Brain Scans Are Sexy, But What Do They Really Reveal?

| 1/15/2009 3:01:53 PM

Tags: Science and Technology, politics, psychology, brains, neuroscience, brain scans, neurolaw, New Scientist,

brain MRI

Neuroimaging grabs headlines, but a recent study, highlighted in the New Scientist, questions the reliability of brain scan research, particularly when it’s used to make claims about human emotions and behavior.

Hal Pashler and his colleagues looked at more than 50 studies that used fMRI scans to link activity in specific brain regions to feelings. They argue that many of the studies—nearly 30—have inflated these correlations or created one where none exists. The problem has to do with methodology. Pashler’s team contends that for any given brain image, researchers should cross-reference two sets of scans in order to accurately judge the strength of a correlation. The studies they criticized relied on only one.

Not surprisingly, the scrutinized groups have already begun to defend themselves, but there’s more than scientific integrity on the line. Studies like the ones in question are already being treated outside scientific circles as fact. As both the New York Times and Justice Talking (pdf) reported, the scans been used as evidence in legal cases for years.

Image by Mikey G. Ottawa, licensed under Creative Commons.

lance winslow
1/22/2009 2:06:35 AM

Interesting thoughts here, I tend to agree that there is a bit of "scientific reaching" in eeg and fMRI scans. Still, we need to push this science forward to fully understand how and why the brain works so well, it's a cool device indeed.

miranda trimmier
1/16/2009 1:45:18 PM

And another related link, thanks to the American Journal of Bioethics. An article suggesting that brain imaging could eventually be used to track kids academically. There's not a ton of elaboration (it sounds like this is mostly speculation at this point), but it's still food for thought in the continuing discussion:

miranda trimmier
1/16/2009 1:30:05 PM

Here's a link to the paper, provocatively titled "Voodoo Correlations in Social Neuroscience":