Book Review: Do Fish Feel Pain?

Betteridge's Law of Headlines, averted

I have learned what the gray aliens are doing when they abduct people and probe their anuses. They’re trying to test whether we’re sentient. I know this, because the experiments done to fish to test whether they’re sentient come straight out of an alien abduction testimonial.¹

Some nonhuman animals almost certainly experience pain. Pain is evolutionarily advantageous, as we can see in the case of humans with congenital insensitivity to pain, who often accidentally severely injure themselves. However, not every organism which has a negative reaction to certain stimuli genuinely has the subjective experience of pain. The plant Mimosa pudica recoils when touched, but it doesn’t feel pain. It doesn’t have a brain, so it has nothing to feel pain with. Humans also react negatively to stimuli without consciously feeling pain: if you touch a hot stove, you’ll pull your hand away well before you subjectively experience pain. Negative reactions to stimuli (with or without a pain experience) are called “nociception.”

Some people think that, like Mimosa pudica, fish can’t feel pain because they don’t have anything to feel pain with. Fish brains look “naked.” They don’t have a neocortex that humans use to feel pain. But different animals often use different structures to achieve the same goal. For example, both birds and bats can fly, even though only birds have feathers. So we had to study fish more closely (that is, do horrifying alien abduction experiments) in order to understand them.

In order to detect nociceptors in trout, the author of Do Fish Feel Pain? and her colleagues completely anesthetized the trout, put them in a special cradle that kept them upright, and washed an anesthetic solution over their gills so that they wouldn't wake up. They removed the skin and bone of the head so they could poke the brain, then removed the cerebellum and the olfactory and optic lobes. They applied a recording electrode to the trout's trigeminal ganglion, where the nerves from the face meet.

The fish is still alive at this point, to be clear.

Then the researchers gently applied a very small probe to parts of the face and checked whether there was an electric signal in the ganglion. They identified fifty-eight receptors. They poked each of the receptors with three different things: a von Frey filament (a hairlike strand of metal) to test their sensitivity to touch; vinegar to test their sensitivity to chemicals; and a narrowly focused quartz light to test their sensitivity to heat. They also dropped some water on each receptor to check that the vinegar response wasn't a response to a liquid being dropped on the receptor. Of fifty-eight tested face receptors, twenty-two were nociceptors. Some responded to all three stimuli; the rest, just two.

Later research discovered that the organization of the fish brain is very different from ours. From our perspective, the fish brain is inside out. The parts on the inside of the fish’s brain are on the outside of a mammal brain, and vice versa. Researchers (not the author of this book) gave goldfish specialized brain damage to see what would happen. Goldfish whose hippocampus-equivalents are no longer functioning can't solve mazes. Goldfish whose amygdala-equivalents are no longer functioning don't avoid things associated with electric shocks. These effects are specific to the area: goldfish whose hippocampus-equivalents are no longer functioning can still avoid things associated with electric shocks, and goldfish whose amygdala-equivalents are no longer functioning can still solve mazes.

Back to trout abduction. Next, the author and her colleagues poked trout² with noxious chemicals. Chemicals are good because you can use exactly the same amount of chemical on exactly the same spot on the fish and make sure that you’re giving the same injury with different fish. They injected the trout with vinegar and bee venom. Trout are stressed by being abducted by aliens; to control for this, researchers abducted two control groups of trout, one of which wasn’t injected with anything and one of which was injected with saline.

Trout, like humans, breathe quickly when they’re stressed. At rest, a trout’s gills beat at 50 BPM. After abduction, control trout’s gills beat at 70 bpm; treatment trout’s gills beat at 90 BPM. Fish in all groups sat on the bottom of the dark half of the tank.³ The trout injected with bee venom or vinegar rocked back and forth and made darting movements. The ones injected with vinegar also rubbed their snouts on the glass walls or the gravel in their tank. After about 80 minutes, control fish's gills had returned to 50 bpm and they were willing to swim to the scary bright half of the tank to get food. The treatment trout took three and a half hours to return to 50 bpm gill beating and to be interested in food.

