MRI Killed The Radiotracer

By Neuroskeptic | August 30, 2013 11:27 am

Positron Emission Tomography (PET) was once among the most powerful techniques available for researching human brain activity.

By injecting a volunteer with a radioactive tracer, such as a glucose derivative, and monitoring the radiation emitted from the brain over the next few hours, neuroscientists could see where in the brain most glucose was being absorbed – and where neural activity was happening.

But in recent years, the influence of PET in neuroscience research has fallen dramatically, as shown in a new paper by Neuroimage assistant editor Paul Cumming: PET Neuroimaging: The White Elephant Packs His Trunk?

Here’s the numbers of research papers published per year (note the exponential scale.)

PET studies of the brain have been declining since about 1985, despite the fact that research about the brain in general has grown.

It seems it was fMRI scanning that killed PET. Cheaper, safer (no radiation), and more versatile, the growth of fMRI has been rapid and constant since 1990 in proportional terms.

Still, fMRI isn’t everything. While PET’s role as a measure of brain activity is over, it remains the only method that can probe certain aspects of brain chemistry.

Anyway. I’m very fond of these kinds of bibliometric graphs. Some that I’ve done previously include the rise of the mouse as a laboratory animal, the fall of Freud (and psychoanalysis) and which brain regions are most researched right now.

ResearchBlogging.orgCumming P (2013). PET Neuroimaging: The White Elephant Packs His Trunk? NeuroImage PMID: 23959198

CATEGORIZED UNDER: fMRI, graphs, methods, science, select, Top Posts
  • Enzo Tagliazucchi

    why there are more counts for “fMRI” than “PET” in the 80’s? I thought that fMRI started to emerge and become mainstream around 1990, is that right?

    • Neuroskeptic

      You’re absolutely right, well spotted. The first “fMRI” paper was in 1993 (there had been research before that, but that’s when the acronym was coined):

      I think the problem with the graph is that it searched for fMRI AND brain, but if you do that, PubMed treats fMRI as an alias for other terms including “NMR”.

      He should have searched for “fMRI” AND brain to make it literal.

  • MedPhys

    You really have no grounds for calling fMRI “safer” as there is no evidence that the small amount of radiation from nuclear medicine studies is detrimental in any way. The assumption that any amount of radiation regardless of how small is harmful is based on the outdated linear no-threshold model.

  • Pingback: Are We Heading For "Peak Autism"? - Neuroskeptic |

  • Pingback: Are We Heading For "Peak Autism"? | Today Health Channel()

  • Jack Smithrod

    This result, one I can believe, within the limits of the comment about fMRI and MRI being conflated by PubMed, is really only applicable to brain blood flow/oxygenation. I think “safety” is a small issue, given than the radiation dose really is fairly small. The more important feature of fMRI is its ability to capture the oxygen levels (reflecting use-related blood flow) of multiple brain states in a single session. This is not possible with PET using FDG since you really only get a shot at one brain blood flow state per day (maybe two). fMRI is much less limited in this area and because of this has met with much greater acceptance among brain researchers.

    In other areas, PET has clear advantages not seen with MRI. PET is exquisitely sensitive to biochemistry and can reveal and quantify extremely small concentrations of a wide array of receptors and other tissue markers. In cardiology, new tracers are currently under development that may yield a substantial increase in the accuracy of myocardial perfusion imaging. This is a very underdeveloped area in MRI and very far from being ready for the prime time of clinical practice. Even if you were to develop an MRI tracer for targets in the tissue, which is certainly possible, they would have to be much more concentrated as the sensitivity of MRI to any of the currently known tracer materials (Gd, for example) is about 5 to 6 orders of magnitude poorer than that of PET.

    Don’t count PET out yet – it’s a powerful technique and yields information that is not obtainable by any other means.

  • Pingback: DDR pg 127-128 Question 1, 2, 3 – PSYCH 250()



No brain. No gain.

About Neuroskeptic

Neuroskeptic is a British neuroscientist who takes a skeptical look at his own field, and beyond. His blog offers a look at the latest developments in neuroscience, psychiatry and psychology through a critical lens.


See More

@Neuro_Skeptic on Twitter


Discover's Newsletter

Sign up to get the latest science news delivered weekly right to your inbox!

Collapse bottom bar