Jump to content


Getting an "Insecure Connection" warning for Exisle? No worry

Details in this thread

Something Fishy about fMRI


  • Please log in to reply
No replies to this topic

#1 Orpheus

Orpheus

    I'm not the boss of you!

  • Administrator
  • 17,759 posts

Posted 06 October 2012 - 12:38 AM

Functional Magnetic Resonance Imaging (fMRI) is probable the main source of most of the colorful "brain activity scans" you've seen in the lay literature in recent years under stunning headlines that probably inspire skepticism in most of you regular EtUers, and are often either dismissed by further study, or (more often) quietly disappear under the waves -- to be cited, unrefuted, by all sorts of people for many years to come. Why unrefuted? Because confirmational studies aren't sexy or easily funded, labs that use fMRI aren't in a rush to discredit the technique, and negative findings are always harder to publish than affirmative findings.

One of this year's IgNobel Prizes went to a study that did fMRIs on a dead salmon as it was shown photos of inclusion and exclusion in human social settings. The soon-to-be-dinner did predictably poorly on distinguishing between the two (perhaps explaining their notorious inability to work the room at faculty dinners), but bits of the brain did "light up", as if the dead fish were thinking about it!

Now the study was really a good one, when seen as an abject lesson in proper statistical analysis of the data from fMRI voxels, which fully 40% of studies at that time failed on the most basic level, but I've always been skeptical of the test itself (at least as it is commonly used). It relies on too many assumptions, like "differential blood oxygen flow is a sensitive indicator of brain activity" or even "oxygen consumption is a major indicator of the important processing steps" [Perhaps the results of certain fixed core "computations" are more important than the location of changing ones]

Though it is completely acknowledged that there is a variable (several second) delay between a change in metabolic activity and its appearance in fMRI, the exact timing in various parts of the brain under various conditions hasn't been well studied. Researchers counter this by running their tests slowly, but what if the crucial determinations are made very quickly [wash or dry clean] leading to much less important differences in persistent secondary processing [which dry cleaner is most convenient to my errands this week]. This likely varies from task to task, and there may be an innumerable number of tasks. It may very easily matter that the "color" of one part of the scan reflects an mere flicker from a quick, critical decision  2 secs ago, while another 'brighter' part of the scan is dwelling on a slower, less important thought 6 sec ago.

Overall, I think you'll find that there's a LOT more said about fMRI's convenience and noninvasiveness than its proven validity, yet brain activity techniques are already accepted by courts in some countries despite the objections of their own scientists, and attempts are underway to have fMRI accepted in American courts as a lie detector, in child abuse cases, and more.

I'm sure we'll have a lot more fMRI articles to discuss in this thread soon enough.


0 user(s) are browsing this forum

0 members, 0 guests, 0 anonymous users