Monday, June 23, 2008

This is your brain on nonsense

Remember that idiotic NYT editorial from November last year that claimed fMRI scans of swing voters' brains could be used to read their minds? Well, the same company that pulled off that stunt is back with an equally egregious piece in The Atlantic. Slate's suitably skeptical,
The neuropundits are more like tarot readers than scientists: They claim to read specific mental states from patterns of blood flow and brain activity, but the narratives they invent are arbitrary, equivocal, and inconsistent. And whenever the imaging data happen to contradict reality, they change their interpretation without a second thought.
Indeed. Another example of cognitive neuroscience gone badly awry...

7 comments:

  1. Steady on, dude. The piece in the Atlantic is never fully serious itself, and it isn't actually by the guys who did the earlier swing votes thing. And the whinge-job in Slate is a poor ally - the image it presents of how brain scans are interpreted is worse than caricature, and endorsing Brooks' gag that "[brain-imaging] is like flying over Los Angeles at night, looking at the lights in the houses and trying to guess what people are talking about at dinner" condemns whole tracts of the best science of the past 15 years or so to the trash on the basis of plain awful reasons.

    ReplyDelete
  2. Perhaps the dumbest, most ill-informed and falsest claim in the Slate piece is the one that claims relating brain activity to psychological states are "arbitrary".

    ReplyDelete
  3. The Atlantic article was written by Jeffrey Goldberg, but the analysis was done by FKF Applied Research and UCLA's Marco Iacoboni, exactly the fools who did "This is your brain on politics" bollocks. (See http://www.nytimes.com/2007/11/11/opinion/11freedman.html). While Iacoboni is a proper scientist, but he's been notorious for running stupid 'studies' to support FKF (see: http://www.mindhacks.com/blog/2007/11/election_brain_scan_.html and http://www.nature.com/nature/journal/v450/n7169/full/450457a.html).

    While I agree the Atlantic article is light-hearted, it DOES present the findings as SCIENCE, which they most certainly are not. The fact that, say, the amygdala lights up when you view X tells you almost nothing.

    Note that the Slate article does not dismiss cognitive neuroscience as a WHOLE, it attacks particular example of it. And it's not that brain activity doesn't cause the mind, it's that the neuropundits create arbitrary narratives linking the two.

    ReplyDelete
  4. "Almost nothing"? Hell no, not by a long way. Here's a tiny part of why. We know that the Amygdala is one of the many brain areas that gets its own set of retinal data, independently of the visual cortex. We know that Amygdala is preferentially activated for various emotionally salient visual stimuli. (And that it ignores some too.) We know that part of how this is cognitively effective is that there are monosynaptic projections from the Amygdala to a series of stages of the visual cortex, which modulate processing there. Also, we have no idea at all how much of this background knowledge might have informed the remarks of the people who let Goldberg have his ride, since he doesn't report it. I agree that they're stirring up hype, I disagree that you can draw good inferences about how bad they are from a second hand and incomplete report, or by using arguments so sweeping that they damn whole branches of science.

    ReplyDelete
  5. I was unclear: I meant that the amygdala lighting up when someone sees X tells you relatively little about that person's specific mental states. That is, you can't conclude "oooo, you fear Jimmy Carter" because Goldberg's amygdala lit up. I honestly think I'm on the side of scientific consensus here: fMRI tells you a lot of cool stuff, it doesn't allow you to read someone's mind. (See the Nature paper I emailed you).

    I'm not concluding they're bad from this single instance - they've done crappy studies like these for years (see the Mind Hacks link above), and have received a lot of flak for it. Including, I might add, from a Nature editorial I found out about because you posted it on Facebook...

    ReplyDelete
  6. I think we're agreed on most substantive issues. FKF are hype-meisters, peddling fairly shallow stuff. I'm not even sure we differ that much over whether you can responsibly justify claims about what someone actually does from the second hand report of a non-expert with a clear interest in readability over accuracy, etc. That said, brain measuring of a variety of sorts is getting closer to mind-reading all the time, for the reason that there are many, many regular relationships between psychological state and brain state, and we're getting to know more and more about them all the time.

    ReplyDelete
  7. Yeah, we mostly agree. I take it you're more enthusiastic about current cognitive neuroscience than I am, but it's a matter of degree and I'm nowhere near denying the importance/coolness of the sub-discipline.

    It seems clear to me, however, that currently technology does not allow the kind of mind-reading FKF claim they can do. FKF's response to Engber, by the way, was "We are confident that he will have a better understanding and appreciation for fMRI brain scanning and its application for commercial purposes if he were to be scanned himself. Would it be possible to pass along our offer to Daniel to be scanned (and subsequently analyzed) in Los Angeles at his convenience." (http://jeffreygoldberg.theatlantic.com/archives/2008/06/a_challenge_to_slate_magazine.php)

    I'd be more convinced if they produced some peer-reviewed papers... (I looked for but could not find a single reference to an academic paper. While I don't want to commit myself to definite statements, FKF have all the hallmarks of a company built to sell pseudoscience).

    ReplyDelete