Balancing Brains

Researchers are learning new things about the human brain by studying how astronauts regain their balance.

Balancing is not as easy as it seems - just try to stand on one foot for a full minute, and you'll get a sense of the constant effort involved.

It's one of those complex skills like reading that becomes so automatic with practice, we simply forget how tricky they were to learn. And, like reading, you might suppose it would take something extraordinary to make you forget.

Indeed it does. Like travelling to space.

Researchers have found that astronauts who return from a space voyage can still balance, but they find it far more difficult. That's because, explains neuroscientist Bill Paloski, their brains are no longer sure how to interpret the information that comes from their senses.

When you balance, he says, you use information from as many as three sources: the proprioceptive sensors in your muscles, which tell you where your body parts are in relationship to each other, the vestibular system in your inner ear, which tracks the position of your head in space, and of course your eyes.

The brain deals with all that information by building "a model." Computer programmers might call it a mental subroutine, but it's more than an algorithm. Models provide context for interpreting and reacting to sensory data. The brain generates such models all the time - it's the way we learn and adapt. We do it on Earth, say, when we learn a new language, or even when we get accustomed to new prescription glasses.

Astronauts do it, too. On Earth, their brains have already constructed a model that tells them how to manage their bodies in 1-g (normal gravity). In space, they must build a 0-g (weightless) model. Then, back on Earth, they have to figure out that it's time to switch to the 1-g model again.

The transition isn't always easy.

Balancing in space requires new ways of thinking.

When you encounter a completely new context like space, your brain has some work to do. It has to decide whether this will be a persistent context or not - whether it's worth building a model. And if it is, then it has to develop one.