The octopus, the maze, and why it matters: behavioral flexibility and sensory-motor integration

Shallow-water octopuses are generalist predators – this means that they can eat a variety of other animals – and good ones too. They have a few different hunting strategies, with the commonest ones involving the octopus groping along the reef, feeling for food with its arms (although octopuses have been reported to hunt by ambushing (pdf link) as well, striking their prey after spotting it.) You can see the groping strategy at work in this video:

It is clear from previous research that octopus arms are capable of movement, even relatively complex movements, on their own. Thus, when an octopus gropes its way around a reef, it might be that it’s central nervous system is doing very little to control its arms; rather, it seems likely that they move mostly “on their own”. Tamar Gutnick and her colleagues at the Hebrew University of Jerusalem recently published a study that investigated if and how octopuses (of the species Octopus vulgaris) can use information from their central nervous systems to control the movement of a single arm. I’ll let them tell you about it:

(By the way, I love video abstracts/experiments. Thanks, guys!)

The researchers took 7 octopuses and trained them to reach into a clear plastic “maze” where they could choose to put their arm into one of three areas. One of the arms of the maze had a piece of food in it. Since they were only given one chance in each session – if they chose the wrong arm in a session, they weren’t allowed to try again – the octopuses learned to find the food by looking at it through the clear walls of the maze and then make the appropriate arm movements to get it. After the researchers covered the clear maze with masking tape, the octopuses, who could no longer see the food, weren’t able to do the task any more – they got about 1 in 3 trials right, exactly what you’d expect if they were choosing randomly.

The results of this study tell us that octopuses can use visual information to direct the movements of their arms, and that they seem to get more accurate with practice. What we don’t know, however, is how an octopus’s brain could pull this off. It’s clear that simple movements are controlled within the arms themselves, as a disembodied octopus arm can make some movements by itself, but it’s unclear how the “higher-up” parts of the brain that receive visual information from the eyes could mix it with tactile information from the arm to direct these sorts of movements.

The skeptic in me says that there might not be much to be excited about. After all, we’ve known that that octopuses use their vision to do things like find their way around, and size up potential food/predators for a long time. The procedure used, even if it’s new, is sort of limited; it’s essentially a simple detour task, where the animal can see its reward but has to take a complicated route to get to it. As Zen Faulkes pointed out in his post on this study, (which is so cleverly titled as to put me to shame), the octopuses weren’t even very good at learning such an apparently simple task. Compare this to the scores of learning tasks that other laboratory animals like rats (and people, for that matter) whiz through, and it seems like a small step. Some experiments using tasks like this fail while others succeed, and there’s no clear consensus as to how and why octopuses learn (or fail to learn) in certain situations, making it even harder to say anything about how octopuses learn.

Nevertheless, there’s some room to be excited; it’s a small step into an mostly unexplored field. Think about just how foreign an experience this was for the octopuses in the experiment – not over the time scale of the experiment, but over evolutionary time. For millions of years, the ancestors of this species have been hunting on the seafloor in shallow waters, where it’s very unlikely that they’d ever encounter a hard, transparent surface that they might have to move around to get food. Even still, when they’re presented with such a situation, they can navigate it, even if they do it with some difficulty. The behavior of these octopuses, then, seems to me to have evolved not only to work well in a specific situation, but to work (at least minimally) in a wide range of situations – their behavior has evolved to be somewhat flexible. In fact, this is a strategy that is used by all animals that can learn (which seems to be most of them) that helps them deal with the fact that there is no such thing as a perfectly stable and predictable environment, and that behavior needs to adapt to deal with this. For example, your ancestors (if you were an octopus) might have fed on a few specific species of crab for the past few hundred years – if something about the environment changes, you need to be able to learn to hunt something else, or you (and your species) are doomed. Looking at it in this light, it’s not very surprising that a laboratory filled with mazes and puzzles built by scientists would push the limits of a cephalopod’s behavioral flexibility – this is a huge change from the environment the animal evolved in. To quote Zen Faulkes, “the point is not that the animals are slow to learn; the point is that they can learn to do this at all.”

This research is also exciting because it begs questions about how the nervous system of the octopus can do this task. In more familiar research animals (that is, mammals), we know that specific parts of the brain (areas of the motor cortex) control the contraction of specific muscles. Besides this, we’ve identified a whole host of brain structures that play various roles in putting together these movements and in using information from the muscles, skin, and eyes to control and refine them. In mammals, both motor and sensory systems are put together in a such a way that their arrangement in the brain corresponds to their arrangement in the body – this is called somatotopy. (Check out this neat little demonstration of the concept by Jaakko Hakulinen.)

According to another study published in 2009 by researchers from the same university, this doesn’t appear to be the case with the octopus. The investigators in that study couldn’t find any clear relationship between activity in different parts of the octopuses’ brain and different movements. While we know where the information from the eyes goes in the octopus brain (to the sensibly named “visual lobes”,) it’s unclear where it goes from there or how it might interact with the neurons that control the arms, or how this information might be put together with sensory information from the arms. How exactly an octopus’s brain uses vision to control ongoing movements, then, is the most exciting kind of scientific problem: an unsolved one.

Thanks for reading! Zullo, L., Sumbre, G., Agnisola, C., Flash, T., & Hochner, B. (2009). Nonsomatotopic Organization of the Higher Motor Centers in Octopus Current Biology, 19 (19), 1632-1636 DOI: 10.1016/j.cub.2009.07.067

Gutnick T, Byrne RA, Hochner B, & Kuba M (2011). Octopus vulgaris Uses Visual Information to Determine the Location of Its Arm. Current biology : CB, 21 (6), 460-2 PMID: 21396818

Trackbacks / Pingbacks

  1. News » Blog Archive » Editor’s Selections: Colin Firth, Beer, and Octopodes
  2. The latest from around the Gam April 14, 2011 « The Gam

Leave a Reply