A Paper on Pain and the Power of Negative Data

Apr 28 2010 Published by under Uncategorized

So Sci was scrounging around the internet for blog topics recently. Every time she does this, she is of course completely overwhelmed by the piles of cool and bloggable science out there (seriously, I've got like 20 topics for the next week, of course not all will make it). But she's also surprised to see some familiar faces. A lot of times it's a familiar face related to her field and the stuff she's been recently interested in (for example, Yavin Shaham just wrote an interesting editorial on binge eating and food addiction in Nature Neuroscience which Sci might have to blog), but everyone once in a while it's something like this:
zen blog1.png
OMG! Sci knows that guy!!
So of course Sci had to go and ask if I could blog it. And of course the gracious Zen said yes!
ResearchBlogging.org Puri and Faulkes. "Do Decapod Crustaceans Have Nociceptors for Extreme pH?" PLoS ONE, 2010.
And first of all, I would like to introduce you to today's three species:
Louisiana red swamp crayfish (Procambarus clarkii)


Nociception is, technically speaking, the detection of stimuli that are noxious. Colloquially speaking, we like to call it pain, but that's really not being specific enough. What it actually is is the ability to detect a stimulus which might bring harm to the creature involved. This can be anything from something mechanical (excess bending or tearing), to something temperature related (like excess heat or cold), to something chemical (like high acidity).
Of course I'm sure everyone can think of all sorts of good reasons for studying pain. But why study pain in invertebrates? Because studying nociception in invertebrates often translates really well to the human condition. Finding receptors responsible for nociception can sometimes lead to discoveries in humans for similar receptors, and so can change the way that we study pain in humans.
And why decapods? Well, obviously, because they're cute!

First off, you all should know that decapod means "ten footed", meaning that these little dudes have 5 pairs of legs. Sci is drawn to these little guys, in part because she studied one of their relatives, the amphipods (not closely related and a lot more legs) in her undergrad days. And also because they are delicious.
But why decapods in particular, you might ask? Because nociceptors, the receptors that process these noxious stimuli, have been found in mammals, worms, and insects, but not a lot has been done in decapods. One study found what could be nociceptors in prawns, but the results were kind of ambiguous. Not only that, nociceptors often are similar in broad groups of species but different in specific species (like the naked mole rat). And so to determine whether decapods in general have nociceptors, Zen and his grad student decided to look at the three species shown above. They first looked at the behavior of the animals in response to known noxious stimuli, really basic and really acidic substances, and then they went in with electrophysiology to see what the neurons were doing.
So how, you might ask,does one study the behavior of decapod crustaceans and their response to noxious stimuli?
Check the antenna.
If a crustacean is exposed to noxious stimuli (like acids or bases) on their antenna and it is capable of detecting them, presumably it'd want them OFF, right? So a good measure of whether or not the animal behaves as though it can sense noxious stimuli is whether or not is grooms its antenna when noxious stimuli are dropped on it. You basically grab your crustacean, take it out of the water (I guess you'd probably dry it off a little), and use a cotton tipped swab to swab some acid, base, or control on one of the antenna, with the other antenna serving as a control. Then, watch for grooming. You can see the results here:
zen blog6.png
You can see that the three species, when exposed to acid (on the left) and base (on the right), they didn't show any real difference in the amount of antenna grooming. So it looks like, behaviorally, the crustaceans can't sense the acid or base. They even tried a local anesthetic, benzocaine, which apparently produced grooming in another paper, and didn't get a response.
They then checked out the neural activity. And this part is the part Sci thinks is really cool. You take a crawdad. You put it on ice to anesthetize it (lobsters on ice, anyone?), and cut off one of their antenna. You put that antenna in a dish, and you can then DISSECT out the nerve and record from it!!! Sci wishes she could do this...in a brain electrophysiology is considerably more difficult and annoying. But anyway, you dissect out the nerve, and bathe it in solutions, either innocuous or noxious, and see what happens to your recording.
And they found what they expected to find here, a good basal level of cell firing. But like the behavioral responses, it appeared to be unaffected by acids, bases, or benzocaine. Bummer.
This paper actually goes AGAINST a previous paper which indicated that decapod crustaceans have nociception. So who's right? Sci doesn't know, but give the variety of species tested here and the variety of tests, she's putting her money on Zen.
Now, you're going to say, does this mean decapods can't feel PAIN?! No. Not at all. It just means that they don't response to noxious stimuli in the same way that other species do. They could still have nociceptors that work in a different way. After all, how often is a crustacean going to run into a solution that is strongly acidic in its life? In many species changes in pH hit the same nociceptors that things like heat and mechanical stress do, but it's possible that decapods could be an exception. They could have nociceptors that respond to bending or pressure or heat or cold, and not to acids and bases.
And here's where Sci is going to go off on a tangent. Yes. This paper was a big pile o' negative data. You know what? That's GREAT. We NEED this data. Sci personally feels that negative data is not published half as often as it should be. This is for two reasons: (1) it's really hard to publish negative data because it's often seen as boring, and (2) it's often not what you expected. As far as point (1), that's an issue with journals and reviewers. Negative data often doesn't have the "impact" of a huge wild positive result to a hypothesis. And for (2), well, scientists are human, and we get very attached to our hypotheses. When something just...constantly fails to work, people are often more likely to blame the person working on it than admit their hypothesis is wrong.
BUT, that doesn't mean this data isn't important. On the contrary, it is JUST as important as the possible positive result would have been. And we produce a LOT of it. Negative findings are still data, and still say something important about the world. There is a possibility that decapods do not sense "pain" or noxious stimuli the way other species do. This could have a lot of implications for nociception and how we study it, for decapods as a group of species, and in other species of animals. And who knows, this paper full of negative data could one day turn an entire field on its head. You never know.
So Sci just wants to take a moment to stand up for negative data. Don't be afraid to publish it. It deserves to see the light of day. No matter what, you've found something important about the world (or a tiny part of the world, studying an effect that is true under certain specific conditions. It's still the world, dangit). Be proud of your negative data! And Sci is proud of Zen and Sakshi, for a good job, with some pretty data, and for showing something interesting, even if it's something the world didn't expect.

