“Neural Correlates of Interspecies Perspective Taking in the Post-Mortem Atlantic Salmon: An Argument For Proper Multiple Comparisons Correction”.
Please. Let that title sink in for a minute.
Even as a graduate student in my first semester, I’ve already read a metric butt-ton of articles (yes, that’s a scientific unit). Most of these have been for class; some have been to beef up my knowledge for a lab rotation; and still others have been read purely out of curiosity.
Reading scientific papers is not for the faint of heart. Even for someone keenly interested in their subject, poor writing and dry content can put the reader right to sleep. That being said, it’s refreshing to find a paper that is, for a wonder, entertaining– such as that by Bennett et al. 2012. What could have been a dry treatise on the importance of using proper statistical and filtering techniques for functional imaging is instead an experiment to end all imaging experiments: running a dead fish through an fMRI. But what does a dead fish have to do with statistics?
For those of you who aren’t familiar with functional imaging, let me mince words: it’s a tricky business. There are a veritable boatload of reasons why, but Bennett et al. (2012) were interested in one specific pothole of fMRI: false positives.
Think about it: you’re a scientist studying a particular behavior, using fMRI to look at what parts of the brain “light up”; but therein lies the first challenge: all of the brain lights up. Even after you subtract out steady-state activity from task-oriented activity, you still have a great deal of “activation”. So, you are faced with the question of how to set your threshold.
This is where Bennett & co. take their colleagues to task. A multiple comparisons correction is basically a method of comparing voxel (volume elements) of an fMRI image to determine if there is a real difference between them. It helps eliminate false positives. Bennett et al. (2012) are wagging their fingers because, guess what? A lot of neuroimaging studies fail to do this.
And what better way to slap someone on the wrist than with a dead fish?
Improbable as it might be, Bennett et al. (2012) procured their subject from the local market, told the fish to perform a behavioral task, and ran an fMRI. Should a dead fish show any neural activity? It’s a no-brainer. Of course not; but, lo’ and behold, it did indeed. Parts of the dead fish brain lit up.
I won’t spoil the whole article for you; it’s quite a gem, and you should definitely read it for yourself. I will, however, share with you this tidbit:
“Either we have stumbled onto a rather amazing discovery in terms of post-mortem ichthyological cognition, or there is something a bit off with regard to our uncorrected statistical approach. Could we conclude from this data that the salmon is engaging in the perspective-taking task? Certainly not. By controlling for the cognitive ability of the subject we have thoroughly eliminated that possibility.”
The sass in this article is palpable. It’s truly a literary treat. But where oh where could you publish such a finding? Sadly, the journal in which this article was originally published seems to no longer exist. But, if you ever find another article published by the Journal of Serendipitous and Unexpected Results (JSUR), I recommend you give it a read.
Bennett, C. M., A. A. Baird, M. B. Miller, and G. L. Wolford. 2012. Neural Correlates of Interspecies Perspective Taking in the Post-Mortem Atlantic Salmon: An Argument For Proper Multiple Comparisons Correction. Journal of Serendipitous and Unexpected Results 1:1-5. (full text)
For more information on the project: