Op, Op, Op. The Neuroscience of Gangnam Style?

By Neuroskeptic | January 16, 2017 3:29 pm

“Our results revealed characteristic patterns of brain activity associated with Gangnam Style”. So say the authors of a new paper called Neural correlates of the popular music phenomenon.

The authors, Qiaozhen Chen et al. from Zhejiang in China, used fMRI to record brain activity while 15 volunteers listened to two musical pieces: Psy’s ‘Gangnam Style‘ and a “light music” control, Richard Clayderman’s piano piece ‘A Comme Amour‘.

Chen et al. say that Gangnam Style was associated with “significantly increased fMRI BOLD signals in the bilateral superior temporal cortices, left cerebellum, left putamen and right thalamus cortex”. They conclude that these results reveal something about the mechanisms for the “Gangnam Style-induced” positive emotional response. But I don’t.


For one thing, only the temporal cortex blobs survive multiple comparisons correction (see their Table 1), but there’s an even bigger problem: the stimuli are poorly matched. The two songs used differ in many fundamental ways: amongst other things, ‘Gangnam Style’ has lyrics while ‘A Comme Amour’ is purely instrumental, so any brain activation differences might reflect speech processing.  Also, ‘Gangnam Style’ is one of the most famous songs of all time, while few people will have heard ‘A Comme Amour’ before, so the differences might reflect familiarity.

Chen et al. do acknowledge these issues, but to my mind they are pretty big ones.

Yet there’s more. Chen et al. also conducted a PET study. Using the tracer 11C-NMSP they sought to see whether listening to “Gangnam Style” for about 45 minutes altered dopamine and serotonin receptor levels, compared to 45 minutes of ‘A Comme Amour’. It turns out it did: “11C-NMSP binding was found significantly increased in the left superior temporal gyrus and left putamen, but decreased in the bilateral superior occipital cortices under Gangnam Style music condition”.


The authors interpret this as showing that Gangnam Style increases dopamine D2 receptor expression in the putamen, perhaps reflecting music-related pleasure. However all of the problems that affected the fMRI study also arise here, including the fact that, as far as I can see, none of the PET results survived multiple comparisons correction.

Overall this paper is a bit worrying in my opinion. The fMRI study was, probably, a waste of money, and PET is even more expensive. Worse, PET involves the injection of a radioactive tracer which is believed to increase the risk of cancer. Now this risk is thought to be very small (of the order of 1/10,000 per PET scan) but even so, it’s only ethical to expose volunteers to this risk if there is a scientific benefit to doing so, and I can’t see the scientific value of this study.

You know what I’m saying?

ResearchBlogging.orgChen Q, Zhang Y, Hou H, Du F, Wu S, Chen L, Shen Y, Chao F, Chung JK, Zhang H, & Tian M (2017). Neural correlates of the popular music phenomenon: evidence from functional MRI and PET imaging. European Journal of Nuclear Medicine and Molecular Imaging PMID: 28083689

  • Nitric-X

    Actually I see a much larger ethical risk than PET scanning: they made people listen to either Gangnam Style or Richard Clayderman for 45 min! By any means, this is inhumane treatment and nobody should be forced to experience this… 😉

  • Nick

    This looks like one of those (depressingly not rare) cases where the authors started with the press release and worked backwards.

    • OWilson

      More neuroscience by “Rorschach Test” ?

  • Денис Бурчаков

    I agree with Nictric-X. Making people listen to Psy for so long is comparable to the use of the children’s song “Barney & Friends” on Iraqi prisoners-of-war (“The Men Who Stare at Goats”).

  • Jamie

    How much money would this have cost, approximately? Also, what were the authors even trying to investigate/shed light on?

    • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

      Good question. I can only say how much it would have cost for me to do this study in the UK – it might be different in China.

      The study involved 15 MRI scans (about 15 minutes scan time each) and 28 PET scans (about 45 minutes each).

      The MRI would cost about £200 per scan = £3,000.

      The PET would cost approx. £2500 per scan = £70,000

      Then there are salary costs, I’d say ballpark estimate this study would occupy 50% of full time for 2 researchers working for 2 years from start to publication, assuming they each earn £40,000/yr this makes £80,000.

      So roughly £153,000. But there’s a lot of uncertainty, in particular PET scans can cost either more or less than £2500 depending on the centre.

      • Jon Mellon

        I’m intrigued about this. Do you really think this paper would take 2 FTE years?

        The scan time is only 25 hours (based on what you’ve listed), even with several multiples for the laboratory work, that’s presumably a fairly small proportion of two years.

        I’m guessing analysis-wise they probably pushed the data through some standard procedure without thinking too hard about it (based on the design of the rest of the study). In terms of writeup, it’s 9 pages and looks fairly standard.

        Is there something I’m missing that adds so much time to writeups in neuroscience?

        • Jamie

          I don’t know how long it would take, but unless the Chinese are super efficient at this (quite possibly they are) I often see that research takes a very long time just because of booking in willing participants, conflicting schedules etc. especially if they had to have preparatory sessions and stuff. 2 years might be a bit long though. I agree they probably weren’t spending months working out the optimal analysis strategy.

        • http://blogs.discovermagazine.com/neuroskeptic/ Neuroskeptic

          You’re right that in theory this study wouldn’t take 2 FTE years. In theory it could be done in a couple of months. However in my experience nothing in neuroscience (especially where PET is involved…) is ever as simple as it seems. In the UK for example it could well take a FTE month of doing paperwork (and 6+ months of waiting for it to be processed) just to get the ethics approvals necessary to start a PET study.

          The writeup likewise could take much longer than you’d think. Papers often go through dozens of revisions.

          So while 2 FTE years seems a lot, I do think it’s realistic for the UK. Maybe in China everything is easier…

  • http://www.mazepath.com/uncleal/qz4.htm Uncle Al

    “It’s a Small World” at Disney literally drove workers mad. Instrument them and do it again. Write a series of papers at the mumble factory, then conduct studies and cherry pick data to fill in the p values.

  • Doomroar

    ” it’s only ethical to expose volunteers to this risk if there is a scientific benefit to doing so, and I can’t see the scientific value of this study.”

    Say, is this really true? sure the increased risk in this case doesn’t seem to be much, but since you are touching the topic, where exactly lies the line between too much or too little harm done in order to advance scientific knowledge? clearly we part from an utilitarian position, so there must be a balance that allow us to get away with stuff, which is something that other ethical system just don’t allow.

  • Pingback: Signal Conduction – 20 January 2017 | Cognitales()



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