Lucid Thoughts

I haven’t picked on the fMRI folks for awhile, but was inspired today by a new post over at the excellent neuroscience-of-language blog Talking Brains entitled “Functional brain imaging, it’s not always where you think it is.”  And that reminded me that I also wanted to write something about an excellent article in Scientific American Mind from about a year ago entitled “Five Ways Brain Scans Mislead Us.”

I’m taking this opportunity for a “two-fer” because both these items draw attention to a central dogma of fMRI interpretation that is particularly prevalent in the popular media.  I call it “locationism” because it’s the belief that thoughts and mental functions happen at specialized locations in the brain, so if you just watch a particular location “light up”, it tells you what someone is, quite literally, thinking.

Now reputable neuroscientists have been fighting a rearguard battle against this kind of interpretation for awhile – see for example:

Knight, R. (2007). NEUROSCIENCE: Neural Networks Debunk Phrenology Science, 316 (5831), 1578-1579 DOI: 10.1126/science.1144677

But despite scientists’ best efforts, the simplicity of locationism seems to regularly outrun its stodgy companion, accuracy, not only in the annals of popular neuroscience journalism, but also, sadly, in the rhetorical flourishes of our neuromarketing brethren.  A recent “research article” published by a neuromarketing firm states rather breathlessly (exclamation point in the original):

notice that for each of the three meaningful images different parts of the brain light up! This is strongly suggestive that different types of visual content in an ad are processed in different parts of the brain and, ultimately, may be stored in different memory systems of the mind.

And that’s locationism in a nutshell.

The Talking Brains post is pertinent because it describes a dilemma every fMRI study faces when it combines results across multiple subjects – in order to find average activations, researchers must “normalize” unique human heads into a common “space.”  As the author, neuroscientist Greg Hickok, puts it:

It is a standard practice in functional brain imaging studies to spatially “normalize” one’s data to a common space such as Talairach space or MNI space. The reason for doing this is to allow data from multiple subjects to be averaged. … It is acknowledged that there is error in the process due to variation from one brain to the next. Yet at the same time the (also standard) practice of overlaying warped and group-averaged brain activations on a single high resolution structural MRI temps [tempts - sic] investigators and readers alike to interpret the location of the blob of activations very literally. More than once I have heard comments about this or that activation being on the upper or lower bank of a sulcus, for example. We simply don’t have the spatial resolution to tell in this sort of fMRI analysis.

So, the most-often proclaimed virtue of fMRI, its high spatial resolution, proves to have some problems of its its own.  The post continues:

The figure below shows activations in native space (left side sagittal and coronal images), that is, non-normalized and on the subject’s own MRI (also non-normalized) and after warping to a standard space (right side sagittal and coronal images). The activation of interest is in the crosshairs.

Notice that in native space the focus of activation is within the Sylvian fissure and after standardization it moves up above and is on the SMG. This happened consistently enough across subjects in two independent datasets that the group focus ended up in the SMG despite individual subjects showing activation in the Sylvian fissure. This bit of spatial error may not seem like much of a problem, but it is. This jump crosses cytoarchitectonic boundaries and could lead to dramatically different functional interpretations based on what is known about these areas.

What this implies is that great spatial resolution in a single head does not necessarily translate into great spatial resolution across a sample of heads.  And this, in turn, means that a location “lighting up” on average may not be precisely the same location for each and every subject in the study.

Michael Shermer’s November 2008 article in Scientific American Mind is a more general and accessible read, and goes much more directly to the point of debunking the locationism error.  Shermer attributes this error to an overplayed metaphor, the brain module:

Many neuroscientists have employed the module metaphor to describe specific regions of the brain “for X,” with X being whatever happens to be the task given to subjects while a machine scans their brains. Such tasks might include selecting brand logos they prefer (say, Coke or Pepsi) or political candidates they would vote for (conservatives or liberals).

Scientists often use metaphors such as these as aids in understanding and explaining complex processes, but this practice necessarily oversimplifies the intricate and subtle realities of the physical world. As it turns out, the role of those blobs of color that we see in brain images is not as clear-cut as we have been led to believe. “There are no modules that are encapsulated and just send information into a central processor,” declares philosopher of the mind Patricia S. Churchland of the University of California, San Diego. “There are areas of specialization, yes, and networks maybe—but these are not always dedicated to a particular task.” (emphasis added)

Shermer goes on to explain a fundamental point, that brain areas do not have single functions.  On the contrary, “brain areas activate for various reasons”:

Interpreting fMRI scans is as much an art as a science, [neuroscientist Russell] Poldrack admits. “It is tempting to look at one of those spots and say, ‘This is where X happens in your brain,’ when in fact that area could be lighting up when involved in all sorts of tasks,” he explains. “Take the right prefrontal cortex that lights up when you do almost any difficult task. One way to think about it is in terms of networks, not modules. When you are engaged in thinking about money, there is a network of several different areas involved in communicating with one another in a particular way. Thus, the prefrontal cortex may be involved in many different tasks. But in communication with other brain networks, it becomes active when engaged in one particular task, such as thinking about money.” Teasing these differences apart requires making relative comparisons across a spectrum of tasks. Certain experiments work especially well with fMRI because decisions provide contrasts between tasks, giving the neuroscientist something to compare.

What about research showing differences in rational versus emotional parts of the brain, as in the “emotional low road” in the deeper and more ancient parts of the brain and the “rational high road” in the cortical regions of the brain? “There are rational and emotional ways of thinking,” Poldrack says. But “it turns out that they interact with one another a lot.” The amygdala, an area typically associated with processing the fear response, also is activated by arousal and positive emotions: “If I put you into a state of fear, your amygdala lights up. But that doesn’t mean that every time your amygdala lights up, you are experiencing fear. Every brain area lights up under lots of different states. We just don’t have the data to tell us how selectively active an area is.” (emphasis added again)

So locationism is just bad science.  And fMRI has other challenges, as described in Shermer’s “Five Ways Brain Scans Mislead Us”.  But I will leave those for you to review on your own, as Scientific American would like to charge you $7.95 to read the whole article, so I shouldn’t give away more than I already have.

Our conclusion?  The next time the “metaphor-loving center” of your brain lights up in response to the enticing metaphor of brain centers, activate the old prefrontal cortex for a little reality check.  Locationism is a very misleading way to think about your brain in action, and fMRI brain scanning, although a wonderful technology for many purposes, unfortunately makes the myth of locationism very hard to resist, especially for non-scientists who are buying (or selling) fMRI scans as a window into the brain at work.