Enrichment in Captive Cephalopods

To get things started, here’s a video of an octopus with a Mr. Potato Head Toy (and other things):

You’ll see why this is relevant in a minute. Now on to the post!

“Enrichment” is a psychological term that’s been thrown around a lot. It’s become a buzzword in publications about education, perhaps rightly so given its huge impact on the field of developmental psychology. It has been the subject of intense study in psychology, and continues to be the subject of study. But what exactly is it that we are referring to when we use the term “enrichment”?

Before I get to that, let’s take a step back and consider how brains work. Brains, in the sense that I am talking about, are simply big networks of neurons (there are other types of cells, but we can ignore them for now.) These neurons get sensory input from the body, talk amongst themselves, and send out signals that cause the organism to behave in a certain way. This is, in a nutshell, what brains do – they generate behavior. Importantly, though, they don’t just generate any behavior at a given time; they generate the appropriate behavioral response to the immediate situation. Even more impressively, in some animals, they generate the response that will lead to a positive outcome in some hypothetical future situation (for example, when birds hide food to recover later.) Thus, brains work because they can process incoming sensory information into relevant and adaptive behaviors. Neurons can do this because they are connected to each other in complex but well-controlled and highly-specific ways. Much in the way that an electrical device will only operate if all of its components are connected correctly, brains will only work if all of their neurons (or at least most of them) are “wired” together into functional circuits.

It may come as no surprise, then, that brains only wired the way they are because of the sensory input they receive. During brain development (and, to a lesser extent, throughout life,) neurons require sensory input and behavioral output to form proper connections as well as get rid of improper connections. At the cellular level, this phenomenon is called activity-dependent or experience-dependent plasticity. One can think of it this way: connections between neurons (called synapses) that are used a lot and produce functional behaviors get stronger; that is, they become more efficient and transmit larger signals from neuron to neuron. Synapses that are not used, or that don’t contribute or contribute negatively to the function of a certain circuit become weaker; that is, they become smaller and less efficient, and may even disappear altogether.

This can be demonstrated directly by looking at neurons that process incoming visual information from the eyes. Suppose we were to raise one group of animals (lets say, ferrets) completely in the dark, and another, otherwise identical group of animals normally. Then, we kill both groups of animals and look at the neurons in their visual system (for example, in visual cortex.) The group that was raised under normal light will have, by definition, normal connections between the neurons in the visual system – specifically, these neurons will form functional circuits that allow animals to sense and respond to their environment. In the animals that were reared in the dark (called, sensibly, “dark-reared” animals,) we’ll see a host of abnormalities – cells will have the wrong shapes, end up in the wrong places, and have the wrong connections. Besides this, we could demonstrate some deficiencies in visual perception in the dark-reared animals. Taken together, these results, which have been replicated in a variety of vertebrates, strongly support the idea that sensory systems need input during development in order to develop properly.

It’s a small logical step to tentatively extend the principal of experience-dependent plasticity to brain functions that are more complex than immediate sensory analysis. In the mammalian cortex (which is involved in such complex behaviors as navigation, auditory and visual perception, social behavior, speech, thought, and so forth,) normal development is dependent on the experience of the animal in question during development. The input that these brain areas deal with is almost unthinkably complex. Their function often involves the monitoring the interaction of large parts of the organism with the environment over an extended period of time. For example, cortical areas involved in directing complex hand movements (say, playing a musical instrument) receive their input in the context of a feedback loop that involves planning a movement, executing it, and then determining whether the movement was successful and, if it was not, how it needs to be changed in order to be successful the next time. More abstract cognitive processes become even more complicated, as the brain areas responsible for these have to take into account such complex “stimuli” as the animals current and previous emotional states, inferred emotional states in other animals, the value of various hypothetical outcomes to the organism, prior cognitive states the animal has had, and so forth.

Here’s where the idea of environmental enrichment comes into the picture. The logic goes like this: because neural development is facilitated by the nervous system’s interaction with the environment, and because the more integrative areas of the brain (like the cortex) interact with the environment in very complex ways, normal brain development requires interaction with a very complex environment. If this theory is true, animals raised in very restricted environments (we might call them “environmentally impoverished”) should show maladaptive or sub-normal behavior as adults. In fact, this prediction has been borne out in many studies, which find that (among other things) environmentally impoverished animals (and, although it has been less directly demonstrated, people) have altered learning abilities, exploratory behavior, cognition, stress reactivity, and social behavior.

