A critique of the paper:
“Do fish have nociceptors: Evidence for the evolution of a vertebrate
sensory system” published in Proceedings of the Royal Society, 2003 by Sneddon,
Braithwaite and Gentle.
James D. Rose, Ph.D.
Department of Zoology and Physiology
The paper by Sneddon, et al. is flawed
and does not provide any legitimate evidence that trout are capable of feeling
pain. There are numerous problems with
methods and data interpretation in this paper but this critique will focus only
on those of greatest significance.
First, an explanation of the invalid claims for evidence of pain will be
presented, followed by an account of the misinterpretations of the behavioral
results.
Flaws in the
argument for a demonstration of pain.
1.
The authors’ definitions of pain and nociception are invalid,
consequently this paper does not actually deal with pain (a conscious
experience), it deals only with nociception (unconscious
responses to noxious stimuli). Pain, as defined by the International
Association for the Study of Pain is purely a conscious experience, with a
sensory component and a component of emotional feeling (suffering). In contrast to this conscious experience of
pain, the unconscious detection, transmission and response to noxious
stimulation by lower levels of the nervous system is and defined as nociception - not pain. According to Sneddon, and associates, any
behavior that is a reflex would be evidence of nociception but any behavior
more complex than a reflex would be evidence of pain. This way of distinguishing pain from
nociception is invalid because there are clearly complex, non-reflexive
behaviors that can be purely nociceptive and unconscious. For example, humans
with extensive damage of the cerebral hemispheres can still make a complex of
responses including facial grimaces, vocalizations, struggling and avoidance
reactions to noxious stimuli, but they are unconscious and unable to experience
pain. From the definition of pain used by Sneddon and associates, it would be
concluded that these unconscious humans are feeling pain rather than making
purely unconscious, nociceptive responses, which is
clearly erroneous. There are many other
examples of complex, non-reflexive, even distress-like behaviors that can be
performed unconsciously. A person
having a night terror, for instance, will show a compelling fear-like display,
including a scream, terrified facial expression, elevated heart rate, sweating
and dilated pupils, even though they are unconscious and in such deep sleep
that they are difficult to awaken. The
point is that complex behavioral displays that seem to reflect distress can be
purely unconscious – even in humans. It
should not be hard to appreciate that the behaviors of which a fish is capable
could be unconscious as well.
2.
In order to show that a fish experiences pain, it is necessary to
show that a fish has consciousness.
Without consciousness, there is no pain.
None of the fish behaviors in this
paper require the involvement of consciousness and the authors don't even deal
with this essential issue. Furthermore, as I have shown in my 2002 Reviews in
Fisheries Science paper, there is extensive scientific evidence that pain and
consciousness depend on very specific brain regions, namely specialized
neocortex regions of the cerebral hemispheres.
These specialized neocortical regions perform the additional levels of
neural processing, beyond unconscious nociception, that make the experience of
pain possible. These brain regions are absent in fishes and there are no
alternative brain systems to perform the same functions. Consequently, there is no neurological basis
for assuming that a fish might have a capacity for consciousness or pain. Thus, the burden of proof that trout are
conscious and potentially capable of feeling pain remains on these
authors. They dealt with this issue only
by citing previous studies that also used invalid criteria for pain, such
avoidance learning, which actually occurs unconsciously. Only anthropomorphic speculation would lead
one to conclude that the trout in this study were experiencing pain.
The behavioral results
allegedly showing evidence of pain were misinterpreted.
1.
The behavioral studies were done by
injecting large volumes of one of three solutions: bee venom, acetic acid
solution or saline, into the jaw of rather small trout. For the sizes of the fish used, these injections of liquid would have
been equivalent to injecting 100 milliliters (more that 3 ounces) of solution
into the lip of a human. Bee venom
contains a great variety of toxins that affect the nervous system and cause a
hormonal stress response in addition to stimulating receptors signaling tissue
injury. In spite of the large dose of venom or acid, the activity level of
these fish was not affected, they did not hide under a shelter in the tank and
they resumed feeding in less than three hours. Furthermore, fish that received no injection
at all or fish that received a saline injection did not feed, on average, for
an hour and 20 minutes, showing that a large saline injection produced no more
effect than just handling. The acid and
venom-injected fish also showed an infrequent rocking behavior that may have
reflected a difficulty by the fish in maintaining an upright posture, given the
magnitude of the toxic chemical trauma created by the injection. But, even if the infrequent rocking was a
response to nociceptive stimulation of the mouth, there is no reason to believe
that it is any more than an unconscious nociceptive response, rather than an
indication of “pain”.
2.
