Our tax dollars at work

He's not dead, he's just sleeping.

Those scans are incredibly expensive. Did we need this study to show something we know mathematically?

"The point of the study was to show that false positives are commonplace in neuroimaging research."

Perhaps a longitudinal study with the same fish. . . ?

A salmon in a social setting ... was he the hors d'oeuvre?

Is this April 1, April fool's day?

If it works on dead salmon, will it work on undergraduate students?

This could explain among academic administrators and executives the propensity for meetings.

I have two responses:
First, the fish wasn't dead, it was just tenured.

Second, I've known many a dead fish to give lectures to undergraduates, amazingly enough, their evaluations showed false positives as well.

I have two responses:
First, the fish wasn't dead, it was just tenured.

Second, I've known many a dead fish to give lectures to undergraduates, amazingly enough, their evaluations showed false positives as well.

False premise about false positive, perhaps it was the fish's ghost or spirit responding...

False premise about false positive, perhaps it was the fish's ghost or spirit responding...

billyminch - No tax dollars were used in the scanning of the salmon.

ksledge - The scans were done during routine testing of our experiment protocol, which means they would have been done on a jar of mineral oil if not done on our salmon. Further, we do believe that the salmon is a worthwhile commentary on the multiple comparisons problem in fMRI. Not all studies use correction when reporting their statistics. We argue, by way of the salmon, that they should.

landrumkelly - It would be difficult to complete longitudinal investigation of the fish due to the fact that we ate it for dinner that night. It was delicious.

marklarson - You laugh, but I have received emails to that effect over the last 48 hours...

Hmm, I sense a salmon appetizer, Cornish hen for the main course, and perhaps pumpkin pie for dessert. But why then no scans of a bottle of wine, perhaps a good California Chardonnay?

You have got to be kidding me.
This guy is going to get credit for this?

You have got to be kidding me.
This guy is going to get credit for this?

Do you really think anyone actually did this? I smell a hoax, and it isn't even April 1.

prefrontal -- good to know! In that case, I wholeheartedly support the test. I like that the fact that you scanned a dead fish made the result more newsworthy, because sadly we do need people to open their eyes to the multiple comparisons problem in fMRI data analysis.

"It just so happens that your fish here is only MOSTLY dead..."

"I dreamt I was teaching a class...and then I awoke, and I was."

"I dreamt I was teaching a class...and then I awoke, and I was."

Congratulations on finding content to occupy this otherwise blank space. The effort put forth in this endeavor was obviously great. Suggestions for future content of this nature include photos of Kentucky Fescue grass in its dormant phase, a deflated automobile tire, and a bowl of Jello pudding.

I don't mean to sound petty, but really...in the history department we have to absorb funding cut after funding cut, lay off staff, etc., but this dead fish thing is worthy to fund? So, history is less important than a dead salmon?

Some of these posts are an affront to dead salmon everywhere! One need look no further for proof that "salmonism" is alive and well in the hallowed halls of academia. For shame! Its obvious that salmon -especially dead salmon - are not being accorded the respect they deserve. Dead Salmon sensitivity training is sorely needed (evidently) . . .

A little science literacy training wouldn't hurt either. The point is the validity (or lack thereof ) of the technique. They just as well might have used a dead rutabaga, if it had a brain ( which it doesn't so never-mind ).

aedonaldson - No department or grant funding was used for the Salmon project. It was completed during routine fMRI testing that would have been done whether a salmon was inside the bore or not. Further, it was presented at a conference that I was already scheduled to give a poster at.

I understand your position on declining department funding. We are going through 20-30% cuts here in the UC system, losing more staff and resources every day. Still, I do not think that the salmon data should be a rallying call to inter-department feuding.

Prefrontal: Isn't it possible that the salmon was not entirely dead, and that it was just registering some activity in response to stimuli? We usually know that "organismic death" and total tissue death are not coterminous. If they were, tissue transplantation would require the killing or mutilation of living donors. If I can take living skin from a dead man and graft it to a living patient, isn't it possible that the salmon, though organismically dead, still had some living tissue in the eyes and the brain? If not, how do we know? And if so, where is the error in the test? Don't we run the risk of affirming the consequent in our experiment--one designed to find what the researchers already presumed. Maybe they found that they were wrong in their presumption? Does this sort of thing have scientific implications for our debates over the ethics of organ transplantation?

new_theologian - This issue has come up more often than I would have guessed over the last few days. My position on the matter is that, while unlikely, it may very well be possible that there is some residual electrical activity in the salmon's brain. It had been dead and eviscerated for several days by the time I purchased it, but I cannot unequivocally say that there was absolutely zero electrical activity going on.