A corollary of this theory is the idea that nervous systems which have evolved to deal with the environment in very complex ways function abnormally when they do not have complex environments to deal with, much like the way that muscles atrophy if they are not used. Part of successfully engaging a complex environment is continually interacting with and exploring it – indeed, many animals have evolved a propensity to explore and manipulate their environment. Thus, animals with complex behavioral repertoires require complex environments to interact with in order to maintain normal brain function. This theory predicts that a lack of complexity in some animals’ (or people’s) environments should lead to aberrant behavior, and that such behavior can be corrected by providing an appropriately complex environment to those animals (or people.) This prediction is borne out in a variety of studies that show that, in captive adult animals, providing more complex environments leads to lower stress levels, less aggression, and fewer pathological and stereotyped behaviors.

Dr. James Wood’s article, “Environmental Enrichment in the Giant Pacific Octopus; Happy as a clam?” makes the case that aquariums should provide environmental enrichment for captive octopuses. I won’t follow his arguments exactly, but I will examine some of the same questions that he covers, and make reference to and critiques of his work as I find it relevant.

The big question is this: Should we provide environmental enrichment to captive cephalopods? Let me rephrase this: do the benefits of providing environmental enrichment to cephalopods justify the costs incurred by doing so?

First, let’s consider the possible benefits. Anderson and Wood point out several:

1. Captive animals should be kept healthy and allowed to behave normally. Behavioral health is as important to the longevity and quality of animal’s lives as is physiologic health. The real question here is whether cephalopods have complex enough behavior that they might show pathological behavior in response to captive conditions. Put in a more cognitive frame: are cephalopods smart enough to be hurt by captivity and, consequently, to benefit from enrichment?

Indeed, it is hard to tell whether cephalopods could benefit from enrichment. To the question of how we can tell if giant pacific octopuses could benefit from enrichment, Anderson and Wood succinctly conclude: “Simply put, we cannot.” Because cephalopods have evolved under predation from fishes, they reason, they spend most of the time during which they are not hunting hiding in their dens. It’s hard to tell if this is “good” for the octopus (after all, it seems like it would be nice not to have the daily stress of fleeing from predators and risking death during hunting) or bad (because, if cephalopods can become bored, this behavior certainly seems like it would be really boring.) In addition, little is know about how behavioral pathology might look in cephalopods, because we know so little about their behavior in general.

In sum, cephalopods may or may not benefit from enrichment. Given how complex their behavior appears to be, and the likelihood that at least some cephalopods possess cognitive capacities to rival at least some vertebrates, it seems like there’s a reasonably good chance that they could. This isn’t a convincing case in itself – it really depends on the costs of providing enrichment to cephalopods. If the costs are low, this might be enough of a reason to do it; with increasing costs, it becomes an increasingly bad bet.

2. Animal enclosures should meet the expectations of the public. Showing the public what they want to see – which is, it seems, natural-looking enclosures that animals can interact with) will lead to financial success and public support for zoos and aquariums. Researchers, institutions, and the enterprise of biological science in general stand to benefit from the increased public support that comes with presenting a public-pleasing image of animal husbandry in research. Also, if the public sees naturally-behaving animals instead of pathologically-behaving animals, they learn more about the animal’s behavior. A primary function of zoos and aquariums is to educate the public about animals, and so the potential to improve on this education is an important possible result of providing enrichment to cephalopods.

This is an interesting idea to me. Octopuses are popular aquarium animals, but not very popular research animals. The public face of animal research is usually mammalian, including rats, mice, rabbits, and primates. It’s doubtful to me that researchers would benefit from an improved public image if all octopuses used in research were given lots of enrichment. In zoos and aquariums, though, this seems like a valid concern. Indeed, it could offset the monetary costs of providing enrichment for these institutions, which allows less tangible benefits (like the possibility of relieving the suffering of bored octopodes) to be given more weight in deciding whether or not to provide enrichment for captive cephalopods.

3. Finally, zoos and aquariums often care for animals with the goal of eventually releasing them into the wild. Animals whose behavior is dependent on learning need to practice skills that will allow them to succeed, such as hunting, defending from predators and rivals, and maintaining good relationships with other individuals of their species. The Seattle aquarium, for example, does this with their giant pacific octopuses. They catch individuals, hold and display them for a few years, and then release them into the wild so that they have a chance to breed.

