Saturday, May 19, 2007

Can running improve your memory?

In honor of the Bay to Breakers run tomorrow, today's posting is about running and possible effects on learning. A recent issue of Neurobiology of Learning and Memory contains an article entitled "High impact running improves learning," by Winter, Breitenstein, Mooren, Voelker, Fobker, Lechtermann, Krueger, Fromme, Korsukewitz, Floel, and Knecht (phew!) and conducted at the University of Muenster.

In a nutshell, the idea was to compare the efficacy of learning new materials depending on the level of physical exertion just prior to learning. In a "relaxed" condition, the runners simply were sedentary for 15 minutes before learning. In a "moderate condition", low impact running was done, defined as running for 40 minutes at a fixed individual heart rate (designed to keep energy consumption aerobic). In an "intense" condition, high impact running was done, consisting of two sprints of increasingly fast pace "until exhaustion" (designed to trigger anaerobic energy consumption).

The task after these three conditions (counterbalanced within subjects) was an associative learning task between an object and a novel word, e.g., learning that a picture of ein auto corresponds to /glump/ in a new, artificial language. The learning task was rather implicit; a picture and a non-word would be presented, and the subjects had to decide quickly if the pairing was "correct" or not. (10 times out of 11, it was.) Immediately after, as well as one week and 8 months later, subjects' ability to identify correct pairings of the new words and their German "translations" was tested.

So, does either low- or high-impact running make it easier to learn new words? If you mine the data enough, it sure does. The comparison I was most interested in - whether memory of the words would be better after either running condition at either 1 week or 8 months post-learning - was not too compelling. There was not even a trend for a main effect of condition at either test-time. Fortunately, an "exploratory comparison" between the low- and high-impact running at the one-week test resulted in a significant different between the two (p = .03). I guess we're supposed to overlook the multiple, unmotivated statistical tests that accompany this one "significant" difference, and the fact that memory in the sedentary condition lies between the two running conditions (Figure 2B, shown below).

Interestingly, the data from the learning part of the experiment (Figure 2A) do seem at first to support the authors' claim that high impact running improves learning - the percent correct for the high impact condition climbs more steeply than the other conditions. Responses in that condition were also faster.

But wait a second, what was going on during training again? The subjects were to decide intuitively whether the picture and word were the correct match. 10 times out of 11, this was the case. So in other words, if you can just manage to hit yes on every trial, you automatically would get 90 percent correct. Maybe if you were, oh say, pumped up on epinephrine, might you answer a little more quickly and perhaps have a different response bias?

The same concern may apply to the retention data as well. The authors never tell us what the ratio of correct to incorrect pairs were for this part of the test, and as with the training phase, only report results as percent correct. Got signal detection analysis? (Evidently not.) And I won't even get started on the fact that the pseudowords and concepts to be remembered, although balanced in important variables, do not seem to have been counterbalanced across the three conditions.

Those who are interested in the physiology, and less skeptical than me about the basic learning effects, might be interested in checking out the analyses of lactate, catecholamines (dopamine, epinephrine, and norepinephrin), and brain-derived neurotrophic factor (BDNF), which were taken at baseline, post-intervention, and post-learning. The authors have interesting stories, if not entirely compelling ones, for all the effects.

Oh well, maybe running won't improve our memories for new things. It's unclear whether it's best to remember all the sights from Bay to Breakers in any case.


Winter, B., Breitenstein, C., Mooren, F. C., Voelker, K., Fobker, M., Lechtermann, A., Krueger, K., Frommer, A., Korsukewitz, C., Floel, A., & Knecht, S. (2007). High impact running improves learning. Neurobiology of Learning and Memory, 87, 597-609.

1 comment:

beartoes said...

I have read another post on this same topic. Running to improve memory, which I find still amazing. There are really smart people out there in the world. Now only if we can learn and use running to improve short term memory in children we will be all set. Heck getting them to get out of bed is a feat, let alone to run to learn. Will have to work on that.