Wednesday, May 10, 2006

EGHM II: what the hell is IQ?   posted by JP @ 5/10/2006 01:55:00 AM

Eco de Geus gave a talk on his search for endophenotypes for IQ. A bit of background: when searching for genetic factors involved in a phenotypic trait, it's easiest to look directly at the trait itself-- that is, if the trait is a disease, grab some sick people and some healthy people and compare the two groups.

A more powerful approach, however, is to look at something closer to "gene action", as they say; look at BMI instead of diabetes or QT interval as opposed to heart disease. It's easy to imagine why this works: instead of diluting your cases with people who have disease for multiple, distinct causes, you're looking only at one of them.

IQ, de Geus claimed, is a complex phenotype that would be amenable to this kind of approach. So he's been looking at various measures to bring IQ down to a more physiological level (using twin studies). But, well, he ain't there yet. Here are some results:

1. Speed of signal transmission in neurons: heritable, but not correlated to IQ

2. Processing speed (as measured by one of the waves in an EEG-- this is obviously not something I know much about): heritable, but not correlated to IQ.

3. Cortical connectivity: heritable, but not correlated to IQ.

4. Brain volume: heritable and correlated to IQ. But he said he didn't like this as an endophenotype because he found "high IQ --> big brain" to be a more plausible explanation for the correlation than "big brain --> high IQ".

5. The one test that showed a little promise was a psychological test of visual inspection time, as measured, if I remember correctly, by the Stroop test. But this was conterintuitive-- the (weak) correlation was negative (those who reacted fastest had lower IQ).

So what to make of this? I, for one, know jack about this area, so don't look at me.