It’s doubtful to me that enrichment benefits cephalopods in this way. I can’t say that it’s impossible, but it seems unlikely. While cephalopods are able learners, the basics of cephalopod behavior – feeding, escape, and mating – appear to be largely innate. For example, octopuses learn things in order to adapt to the specific micro-environments they end up in. The characteristics of these environments are impossible for aquariums to predict, and so they cannot be simulated. Feeding a cephalopod common local prey species so that it learns how to eat those species efficiently might help it succeed if it is released, but besides this one example there seems to be relatively little to do in the way of “preparing” a cephalopod for release.

Before I hash out some of the costs of providing enrichment to captive cephalopods, let’s consider what providing such enrichment entail. Anderson and Wood suggest several ideas for giant pacific octopuses, most of which could work with other species of cephalopods. The one that sticks out to me is feeding the animals a variety of live prey. This would provide them with a good deal of activity, a variety of problems to solve (how to catch and eat different species of prey,) and is very entertaining for the public to watch. Some cephalopods (cuttlefish come to mind – I’ve read cuttlefish keepers complaints about this) will only reliably eat live food, anyways.

The authors also suggest that the octopuses be given objects to explore, which can be smeared with fish drippings or have food hidden in them to attract the octopus and encourage exploration. In my favorite quote of the paper, they recount a particular use of this technique:

Wood and Wood… hid food in a play ship – the octopus had to “sink” the ship to get the proffered food. Such a demonstration with a large octopus… would interest the paying customers of a public aquarium by invoking a “sea monster” image.”

The future of cephalopod husbandry.

They also suggest the use of “training” to provide enrichment to octopuses (I put “training” in scare quotes, because some of their specific suggestions, such as smearing fish-smelling fluid on parts of the tank, don’t necessarily involve learning.) This seems like a moot point to me. There are two ways in which you can train a cephalopod – by rewarding it with food, or by punishing it with electric shocks (or some other unpleasant stimulus.) In the first case, you might as well simply give the animal the food, the capture of which seems like it would provide most of the activity inherent in training. In the latter case, well, who’s going to argue that training an animal by electric shock improves its quality of life or reduces its suffering (assuming that the training isn’t absolutely necessary, such as teaching it not to attack people or run into traffic, etc.)?

What would be the costs of providing this sort of enrichment to cephalopods? In my estimation, they would be small. Here’s my reasoning:

It should be relatively easy for aquariums, especially those near coasts, to obtain live prey to feed to cephalopods. By advertising the feeding times (as is done with the animals like sharks, dolphins, and sea lions) aquariums could turn this into a way to entertain visitors and draw more business. The increased popularity of octopus exhibits would likely make up for the extra cost of providing more live food. One of the complaints I hear about the giant pacific octopus exhibit at the Niagara Falls Aquarium is that it just sits there all the time. In the interest of furthering public interest in cephalopods, I make sure to write “Please publicize octopus feeding times!” in their guestbook each time I visit. I’m sure that if they did this, their visitors would be more interested in the octopus and happier with the aquarium overall.

Similar reasoning applies to constructing interesting enclosures that encourage cephalopods to explore. The public likes to see animals doing things, and the increased public interest will more than make up for the expenses of outfitting an octopus tank (which can be as cheap as a few plastic toys) These forms of enrichment are probably very cost effective, at least for public aquariums, even if they have only a small chance of benefit the captive animals.

When we consider animals used for research, the costs become larger. Giving animals a less monotonous environment may exaggerate inter-individual variation. It is usually argued that enrichment produces consistently healthy experimental animals, and so reduces variation in experimental results. The research that is used to make this argument in the case of mammals, however, has not been replicated in cephalopods. Without such evidence, it’s hard to say what effect different kinds of enrichment would have on behavioral experiments with octopuses. Such evidence would be rather expensive and time-consuming to obtain, and providing enriched environments to experimental cephalopods on the assumption that it would improve results could be, if that assumption were wrong, very costly as well. If behavioral research on cephalopods becomes more popular, this will become a more urgent question, and somebody will have to take on the task and expense of answering it.

