Posted by: wolf | March 5, 2008

Brains, video games, and experimental design & analysis
At the beginning of the 21st century, does anyone anywhere doubt that some region in their brain is active when they are doing things?

I have to admit that as a psychologist I am really envious of all the public and scientific interest usually attracted by neuroscience results, compared to the mediocre kind of reception of most of the down-to-earth psychological experiments not involving any huge, costly high-tech brain scanning devices. But even if you discount my biasedness, there are some neuroscience studies that have, well, a kind of “duh!” flavor to them, which becomes even more pronounced in the popular media reports of such studies, like in the following excerpt from a CBCnews article:

Men are more rewarded by video games than women on a neural level, which explains why they’re more likely to become addicted to them…

Well, similar to the discussion of that article at the cognition and language lab one could paraphrase this as saying “men are more rewarded by video games because men are more rewarded by video games” — or do you really believe that reward is somehow distinct from your brain? Again the discussion over there of linguistic confusion when reporting neuroscientific results inspired me to post, but today, I put forth a little more effort and looked up the paper in question. And after reading, I think there’s a little bit more to say about that paper in addition to the comment that it is somehow obvious that, well, something will happen in your brain any time you are doing something (or even doing nothing), so something will also happen when you are playing a video game.

The starting point of Hoeft et al.’s paper is that men are more likely to get addicted to video games than women. Also, it has been shown that playing video games is correlated with activity in the brain’s reward system. So far, so good, even if the envious psychologist inside of me cries out: “isn’t it kind of obvious that the reward system is active when you do something that feels good?” What Hoeft et al. did in their own study was to observe 11 men and 11 women with functional magnetic resonance images (a method to study brain activity in order to locate brain regions concerned with specific tasks) while they played a simple video game (they had to click on “balls” on a black-and-white screen). What the authors present as the main result is that they found a greater activation in several brain regions for men than women during game play, specifically in the right nucleus accumbens, the orbitofrontal cortex, and the amygdala — regions that have been identified by other researchers as dealing with experiencing and predicting reward.

The conclusion the authors draw is that men are more likely to become addicted to video games than women because their reward systems are more active.

At first sight, no serious flaw. But — and make it a big BUT here — at second sight there is a real methodological flaw in this study, a flaw (the envious psychologist is taking over again) that probably would not have gone unnoticed in most psychological journals: the first paragraph of the results section and figure 1 show that men performed significantly better in the video game; they achieved higher scores in total and their performance increased more during the session compared to women. The authors themselves even report in section 3.4 (“Gender specific covariance of behavior and brain activation profiles”) that

“Males showed positive correlation between goal achievement and Game > Control brain activation” (p. 255)

with what seems to be a rather large effect — but I have to admit I don’t know what they mean with

“contrast estimates, i.e., effect size calculated as the linear combination of beta parameters” (p. 255)

so maybe I have the wrong “rho” in mind when they state the effect size as “rho = 0.61”. So they must have noticed that men did not only show a greater activation in brain regions of the reward system, but that men also performed better than women and there is a correlation between performance and activity of the reward system.

Now let me ask: Don’t you think that success in the video game is responsible for the differences in activity of the brain regions of the reward system instead of gender?

That’s a typical example of experimental confounding — the authors failed to separate gender effects from the effects of success in the video game. Of course, it is a more or less interesting result on its own that men performed better. But what is a no-go in experimental analysis is to ignore a potential confounding variable like success in the video game and draw conclusions about gender effects. That is really a serious flaw. Imagine someone tested the intelligence of females at the age of three and males at the age of six and concluded: “male children are more intelligent than female children” — that would obviously a forbidden conclusion, because age is confounded with gender. Hoeft et al. , however, ignored the confounding variable of performance.

HOEFT, F., WATSON, C., KESLER, S., BETTINGER, K., REISS, A. (2008). Gender differences in the mesocorticolimbic system during computer game-play. Journal of Psychiatric Research, 42(4), 253-258. DOI: 10.1016/j.jpsychires.2007.11.010

P.S. At least they controlled computer and video game experience when selecting their participants.


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s


%d bloggers like this: