Friday, April 16, 2004

Short-Term Working Memory and Brain Structure

Here's an intriguing story that's currently on Nature Science Update (found via Slashdot). It appears that some researchers are claiming to have localized the part of the brain responsible for short-term working memory, an important factor in performance on IQ tests like the Wechsler Adult Intelligence Scale (WAIS).

The number of things you can hold in your mind at once has been traced to one penny-sized part of the brain.

The finding surprises researchers who assumed this aspect of our intelligence would be distributed over many parts of the brain. Instead, the area appears to form a bottleneck that might limit our cognitive abilities, researchers say.

"This is a striking discovery," says John Duncan, an intelligence researcher at the Cognition and Brain Sciences Unit in Cambridge, UK.

Most people can hold three or four things in their minds at once when given a quick glimpse of an image such as a collection of coloured dots, or lines in different orientations. If shown a similar image a second later, they will be able to recognise whether three or four of these spots and lines are identical to the first set or not.

But some people can only catch one or two things in a glance, while others can capture up to five.

This very short-term memory capacity is thought to be related to intelligence. In the same way that a computer with a larger working memory can crank through problems more quickly, people with a greater capacity for holding images in their heads are expected to have better reasoning and problem-solving skills.

A person's working memory capacity can be determined using simple psychological tests. But now two teams of researchers report in Nature that they can see it in brain scans too.


Both teams concluded that everything depended on the same tiny spot in the posterior parietal cortex.

Putting aside the question of just what IQ really measures, there's a whole lot more to it than the mere ability to store prodigious quantities of material in short-term memory. Still, that one aspect of IQ should be linked to activity in a specific area of the brain isn't particularly surprising, at least to me. Also worth noting is that neither of the two teams claims to have found a relationship between the size of an individual's posterior parietal cortex and the quantity of material the individual could hold in working memory, which is just the sort of simplistic scenario practitioners of pseudoscience in the Philipe Rushton/Richard Lynn mode would have postulated to hold.

As a personal aside, I had occasion to take the WAIS a few months ago, and made a perfect score on the forward digit-span subscale, correctly recalling 15 consecutive numbers spoken to me after hearing them just once. I wasn't quite so strong on the reverse recollection bit - I could recall 14 consecutive numbers backwards just fine, but when it got to 15 I swapped two adjacent numbers in the sequence. The results weren't stunning to me, as I've known since childhood that I have a freakish (though not photographic) working memory, but it was certainly amusing to see the look on the test administrator's face as I kept long past where nearly everyone would have stumbled ;-) (For the record, the maximum number of digits most adults can recite correctly is 7±2, which is why telephone numbers are rarely more than 9 digits long, and usually grouped into subunits of 3 to 4 even so 6 forwards and 5 backwards - according to the WAIS administrators' guidebook, at any rate. The 7±2 bit appears to be little more than conventional wisdom based on a single flawed study dating back to 1956.)