Generally, providing enrichment for captive cephalopods seems worth it. Given the (even relatively slight) chance that they could benefit from basic environmental enrichment and the small cost of such enrichment, there’s no reason not to do it. The deal only becomes sweeter when you take into account the benefits to aquarium popularity and public education. Even if the cephalopods don’t benefit from it, it can hardly hurt.

Thanks for reading!

Anderson, R., & Wood, J. (2001). Enrichment for Giant Pacific Octopuses: Happy as a Clam? Journal of Applied Animal Welfare Science, 4 (2), 157-168 DOI: 10.1207/S15327604JAWS0402_10

van Praag H, Kempermann G, & Gage FH (2000). Neural consequences of environmental enrichment. Nature reviews. Neuroscience, 1 (3), 191-8 PMID: 11257907


  1. Al Dove says:

    Thanks for a thought provoking post, Mike. I have a couple of observations (no special theme), based on experiences working at a public aquarium and being married to an animal trainer.

    Part of the challenge in deciding whether or not to enrich rests on determining what represents a “bad” behavioural sign for an animal. In the case of bears and elephants pacing and swaying (respectively), the repetitive behaviours are fairly clearly undesirable, but if an under-stimulated octopus just sits there, how do we know if that’s bad or not? Indeed, it may be more stressful to them to continually be poked or have their space invaded by strange objects, even if they are smeared with fish juice.

    Training as enrichment is a more recent idea that seems to work well based on evidence from marine mammals that choose to be trained over getting their food without ‘working for it’. You can safely discount the idea of training by punishment – no decent aquarium would do that, nor should they be allowed to. Essentially all training in aquariums happens by positive reinforcement; temporary witholding of the attention of the trainer (LRS) is about as bad a negative reinforcement as animals get.

    Enrichment to maintain naturalistic behaviours for future release is only a reality for a small fraction of aquariums. You mentioned Seattle – they are fortunate to be near the natural range of GPO’s, but most aquariums are committed to lifelong care of their cephalopods, which generally isn’t very long anyway, so I don’t consider this a major reason for enrichment.

    Ultimately the costs of enriching captive cephalopods are low, at least at the aquarium scale (lab animal facilities might be different). The biggest challenge to me seems to be in balancing the animal’s desire to remain hidden and still for large amounts of time, with the public’s expectation that a healthy animal is an active animal moving around a large exhibit. You can provide an octopus with the biggest tank in the world and she will still likely hide in an impossibly small cave for 95% of the time! There’s a dynamic tension, then, between exhibit designers, marketing folks, husbandry staff and veterinary staff that hopefully meets in a good middle ground where the animal remains happy and the public gets to see them and thus achieve the point of the exhibit in the first place.


    • Thanks for the thoughtful comment!

      “Part of the challenge in deciding whether or not to enrich rests on determining what represents a “bad” behavioural sign for an animal.”

      I entirely agree. Know what an animal does (and what that means) is a prerequisite to determining whether any intervention actually worked. It might turn out (if it ever gets studied) that the differences between pathologically stressed behavior and normal behavior in cephalopods are very subtle or obscure – this would be one good explanation as to why this distinction can’t be made for cephalopods yet. Ethology is useful.

      “You can provide an octopus with the biggest tank in the world and she will still likely hide in an impossibly small cave for 95% of the time!”

      It is difficult to say what this actually means. How important is that 5% of the time the animal spends roaming to its “well-being”? Can we alter the habitat to reduce it to 3% with no harm? 0%? How would we even figure out if cephalopods were being harmed? There are lots of unanswered questions.

      I think that specific problem could be investigated by putting cephalopods in smaller and smaller containers until it interfered obviously with their behavior. We could then see what sort of abberant behaviors developed, and try to detect those behaviors in the ethograms of cephalopods in various captive situations.

  2. Thanks very much for posting a lot of this awesome content! I am looking forward to checking out more posts.

Trackbacks / Pingbacks

  1. Quick Links | A Blog Around The Clock
  2. ResearchBlogging.org News » Blog Archive » Editor’s Selections: Enrichment, Ballerinas, Salmon, and Telepathy
  3. Best of the Network « Arthropoda
  4. Editor’s Selections: Enrichment, Ballerinas, Salmon, and Telepathy [The Thoughtful Animal] » iThinkEducation.net!
  5. Best of the Network: September through October « The Gam
  6. This Week in Cephalopods
  7. Open Laboratory 2010 – submissions now closed – see all the entries | A Blog Around The Clock

Leave a Reply