Thursday, October 18, 2012

Science and social policy

We recently blogged about science denial being a trait we can't just attribute to creationists, but one that we scientists share as well.  We said that to a great extent political views can determine how we pick and choose the evidence we believe.  Right-wingers tend to deny climate change but climate change can make left-wingers very anxious; to a right-winger IQ is real and even a genetically determined group characteristic while to a left-winger it's impossible to measure and is environmentally determined.  So in this light, a story at the BBC website on Tuesday was of interest.

Entitled "Childhood adversity affects adult brain and body functions, researchers find," and written by Alok Jha, the piece describes a number of studies presented at the Society of Neuroscience meetings this week in New Orleans.  The papers aren't yet published but the abstracts are online, and here's the first sentence of one example (E. Pakulak, Y. Yamada et al.):
A large and growing literature documents the profound impact of lower socioeconomic status (SES) on cognitive skills and brain structures and functions in children (Hackman, Farah, & Meaney, 2010).
Well, so if you're a believer in IQ being genetically determined and the idea that people earn their socioeconomic status by virtue of their IQ, this isn't right.  That's because you think the cause and effect are the other way around: cognitive skills have a profound impact on where we end up in the SES hierarchy, children inherit their cognitive abilities and therefore they inherit their place in the order of things.  But if you believe that the brain is plastic, and can be affected by experience, you're perfectly ok with how this abstract begins.  In fact, probably you liked the presentation title and that's why you kept reading.

So, say you wanted to sort out which comes first, cognitive ability or SES (and just as importantly, who's right), how would you do it?  Clearly just declaring the order in which you think things happen isn't enough.  A lot of work on neuroplasticity has been published in the last decade or so, much of which is pretty convincing, which may or may not predispose you to believe findings that brains can respond to environment, but even so let's think about it.  Since it's impossible to determine cause and effect just by looking at outcome, what's required is an intervention study.

You'd need to look at the IQ/brain structure/cognitive abilities -- whatever you think the right measure/outcome is -- before and after some kind of intensive training/attention/input.  Anything from a repeat IQ test to determine whether cognitive abilities have changed to a follow-up MRI or fMRI or PET scan to assess changes in brain structure or functioning.  Though, if you did see improved IQ scores, you'd have to worry about whether repeat testing itself is what improved the scores.  And you then have to do the genome sequencing of each individual to be able, in principle, to separate out prior inborn, and later experiential effects, assuming inborn factors are identifiable.  Clearly, these kinds of studies have to be carefully planned and interpreted, at best. 

Judging from the abstract, the study described by Pakulak et al. was an intervention.  They gave adults from a number of SES backgrounds a battery of memory and language proficiency tests and found that childhood SES was strongly predictive of working memory, language proficiency and attention span.  But, see above as to cause and effect.  Then they did an intervention with parents of lower SES children, after which they measured "attention/executive function" in this group and in controls and found attention improved. They conclude that cognitive abilities are malleable and that neuroplasticity extends into adulthood. They did not test the role of genetics.

At the same meeting, Suzanne Houston reported that
...the size of different parts of the brain could be affected by growing up in different homes. "We found higher parent education, smaller amygdala. The higher the income, the larger the hippocampus."
Her interest is in determining which environmental factors affect brain growth.   Others reported that excessive stress or abuse in childhood affects the functioning of the brain and is associated with ill health in adulthood. And so on. 

Does all this work resolve this debate?  Of course not.  If you think IQ is fixed at birth, you wonder first whether the executive function Pakulak et al. measured improved, and if not why not.  That's not addressed in their  abstract. Does something other than improved cognition explain improved attention?  Say, the desire to please the researchers?  You might wonder whether the sample size (72) was large enough.  A researcher can't do the experiment of having a child grow up in one home and then another and measure the difference in the brain.  Nor adequately control for confounders in studies of, say, stress and adult morbidity and mortality. So, you might think this is all yet more evidence that we should stop pouring public money into programs based on the principle of equal opportunity for all when they clearly can't work for everyone. 

But, if you already like the idea of neuroplasticity, and believe that society has a responsibility to let everyone live up to his or her potential and that everyone shares the same potential, you think this kind of study is yet more evidence of neuroplasticity, that social programs can improve the lot of those in the lower SES and that tax money should be spent equally on everyone.

In all of this, we also know that uterine experience can affect growth, development, gene expression and physiological states, and these themselves can be inherited.  Whether this applies to traits like IQ is an open question today, as far as we know, and indeed gene expression would have to be tested for each relevant tissue, and whether this was only relevant during gestation or remains so during life. The point here is just that uterine experience adds another potentially major source of variables that would have to be measured.

It's much easier to pick apart a study we disagree with than one we like, even if we don't recognize that's what's going on.  We all have our subtle biases and may not even be aware of them.  And then there are the not-so-subtle biases.  And, since there's no such thing as a perfect experiment, any study can be criticized.  Rarely do we learn, when we're taught the scientific method, how much our evaluation of the evidence depends on what we already believe rather than how well the experiment is done. 

No comments: