Hand-hand-hand-hand-hand-hand-hand-hand-eye coordination

We’re smart. Octopuses are smart. But we have different kinds of smart.

Octopuses don’t process information like us. An octopus can tell -[ from ]-, but has a very difficult time telling < from >. There are plenty of task that we find trivial that are very, very hard for octopuses to do. (Many are shown in Wells 1978).

Gutnick and colleagues were interested in whether octopuses could integrate sight and touch. We do this all the time. Almost the entire video game industry depends upon us being able to do this. But octopuses? They’d probably suck at video games. They can learn to attack a particular object they can see. They can distinguish objects based on touch. Nobody had been able to get them to learn a task that needed both.

These sorts of behavioural puzzles are what led Gutnick and colleagues to build this maze.

To get a food prize, the octopus has to reach up into a tube and grab the food. The octopus can’t use any chemical cues to locate the food, because the food is not in water; the octopus has to lift its arm outside the water to get it. The octopus has to figure out which of three possible locations the food is in by seeing it.

Gutnick and colleagues confirmed that vision is the main cue the octopuses are using by blacking out the arms of the maze and letting the trained octopuses try to find the food. As soon as they do that, octopuses fail, doing no better than blind luck.

This task sounds easy... for a human, it is. But the octopuses in this study had to work a long time to get this down. None learned it in fewer than 60 trials. And one of the seven octopuses never learned the task. (I had similar experiences training octopuses; some never learned.) And they never got faster at it.

But the point is not that the animals are slow to learn; the point is that they can learn to do this at all. Several previous studies had not been able to get octopuses to integrate their sight and touch in this way. And at the neural level, there is little evidence for the brain of octopuses being arranged in sensory or motor maps like they are in many vertebrates.

The authors speculate a bit on the way that the visual and touch systems might be coordinated. Some octopus arm movements are purely “top down” affairs: the brain sends a command and the arm follows through, and sensory input from the arm doesn’t alter the movement very much.

The authors suggest that the behaviour they are seeing might be too complicated for that, and think that there is an interplay between the sense of sight and the sense of touch. They don't have any strong evidence for that yet, however. This might be testable by modifying the task so that the octopus was only able to see the food for a limited time, or to change the position of the food after the octopus had started to reach into the tube.

So maybe there’s hope that one day, this will be a reality:

References

Gutnick T, Byrne R, Hochner B, Kuba M. 2011. Octopus vulgaris uses visual information to determine the location of its arm. Current Biology: In press. DOI: 10.1016/j.cub.2011.01.052
Note: The article has a nice 4 minute video that appears to be freely available!

Wells MJ. 1978. Octopus: physiology and behaviour of an advanced invertebrate. Chapman and Hall: London.

Dumbo octopus photo by Anomolous4 on Flickr; used under a Creative Commons license.