Sneddon and associates also state that
the acid-injected fish rubbed their mouths against the gravel (they don’t say
how often), but the venom-injected fish did not. They concluded that mouth rubbing was an
indication of pain because mammals, including humans, rub injured tissues to
alleviate nociceptive input. If so, why
did the venom-injected fish, that were also supposed to be in pain, not perform
this behavior? In addition, injections of irritants into skin tissues is known to cause
hyperalgesia, where skin becomes hypersensitive, like the effect of a
sunburn. Who rubs sunburned skin against
gravel to alleviate the pain? At one
point in the paper, Sneddon and associates say that feeding was suppressed
because the fish were avoiding mouth stimulation, which would cause
“pain.” But later, they say that mouth
rubbing was a way of reducing “pain.”
These are contradictory interpretations and you can’t have it both
ways. Their interpretations of the
mouth-rubbing behaviors don’t make sense nor do they show conscious experience
of pain.
3.
One of the few effects actually
produced by the acid or venom injections was an elevated opercular beat rate
(breathing). This response could have
resulted directly from gill irritation due to leakage or blood borne spread of
the acid or venom injections, but even if increased opercular beat rate was due
to nociceptive stimulation of the mouth, this unconscious movement proves
nothing about conscious pain.
4.
One caveat regarding the behavioral
data described above is the fact that some of the statistical analyses were not
done correctly. Data for opercular beat
rate and for time to resume feeding were analyzed by one-way analysis of
variance, but conclusions were made about specific group differences in these
measures. With this type analysis, it is
not legitimate to conclude that one group (e.g. acid or venom injected differed
from any other group (e.g. handeled or saline injected), but the authors made
such conclusions, nonetheless. Given the
sizes of the standard errors of the means for these data, however, the group
differences reported by the authors would probably have been substantiated
following proper statistical analysis.
To summarize, the most impressive thing about the acid and venom
injections was the relative absence of behavioral effects, given the magnitude
of the toxic injections. How many humans would show little change in behavior or be ready
to eat less than three hours after getting a lemon-sized bolus of bee venom or
acid solution in their lip? Rather than
proving a capacity for pain, these results show a remarkable resistance to oral
trauma by the trout. It comes as no
surprise, then, that many anglers have had the experience of catching the same
fish repeatedly within a span of a few minutes.
Of course predatory fishes, including trout, feed avidly on potentially
injurious prey like crayfish, crabs and fish that have sharp spines in their
fins – which further indicates that these fish are not highly reactive to
noxious oral stimuli.
In addition, Sneddon and associates
claim to have presented the first evidence for nociceptive sensory receptors in
fish, but their results were neither wholly original nor unexpected. In my 2002 Reviews paper, I cited a 1971
study by Whitear that showed the presence of C-fibers in fish. C-fibers are a principal type of nociceptive
receptor, so there was very good reason to assume that trout would have
nociceptive receptors. Another technical
issue arises in the authors’ description of their procedure for decerebration
of trout in order to make them “insentient.”
The term sentience is vague and has no standard scientific meaning, but
apparently Sneddon, et al. were performing this decerebration in order to
eliminate any potential pain that they assumed was within the capacity of the
trout. The usual means of producing a
decerebration is to remove all brain tissue above the midbrain. According to Sneddon, et.
al, however, they removed the “…olfactory and optic
lobes and cerebellum…” This is peculiar
and counterproductive because the entire pathway for nociceptive information
from the periphery through the brainstem to the cerebral hemispheres would have
remained intact in these fish, since the “ofactory lobes” but not entire
cerebral hemispheres would have been removed according to this
description. If fish could feel pain, as
the authors contend (and I dispute), these fish probably would have.
The bottom line of this critique is that any attempt to show pain
in fish must use valid criteria, including proof of conscious awareness,
particularly a kind of awareness that is meaningfully like ours. This is not something that
can be taken for granted, because on neurological and behavioral grounds it is
so improbable that fish could be conscious and feel pain. Furthermore, the behavioral results of this
study show that in spite of very large injections of acid solution or venom,
the fish showed little adverse effect, hardly supporting the claim that they
were in pain.
I wish to emphasize that the improbability that fish can
experience pain in no way diminishes our responsibility for concern about their
welfare. Fish are capable of robust,
unconscious, behavioral, physiological and hormonal responses to stressors,
which if sufficiently intense or sustained, can be detrimental to their
health.
Cited
reference:
Rose, J. D. 2002. The
neurobehavioral nature of fishes and the question of awareness and pain. Reviews in Fisheries Science, 10: 1-38. This paper can be obtained in electronic form
from the author.
For another neuroscientist’s critique of the Sneddon, et al.,
article, see: http://www.spiked-online.com/articles/00000006DD91.htm