Do Bugs Feel Pain?

By Brian Tomasik

First published: . Last nontrivial update: .

Summary

Do bugs suffer? Does a fly caught in a spider's web consciously experience fear and pain? This piece aims to shed some light on that question by presenting quotations and references from a variety of sources. My personal conclusion is that we should give some weight to the possibility of bug suffering, especially until more evidence is available. Thus, considering the 1018 insects that exist at any given time, there is a huge amount of (potential) suffering in nature due to insects alone. We may also want to consider the ways in which humans impact insects, such as through insecticide use, although insecticides could potentially prevent more suffering than they cause if they avert vast numbers of future offspring that would have mostly died, possibly painfully, soon after being born. (Whether insecticides reduce or increase insect suffering on balance seems unclear. And of course, reducing insect habitat permanently would be more humane than simply spraying pesticides.)

See also: "How to Avoid Hurting Insects". One of the easiest tips is to avoid buying silk, since its production boils silk worms alive.

Contents

Suggestions of pain

Nociception vs. pain

Learning and intelligence

"Emotion"

"Consciousness"

Sleep-like states

Non-stereotyped behavior

Cognitive generalization

Menzel and Giurfa (1999), p. 718:

Liu et al. report their results on visual learning by Drosophila and its underlying neuronal substrate. They show that individual flies can do quite complex tasks. The authors first conditioned flies to associate visual patterns (the ‘conditioned’ stimulus) with the presence or absence of heat (the ‘unconditioned’ stimulus). The idea is that the animals should fly towards the appropriate patterns to avoid dangerous levels of heat. Their behaviour is ‘operant’, because the flight course that they choose determines delivery of the heat. The experiments were done under particular illumination conditions, which form part of the general ‘context’ in which the associations are established.

The authors next showed that the flies could ‘generalize’ this trained response to several other, different environmental contexts. As contexts are ill-defined stimuli, comprising individual features from many modalities, the change in context that the authors introduced was a change in the illumination between training and test: between white and monochromatic broadband light; between two monochromatic broad-band lights; and between constant white light and white light interspersed with ‘dark flashes’ (light is switched off for 200 ms). These changes did not affect the performance of the flies. These results show that context generalization — rather than context specificity — guides the insect’s learning. But when the authors impaired the fly’s normal brain function by eliminating the mushroom bodies (a central brain structure), they found that retention of the trained pattern was strictly bound to the context during learning, and that the flies did not generalize to other contexts.

From "Exploring Consciousness through the Study of Bees" by Christof Koch:

Insects, in particular, were long thought to be simple, reflexive creatures with hardwired instinctual behaviors. No more. Consider the amazing capabilities of the honeybee, Apis mellifera. [...]

In humans, the short-term storage of symbolic information—as when you enter an acquaintance's phone number into your iPhone's memory—is associated with conscious processing. Can bees remember task-relevant information? The gold standard for evaluating working memory is the delayed matching-to-sample (DMTS) paradigm. The subject looks at a picture for a few seconds. The test image then disappears for five or 10 seconds. Subsequently, two pictures are shown next to each other, and the animal has to choose, by pushing a lever or moving its eyes, which of the two images was the test picture. This test can be carried out correctly only if the animal remembers the image. A more complex version, the delayed nonmatching-to-sample (DNMTS) task, requires one additional processing step: choosing the opposite image from the one previously shown.

Although bees can't be expected to push levers, they can be trained to take either the left or the right exit inside a cylinder modified for the DMTS test. A color disk serves as a cue at the entrance of the maze, so that the bee sees it before entering. Once within the maze, the bee has to choose the arm displaying the color that matches (DMTS) or differs from (DNMTS) the color at the entrance. Bees perform both tasks well. They even generalize to a situation they have never previously encountered. That is, once they've been trained with colors, they "get it" and can now follow a trail of vertical stripes if a disk with vertical gratings is left at the entrance of the maze. These experiments tell us that bees have learned an abstract relation (sameness in DMTS, difference in DNMTS) irrespective of the physical nature of the stimuli. The generalization to novel stimuli can even occur from odors to colors. [...]

Although these experiments do not tell us that bees are conscious, they caution us that we have no principled reason at this point to reject this assertion.

