Training Working Memory: Bad News, and Surprising Great News

Regular readers of this blog know that I’m very skeptical about training working memory. Despite all the promises, most studies show that WM training just doesn’t do very much.

working memory training

Better said: working memory training helps people do better on other, similar working memory tests. But it doesn’t help students learn to read or calculate or analyze any better.

(Earlier posts on this topic here and here.)

But here’s a tantalizing possibility: what if we could find an even better shortcut to cognitive success?

Training Working Memory: News from Finland

Researchers at Abo Akademi University in Turku wondered why WM training works in psychology labs, but not in classrooms.

(One of the champions of WM training — Dr. Susanne Jaeggi — has spoken at Learning and the Brain conferences. If you’ve seen her, you know she’s an incredibly impressive researcher. You too might reasonably wonder why that research isn’t panning out.)

These Finnish researchers wondered if the WM training simply gave students the chance to figure out a particular WM strategy.

That is: they didn’t have more working memory. But, they were using the WM they already had more strategically.

This strategy applied to the specific working memory task (which is why their WM scores seemed to get better), but doesn’t apply to other cognitive work (like math and reading).

If that hypothesis is true, then we could simply tell our students that strategy. We would then see the same pattern of WM development that comes from the training — only much faster.

Specifically, we would expect to see improvement in similar WM tasks — where students could apply the same strategy — but not on unrelated tasks — where that strategy doesn’t help.

If their hypothesis is correct, then the results that take 6 WEEKS of training might be available in 30 MINUTES. Rather than have students figure out the strategy on their own (the slow, 6 week version), we can simply tell them the strategy and let them practice (the 30 minute version).

The Test, the Results

The Finnish researchers worked with three groups of adults.

Control group #1 did a WM test on Monday and a WM test on Friday. They got no practice; they got no training.

Control group #2 also did WM tests on Monday and Friday. In between, they got to practice a WM task for 30 minutes. This is a mini-version of the WM training model. (If they had gotten the full six weeks, they might have figured out the strategy on their own.)

The study group — lucky devils — were TOLD a strategy to use during their practice session. (More on this strategy below.)

What did the researchers find?

First: As they predicted, the group that was told the strategy made rapid progress, but the other two groups didn’t.

Control group #1 didn’t make progress because they didn’t even get to practice. Control group #2 did practice…but they didn’t have enough time to figure out the strategy.

Only the study group made progress because only they knew the strategy.

Second: As researchers predicted, the group that learned the strategy didn’t get better at WM tasks unrelated to the strategy they learned.

In other words: the group given a strategy behaved just like earlier groups who had discovered that strategy for themselves during 6 weeks of practice. They did better at related WM tasks, but not at unrelated tasks.

We don’t need 6 weeks to get those results. We can get them in 30 minutes.

What, exactly, is this magical strategy?

The precise strategy depends on the working memory exercise being tested.

In general, the answer is: visualize the data in patterns. If you’ve visualized the pattern correctly, you can more easily perform the assigned WM task.

You can check out page 10 of this PDF; you’ll see right away what the strategy is, and why it helps solve some WM problems. You’ll also see why it doesn’t particularly help with other WM tasks — like, for example, understanding similes or multiplying exponents.

Training Working Memory: Classroom Implications

This research suggests that we shouldn’t train students’ general WM capacity, because we can’t. Instead, we should find specific WM strategies that most resemble the cognitive activity we want our students to do.

Those strategies allow students to use the WM they have more effectively. With the same WM capacity, they can accomplish more WM work.

The key question is: what WM strategies are most like school tasks?

We don’t yet know the answer to that question. (I’ve reached out to the lead author to see if she has thoughts on the matter.)

I do have a suspicion, and here it is: perhaps the practice that we’re already doing is the best kind. That is: maybe the working memory exercise that’s most like subtraction is subtraction. The working memory exercise most like reading is reading.

If I’m right, then we don’t need to devise fancy new WM exercises. The great news just might be: the very best WM exercise already exists, and it’s called “school.”

Leave a Reply

Your email address will not be published. Required fields are marked *