skills, and those nuclei predicted learning throughout the 20-hour period," said the University of Illinois' Arthur Kramer, another co-principal investigator.

Arizona State University linguist James Gee, who has studied the linkage between video games and learning but was not involved in the latest study, said in an e-mail that the researchers made "an interesting set of claims." However, he noted that the study involved "an old-style arcade-type game that is not typical of today's games in many respects."

He also noted that other research, going back to the 1980s, has shown that "a short training period made differences between men and women on spatial skills in video games disappear." That runs counter to the claim that there's an unbridgeable gap between good and bad gamers.

Exercise and the brain
Even the researchers behind the latest study stress that brain structures aren't set in stone. "We know that's not true for a lot of nuclei in the brain," Kramer said. "We know that exercise can increase the volume of the nuclei."

He could be talking about physical as well as mental exercise. An earlier study conducted by Kramer and Erickson, for example, found a linkage between physical fitness and memory function in elderly people. A more recent study found that game-playing enhanced brain function. And it could be that the folks with brains suited for gaming got that way by pursuing other activities, such as musical training or even juggling.

Erickson said the study could have an impact on how brain-training services are offered - depending on how much value people place on measuring brain volume. "In the future, we might be able to tailor the training regimens based on pre-existing differences in the brain," he told me. "Some people might be slower at reaching the same level, so that means they might need more time to learn the same process."

He said the findings also could suggest regions to focus on when treating mental disabilities - or regions to watch when looking for the effects of different training strategies.

"What we don't know yet is the extent to which these brain regions can actually change as the result of more experience or training," Erickson said. "It's possible that the more activity you engage in, the more likely you are to change the volume of these structures."

Researchers are just beginning to get a handle on how different regions of the brain interact to foster learning - which brings us back to the tip of the iceberg.

"The fact that we could explain more than 20 percent of the variance in learning rates by measuring the volume of only two or three brain regions is actually quite impressive," Erickson said. "There must be several other brain regions contributing to performance in learning. These other regions are things that other studies will have to track down."

Does all this hint at a new age of neuroscience, or a newfangled revival of phrenology? Feel free to weigh in with your comments below.

Update for 1:30 p.m. ET: Arizona State University's James Gee has already weighed in with additional comments via e-mail:

The paper is interesting indeed, though problematic in certain respects.  Of course, the nature of people's brains affects their learning and performances — this is just another way of talking about individual differences.  So, too, does the nature of people's prior experiences in the world.  And, of course, these two (internal structure and external experiences) interact in complex ways — one way they interact is that sometimes it is hard to tell whether a set of experiences led to a brain difference (e.g., more reading leading to changes in the brain) or a brain difference led to people wanting, getting, and being better at some experiences (e.g., seeking out reading and liking it more and being better at it).

The authors try to control for experience by taking people who have played video games for less than 3 hours a week in the prior two years.  However, this still allows the possibility that some of these people had played arcade-style video games massively when younger (it is not uncommon that college students — especially ones as old as 28, which are included in the sample — played games a good deal before they went to college, but play less in college).  So, without controlling for this factor, we cannot be sure the result is not due to some of these students having played a good deal more video games than others prior to two years and this may have affected their brains.

This would be like doing a reading study on working adults and taking those who have not read many books in the last two years.  Many adults do not read a lot, but some may have been big readers earlier in life, before the demands of work and family, and others may have been non-readers.  The sample would not be a sample of true reading "novices," and we do not know that the sample of gamers in this study are truly game "novices" for the same reason.

The game is not typical even of arcade games — it is a made-up game for research involving a lot of complexity and elements that certainly sound frustrating and un-gamelike — e.g., lots of symbols to pay attention to.  It would certainly be important to know how much different subjects like the game or were motivated to play it in any way that they were paying avid attention and caring enough to learn anything.  Otherwise, the study may just be identifying subjects that were motivated by a rather strange game.

In any case, the study tells us little about the sorts of modern games — like Civilization — that educators have argued hold out great promise for deep learning and school reform.  The game in the study stresses hand-eye and motor coordination — admittedly important for some purposes — but games that want to recruit a larger (and older than teen) player base downplay hand-eye coordination in the favor of strategy.

The most interesting result in the paper — and an important one — is the result that shows the importance and efficacy of variable-priority learning, that is, learning that trains and stresses cognitive flexibility, rather than fixed-priority learning.  But, of course, cognitive scientists already know this from other forms of research.

More on brains and games:

In addition to Erickson, Graybiel and Kramer, the authors of the study appearing in Cerebral Cortex include Walter Boot of Florida State University; and Chandramallika Basak, Mark Neider, Ruchika Prakash, Michelle Voss, Daniel Simons, Monica Fabiani and Gabriele Gratton of the Beckman Institute for Advanced Science and Technology, University of Illinois. The study was funded by the Office of Naval Research.

Join the Cosmic Log team by signing up as my Facebook friend or following b0yle on Twitter. And pick up a copy of my new book, "The Case for Pluto." If you're partial to the planetary underdogs, you'll be pleased to know that I've set up a Facebook fan page for "The Case for Pluto."