Of course, this could all be purest nociception, like Mimosa pudica or pulling your hand from the stove. But further research by the same team suggests that’s not so. Trout are scared of new objects. It’s a pretty complex cognitive process to recognize that an object is new. Trout injected with saline and put into a tank with a Lego tower are like “aaaaaa! A Lego tower! Scary! Swim away!” Trout injected with vinegar are too distracted to notice the Lego tower. But trout injected with vinegar and morphine—to numb the pain—are scared of the Lego tower again.⁴

We’ve done other experiments that provide evidence that fish feel pain. For example, another researcher (not the author of Do Fish Feel Pain?) attached electrodes to goldfish with a long, lightweight lead so the goldfish could swim around. Whenever the goldfish went into a certain part of their tank, they were shocked. The goldfish learned to avoid this part of their tank, which suggests that they learn which situations cause them pain and avoid them.

Trout are capable of making tradeoffs. Researchers taught trout that they will be electroshocked if they enter a particular part of their tank. But trout are social animals who strongly prefer being around other trout. If the electroshock area is the area closest to another trout, the trout will move into the area with the electric shocks.

Other researchers attached electrodes to the skin of carp and electroshocked them an hour before and every five minutes for ninety minutes after they got a dose of tramadol. The carp have a weaker nociceptive response when they’re on tramadol. The strength of the response is dose-dependent. This suggests that the carps experience pain, which is eased by opiates.⁵

Many fish species have complex cognitive abilities. Maze-solving is important for understanding whether fish are sentient. In order to solve a maze, a fish must integrate their memories and their observations to create a world model that will allow them to achieve their goals. On some theories, the thing that consciousness is for is integrating information and using it to make decisions, so this is very important. Fishes are, in fact extraordinarily good at solving mazes. Experimentation shows, for example, that goldfish can navigate mazes using landmarks.

Frillfin gobies live in rock pools at low tide. When a predator approaches, they flip into a neighboring pool, even though they can't see the neighboring pool. At high tide, gobies swim around and look for the depressions that will become rock pools at low tide; at low tide they remember where the depressions are. Gobies can in fact learn the location of depressions with a single high tide of exposure. They plot novel routes that they’ve never used before and even do detours (going the “wrong way” for a bit in order to get to a desirable location).

Frillfin gobies are tiny, incidentally:

You might think there’s no room for a real brain in a fish that small, but they’re wicked smart.

Other cognitive abilities also matter. Cichlids and Siamese fighting fish can remember the identities of fish that they've watched fight. Cichlids are, in fact, capable of transitive inference. In one study, a cichlid "bystander" was allowed to watch other cichlids fight: A beat B, B beat C, C beat D, and D beat E. The cichlid was then placed in a tank with, say, B and D. The cichlid will reliably swim towards D, the weaker fish, even though they’ve seen both B and D win once and lose once.⁶ It’s not a status communication: a cichlid who hasn’t observed any fighting, if placed in a tank with B and D, is equally likely to swim towards either of them.

Some people say that fish don't feel pain because, when fished, they swim and pull away from the angler, which causes the hook to dig in more deeply into the mouth. However, many species (including humans) will endure horrible pain in order to escape.

The same fish can be hooked again and again, which some people argue means they can't learn. However, fish do learn to avoid artificial lures. Studies have shown that carp originally caught with a rod and line need to be caught with a net in the future because they have learned to avoid the rod and line. In areas where catch-and-release fishing is common, fish learn not to bite on hooks. However, if there's a lot of competition for food, fish can't be too picky about what food they eat. Experiments on whether fish learn not to bite on hooks happen on well-fed laboratory fish. Wild fish, who are likely to be quite hungry, make different tradeoffs.

In conclusion: while the subject has been shamefully understudied, the balance of the evidence is that teleosts (bony fish) have sophisticated cognitive abilities and are sentient. We should care about their well-being.

Do Fish Feel Pain? By Victoria Braithwaite. Published 2010. 194 pages. $19.

1. When I refer to “fish” in this post, I’m referring to teleosts and not e.g. sharks.

2. Different trout. As one might expect, the poor horrifying-brain-surgery trout was euthanized before it woke up from anesthesia.

3. Relatable.

4. Of course, morphine has a number of effects on behavior other than relieving pain. But opioids typically prevent behavior (like nociceptive reactions); they don’t actively cause the fish to do things.

5. Were all 18 of those post-tramadol electric shocks really necessary? I think it is a bit excessive to shock a fish 18 times even if the fish is high.

6. And so on for every other pairing.