In a recent issue of Nature, Nikos Logothetis, director of the Max Plank Institute for Biological Cybernetics, wrote some surprisingly harsh sentences about the experimental limitations of fMRI. The piece is especially noteworthy because Logothetis has probably done more than anyone else to document the tight correlation between what fMRI measures (changes in cortical blood flow) and the underlying neural activity of the brain. (His 2001 Nature paper, “Neurophysiological investivation of the basis of the fMRI signal,” has been cited more than 1200 times.)
Although brain scanner technology is often described as a “window into the brain,” Logothetis, in this most recent article, makes it clear that the metaphor of transparency is inappropriate. He cites a long list of factors that complicate the interpretation of fMRI data, from the challenge of distinguishing between excitation and inhibition to the difficulty of measuring the relative activation of different brain areas. If brain scanners are like a window, then the window has some very dirty glass.
The limitations of fMRI are not related to physics or poor engineering, and are unlikely to be resolved by increasing the sophistication and power of the scanners; they are instead due to the circuitry and functional organization of the brain, as well as to inappropriate experimental protocols that ignore this organization.
The one thing missing from this otherwise excellent article was a few choice examples of bad experimental design. I’ve complained before about the reticence among scientists to criticize each other in public forums, but the omission seems especially glaring in an article that is, in large part, about “what we cannot do with fMRI”.
Bonus question: has fMRI research jumped the shark?