Neuroskepticism - Neurotechnology in National Security and Defense - page 186

The language of lie detection is more worrisome still when it is deployed to
describe the use of brain imaging and related technologies that do not patently rely
on external manifestations of anxiety. Newer technologies purport to show us what
is going on in someone else's brain and—if one were to believe much of the press
coverage—in their mind. As the British experimental psychologist Richard Henson
(2005, 228) has succinctly observed, “There is a real danger that pictures of blobs
on brains seduce one into thinking that we can now directly observe psychological
processes.” It is to the seductive power of brain images that I now turn.
NEUROIMAGING AND NEUROSCIENTIFIC IMAGINARIES
Brain images are ubiquitous. They are no longer the sole province of medical and
scientific journals. Viewers of cable news and print media alike are frequently shown
brain images. They usually accompany features breathlessly reporting that brain
imaging has heralded the end of lies or finally lifted the shroud to reveal “how
the brain handles love and pain” (Kane 2004). But functional neuroimages are not
images insofar as that word is used to connote optical counterparts of an object pro-
duced by an optical device. Brain activity does not have an optical component. We
cannot literally see people think—although these images may suggest as much to
those with little or no understanding of how they are produced. Rather, neuroimages
are carefully constructed representations of the brain and its functions. When the
results of fMRI-based cognitive neuroimaging studies are presented to us in image
form (as is almost invariably the case), tiny changes in blood oxygenation levels (less
than 3%) are represented by bright colors (usually reds, yellows, and blues). These
changes are the product of a comparison between levels of blood oxygenation for a
chosen cognitive task and those for an activity considered a suitable baseline. These
changes are interpreted as markers of local activation or inhibition in the regions of
the brain in which they occur.
Many science studies scholars and bioethicists have critiqued the manner in
which brain images are produced, constructed, and interpreted (see, e.g., Dumit
2003; Wolpe et al. 2005; Marks 2007b; Joyce 2008). I do not review all these cri-
tiques here. Instead, I wish to highlight a simple methodological point that is often
not appreciated. An fMRI—the kind so frequently reproduced in glossy maga-
zines for lay readers—is usually not a single image of one person's brain. There are
two reasons for this. First, the changes represented in brain images are often not
those of a single experimental subject. More commonly, they are representations of
composite data from a small experimental cohort. Second, these color images are
superimposed on a higher resolution structural brain image, just as Doppler weather
radar images are superimposed on geographic maps. The higher resolution image is
intended to reveal the topography of the brain—just as the geographic map (on to
which the constructed Doppler image is superimposed) is intended to show that, for
example, the latest hurricane is 50 miles off the coast of Georgia. However, the struc-
tural image of the brain need not be taken (and is often not taken) from any of the
subjects of the experiment. This is important because the neurological analog of the
state of Georgia in my brain is (like the lyrical Georgia on my mind) not necessarily
the same as yours. Put another way, there is considerable variation in the anatomical