What I am able to say is that the version of fMRI that we used in the experiment is not sensitive to electrical activity, but to changes in blood oxygenation that are associated with neural activity. There was little blood left in the salmon's body and absolutely no blood flow through the salmon's brain during our scanning session. While residual electrical activity can be debated, it is unlikely that this would manifest itself as repetitive increases and decreases in the blood oxygen level dependent (BOLD) signal.

"The salmon was asked."

pre-frontal:

Any more info on how sensitive these scans are to electrical activity and how that compares to the electrical energy levels in salmon brains during visiualization?

just curious...
thanks.

D(e) Tro(o)p = French superfluous.
Is this a French postmodern comeback for the Alan Sokol hoax paper?

The experiment with the dead salmon shows that death is relative. It all depends what one means by 'dead'. After all, there are things like the living dead.

Perhaps the salmon was speaking to you from salmon heaven, where all rivers run free to the sea?

Prefrontal: Thank you for taking my question seriously. Theologians are sometimes dismissed for reasons with which I am sure you are familiar. In any event, is there any theory about what is being detected? I know that when I look out into total darkness--at least, as dark as it ever gets around my own head--I still "see" some sort of visual static. I'm not sure why this happens, but I know it's going on within my own neural structure. I suppose that a machine might have the same noise, and might register something even if nothing were there at all. But the assertion was that it did not register any activity in the pumpkin or the jar of peanut butter, or whatever it was that you fed it besides the salmon. I'm pretty doubtful that the salmon was still "alive," but it is a curious thing--and it would seem to me to introduce a level of uncertainty about which we ought to be concerned, when the subject is a human being.

dobbs2000 - While there is an interaction between electricity and magnetism, the scale at which this would be taking place is far too small to influence our scans. Some researchers have been trying to use MRI hardware to detect the magnetic field shifts associated with electrical activity in the brain. They have been calling this approach magnetic source MRI (msMRI). It is an interesting endeavor, but there is still a lot of debate regarding to how effective it can be. See this link if you are interested:
http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=2459244

new_theologian - All the evidence points to the fact that we are simply detecting noise in the fMRI acquisition. Your example of visual static is quite appropriate. When there is little light in the environment our retina will attempt to amplify what light is available. This also increases the noise, which we can sometimes observe as visual static. In fMRI we are similarly attempting to detect very weak signals. Like the retina, we must significantly amplify the signals for them to be useful. This leads to a corresponding increase in noise. We didn't actually compare the salmon results to the pumpkin or other objects, but is is likely that we would find 'significant' voxels in those items as well for the exact same reason. This is why we argue the field should take steps to be more statistically cautious.

Prefrontal,
Is this static or background noise common to all devices using magnetic imaging? Does the amount of "noise" change based on the location of the fMRI or is it a constant that is the same the world over. It would seem that magnetic fields generated by the earth and other objects would have some effect on an MRI. Can the fMRI machine be calibrated in such a way to ignore such "noise" without hampering its ability to perform its function?

In French, an April Fools' joke is indeed a fish -- "poisson d'avril." But I suspect that the authors of the study are more familiar with a different distribution of Poisson, n'est-ce pas?

primaryovertone - There is noise in every measurement obtained through MRI. While it is troublesome, there are many ways that fMRI researchers work around this problem. Some options involve improved scanner hardware, like stronger magnetic fields and better head coils. We can also optimize the parameters of the fMRI acquisition, as certain variables like spatial resolution can have a dramatic impact on noise in the images. Other approaches involve changing we design the experiments, as certain arrangements will have greater statistical power to detect real signals. Noise is a real issue in functional neuroimaging, but most people are quite motivated to do everything in their power to minimize it. The alternative is that none of their experiments would work!

I am not sure what influence geographic location might have, but I do know that each magnet has to go through a rigorous calibration process to compensate for disturbances of all kinds. This is called shimming the magnet, and the technicians have an amazing array of tools to measure very small errors in the magnetic field. As an aside, the magnetic field of the Earth is about 60 microtesla at is most strong point while the main field of a typical MRI machine is 3,000,000 microtesla - about 50,000 times stronger.

Did they ask the salmon questions? I think the whole scene must have been hilarious. Reminds me of the "Dead Parrot" Monty Python sketch.