Sakshi Puri, Zen Faulkes (2010). Do Decapod Crustaceans Have Nociceptors for Extreme pH? PLoS ONE, 5 (4) : 10.1371/journal.pone.0010244

14 responses so far

  • Janne says:

    I'll go out on a limb here and say that one problem here is that people tend to confuse negative data with unsurprising data. This is a good example where this is not the case: You would expect these species to have working receptors, and yet careful examination fails to find them. In fact, this is likely more surprising (=interesting) than if they had found them.
    But of course a lot of negative data isn't just negative; it is unsurprising. A lot of good, solid research is of the form "Hey, what if X was true? Nobody would expect it, so if it were, it'd totally validate our theory!", followed by "Nah, X was false after all. We need to come up with something else". And unsurprising results are indeed usually not worth publishing.

  • Tony Jeremiah says:

    Sci personally feels that negative data is not published half as often as it should be This is for two reasons: (1) it's really hard to publish negative data because it's often seen as boring
    In practice, the main criterion for having data from a quantitative study even peer reviewed (at least in the social sciences), is a statistically significant value of p = .05 or lower. I notice one data point in the study has a value of .028 (a statistically significant result), and am wondering if that is the major reason for publication.
    You can also have studies that meet the statistically significant criterion, but which are practically non-significant. For example, imagine a newly developed drug increases IQ by 1 point and is statistically significant (possible given a large enough sample size). While statistically significant, it would not be worth publication since it's not a practically significant IQ increase.
    Other studies can produce statistically non-significant results (and thus, technically, are not publishable), but can still be interesting. For example, psi phenomena that do not reach statistical significance.
    In general, the point that quantitative data not published should be published more often is an important one. Failure to reach a statistical criterion (also known as the file draw problem) is an important major issue when conducting meta-analyses on particular phenomena. Without access to these unpublished data, meta-analyses can be faulty to the extent that the analyses are based only on data producing statistically significant results.
    Publications that concern themselves with qualitative analyses (the work of Mendeleev and Darwin have strong qualitative content), involve an entirely different set of (non-quantitative) acceptance parameters.

  • Zen Faulkes says:

    There are two kinds of negative data: Negative data that you expect to be negative, and negative data when it should be positive data. The former may not be all that interesting, and might signal a timid hypothesis. But the latter?
    As Isaac Asimov used to write, the big moments in science are often not preceded by "Eureka!" but by, "That's funny..."
    Minor correction: My co-author Sakshi is an undergraduate (graduating this sumemr), not a grad student. We ♥ you, anyway!

  • JerryM says:

    "When you're shown wrong, you become more right."
    re the study itself. Is this showing there is no response to noxious stimuli, or that there is no difference to response to noxious vs other stimuli?
    iow, would the beast preen their antennae just as much when it is touched with a hot rod or a water-temped rod?

  • Zen Faulkes says:

    JerryM: Strictly speaking, we saw no difference between responses to control (water) and noxious (acids or bases) stimuli.
    We're definitely NOT saying, "Decapod crustacean won't respond to any kind of noxious stimuli, ever." I desperately hope nobody comes away with that impression.