Neural oscillations

Prieto-Godino and de Polavieja (2010):

Electroencephalogram and local field potential (LFP) oscillations generally indicate periodic coherent synchronization of neuronal assemblies [1]–[3]. Oscillations have been found in systems as disparate as mollusks [2], moths [4], locusts [5], [6], rats and mice [7], [8], suggesting a fundamental role in computations carried out during higher-order processing. Oscillatory and synchronized activities in the mammalian brain have been correlated with distinct behavioural states or the execution of complex cognitive tasks, and are proposed to participate in the ‘binding’ of individual features into more complex percepts [3], [9]–[11]. Similar oscillations have been observed in LFP recordings from the first and second relay centers for olfactory information in insects. [...] However it is not known whether and how the oscillations observed in insect sensory systems and the ones recorded from mammals during the execution of complex cognitive tasks are functionally and computationally related. A first step in this direction is to find out whether oscillations occur in higher brain structures of insects during the performance of complex cognitive tasks, involving processes such as selective attention [14], contextual generalization [15], or formation of ‘sameness’ and ‘difference’ concepts [16]. Recently two reports have correlated in flies LFP oscillatory activity recorded centrally in the brain with different behavioural states, as it happens in mammals [17], [18]. These studies found that conspicuousness of different visual objects modulates the oscillatory activity recorded from central brain structures in the 20–30 Hz range. [...]

The fruit fly's LFP responses share several key features with physiological correlates in the 40–60 Hz range of visual selective attention in monkeys and humans [31]. For example, amplitude increases with salience, which can be increased by either an unconditioned stimuli or by novelty. [32]. Furthermore, a recent study in rats reported that, during difficult odour discrimination tasks, strong synchronous oscillations appeared in the olfactory bulb field potential. These oscillations were not present when the task was easy and could reflect an increased attention [33]. Therefore in flies, as it has been suggested before in mammals [31], neural synchronization may be a common neural mechanism involved in complex cognitive tasks like arousal, perceptual integration, and attentional selection. Our findings in the olfactory system, together with the findings of others in visual attention [18], indicate that despite the fruit fly's lack of neuroanatomical homology with primates, Drosophila might have analogous mechanisms of establishing salience and directing selective attention to its environment.

Social behavior

What about tissue damage?

See also: "Which Stimuli Are Painful to Invertebrates?"

Benefit of the doubt?

From entomologist Jeffrey Lockwood, "Do bugs feel pain?":

So, given that we can't be sure whether insects experience pain, how should we treat these creatures? When I was teaching insect anatomy and physiology I insisted that the students anesthetized insects before conducting experiments that we would expect to inflict pain on a mouse. [...]

[One reason is] it seems ethically obligatory to guard against the possibility that insects feel pain. If we use anesthetic and it turns out that insects don't experience pain, the material cost of our mistake is very low [...]. However, if we don't use anesthetic and it turns out that the insects were in agony, then the moral cost of our mistake is quite high.

In general I agree with this sentiment. As a matter of detail, I would not use exactly this "precautionary principle" approach of giving insects the benefit of the doubt. I would instead multiply possible insect suffering by a probability of sentience; this makes the ethical tradeoffs between insects, which may or may not suffer, more fair against animals we know can suffer. However, given that insects have a probability of sentience that isn't too small, their potential suffering still tends to dominate calculations even when multiplied by 50%, or 10%, or even 1%. My own probability for sentience is ~40%.a Considering how easy it is to avoid causing harm to so many insects, on a macro scale and even in our daily lives, any reasonable sentience probability will imply significant consequences for our actions.

Further reading



Footnotes

  1. I now don't believe that whether insects are sentient is an objective factual question, so talking about probabilities isn't quite accurate. Fundamentally, whether we care about insects is a moral choice. However, facts are very relevant in informing us about (a) what abilities insects actually have and (b) what types of abilities in humans correspond to conscious emotion. For instance, we might have thought that a given cognitive trait was important for human emotional experience but later learn that it's completely irrelevant, and in that case, the fact that insects have it would not be as pertinent as we had presumed. And the same could be true in the opposite direction. So one way to continue using probabilities is to ask "How likely would I be to care about insects if I learned and thought more about the topic for a long time?"

    Beyond this, I think sentience should be seen to come in gradations, in which case the question is less whether insects matter or not in a binary fashion but how much they matter. I think it's very likely I would care to some degree about insects upon further reflection, but whether I would care a little or a lot remains uncertain. At the moment I would guess that I value preventing one dog from suffering at around the same as preventing ~100 insect from suffering in an analogous way. This assessment is likely to change over time.

      (back)