  • Tony Jeremiah says:

    If a crustacean is exposed to noxious stimuli (like acids or bases) on their antenna and it is capable of detecting them, presumably it'd want them OFF, right? So a good measure of whether or not the animal behaves as though it can sense noxious stimuli is whether or not is grooms its antenna when noxious stimuli are dropped on it.
    Sounds very much like an operant conditioning procedure. In which case, acids and bases (presumably) serve as reinforcing stimuli for grooming behavior. However, it looks like the data is trending toward a neutral pH being the noxious stimulus (i.e., a number of the control conditions seem to be evidencing more grooming behavior when it seems that it should be the reverse).
    Questions
    (1) Is it known what pH the various decapod species are most comfortable with (perhaps based on their physiology)?
    (2) Has a study been conducted showing what the grooming behavior of each species looks like from pH 1 through 14?
    (3) Were t-tests the only stats conducted for the analysis? If not, it would be interesting to run an ANOVA to see if significant interactions exist

  • Hugh Clements-Jewery says:

    I was trained by my PhD advisor that good science involves testing a specific hypothesis, formed with good rationale, using an appropriate method. If you do this and there were no significant differences when you analyze the results, the hypothesis is proven to be false (or, at least, there is no evidence in support of the hypothesis). However, this is still useful and reasonable information that advances the field. Therefore, I have no problem with a lot of 'negative' studies.

  • Zen Faulkes says:

    Tony: One thing that will help to understand out paper is to have a peek at the paper by Barr and colleagues (2007. Animal Behaviour 75(3): 745-751.) I am pretty sure they did not think that acids and bases were reinforcing stimuli for grooming.
    1. I don't know if there are any behavioural tests for "favourite" pH for different decapod species. Physiological saline tends to have a pH of 7.2-7.5 or so. You might reasonably expect marine species to prefer the pH of the ocean (about 8); freshwater species might be more variable.
    2. Not that I'm aware of. But that kind of pH detection would be different from nociception, just like thermoreception is different from nociception. For example, you may have a preference for 20ºC over 10ºC, but that preference isn't mediated by nociceptors.
    3. We only compared two conditions (control and noxious stimuli) at a time, so just t-tests. An ANOVA would work if we planned for it, but the experiments unfolded in series rather than being cut from whole cloth at the start. The chain of events was sort of like this:

    "What? There's no difference? There should be a difference. Let's try this variation..."
    [Some weeks later, after experiments run and data analyzed]
    "No difference again? Let's try this variation..."
    [Some weeks later, after experiments run and data analyzed]
    "No difference again? Let's try this variation..."

    Because the experiments were somewhat drawn out over time, it would be problematic to pool the results and try to analyze it with an ANOVA.

  • Neuroskeptic says:

    One thing that's good for causing pain in mammals is capsaicin. What makes chilli peppers burn. Could you try capsaicin?

  • Zen Faulkes says:

    Neuroskeptic: Not only can we try capsaicin, we have tried capsaicin. 🙂 Here’s an excerpt of our abstract from the 2009 Society for Integrative and Comparative Biology meeting (doi: 10.1093/icb/icp002, pg. e139):
    Do crayfish like spicy foods? and other tests of crustacean nociception

    To test for responses to capsaicin, foods containing variable amounts of capsaicin were given to P. clarkii. The crayfish were given Anaheim peppers (low capsaicin content) and habaneros (high capsaicin content) first separately, then together. Crayfish ate both habaneros and Anaheim peppers. When given both types of peppers at the same time, the crayfish preferred habaneros.

  • DDeden says:

    Antennae may not be pH sensitive, but gills may be. Among human marine freedivers, breathing is dependent on pH of blood, 'air hunger' and diaphragmatic contractions are triggered by accumulating carbonic acid in blood affecting internal chemoreceptors.

  • Zen Faulkes says:

    A new paper in press has evidence suggesting that crayfish are not pH sensitive.
    Bierbower and Cooper were testing whether crayfish can detect carbon dioxide (they can). Because putting carbon dioxide into water creates carbonic acid, they used acidic water (pH 4.85) as a control in addition to regular water, and the crayfish showed no change in behaviour at all.
    I've written a blog post about this over at NeuroDojo.
    Reference
    Bierbower S, Cooper R. 2010. The effects of acute carbon dioxide on behavior and physiology in Procambarus clarkii. Journal of Experimental Zoology Part A: Ecological Genetics and Physiology DOI: 10.1002/jez.620

  • This_Is_Not_An_Exit says:

    Unfortunately it seems the authors of this paper have missed out on the key concepts of nociception and pain. By definition, nociception results in an immediate reflex response and as such prolonged grooming behaviour cannot be classified as a nociceptive behaviour. Barr et al. only refer to immediate tail flipping as being consistent with nociception. The pronlonged, focussed grooming patterns they observed were postulated to be part of a subsequent pain coping strategy. Suggesting that Barr et al. have mischaracterized grooming as nociception is thus a moot point.

  • NB says:

    Seriously, there /is/ a journal for negative data:

    Journal of Negative Results in BioMedicine
    http://www.jnrbm.com/

    also
    Journal of Negative Results - Ecology and Evolutionary Biology
    http://www.jnr-eeb.org