Way to work a room: How fish are like chickadees

You're trapped inside a rectangular room with four doors, one in each corner. You try the first door. It's locked. You try the second and the third door -- locked again. Finally the fourth door opens. You make a point of remembering which corner of the room it's in, which turns out to be useful, because before you know it you're trapped in another identical room. Now, how did you remember what door opened last time? You can't rely on your physical orientation because you might be facing a different direction.

Suppose one wall in both the rooms was a different color from the rest. Then it would be easy to locate the correct door based on this feature. If not, you'd have to rely on the shape of the room. Take a look at the two rooms below. If door 4 is always unlocked, then it should be easier to find in room B than in room A.

If you're placed in a series of rooms like room A, facing a different direction each time, then you wouldn't be able to tell the difference between door 4 and door 2: both doors have a short wall to the left and a long wall to the right. If the rooms always look like room B, it shouldn't take long to learn to pick door 4 every time -- it's always just to the left of the green wall. But still, even in room A, eventually you'd learn to avoid doors 1 and 3 -- they're never going to be unlocked.

Researchers have trained both humans and animals in similar tasks, and found a curious result: most animals only pay attention to the shape of the room, even though it provides less information about the location of the door than the colored wall does. The only exception that had been found until recently was the wild mountain chickadee, which responded better to features of the room than to its shape. There has been some speculation that the reason for this is the environment that the chickadees were raised in: if they've never been inside of a room, they don't know how to navigate based on room shape. The experiment hasn't been attempted on chickadees raised in captivity, however, so it's uncertain that this is the real reason for the difference.

A team led by Alisha Brown has now tried the experiment on a different animal entirely: the convict fish. They raised the fish from infancy in two different tanks: one rectangular, and one round, with no obvious features to use for navigation. After four months, when the fish were mature, they trained them to find the one open doorway in a brightly lit rectangular tank inside of a larger, round tank. The doorways could be closed with a transparent door, which previous experiments had shown were invisible to the fish. The round tank featured four ceramic bowls, one near each doorway, that offered the fish much-preferred shade, so they had an incentive to try to escape from the rectangular tank.

Half of the fish were trained in a tank with four white walls, while the other half were trained in a tank with three white walls and one blue wall. Once they could accurately find the open doorway, the testing phase began. During testing, all four doors were closed, and researchers recorded which door the fish bumped into first. The fish that had been trained in the all-white tank were given just one test -- the same white tank. As you might guess, nearly all the fish headed either for the correct door or the door opposite that was indistinguishable from it (in our example, door 4 or door 2). The rearing tanks -- circular or rectangle -- made no difference in the results.

But the fish that had been trained in the tank with one blue wall were given three different tests. In an all-white tank their results were similar to the white tank group. In a tank identical to the training tank, with one blue wall, nearly all the fish headed for the correct door, again, regardless of where they had been reared.

In one final test, the position of the blue wall was changed

Now the fish faced a dilemma. If door 4 had always been open in training, then which door was correct? The door to the left of the blue wall (#3), or the door with a long wall to the left and a short wall to the right (#4 or #2)?

This time, the results depended on where the fish had been raised:

Nearly all the fish raised in the round tank went to door 3 -- the door to the same side of the blue wall. While some of the fish raised in the rectangular tank also tried this door first, more than half tried door 4, or its geometric equivalent, door 2. For the first time, laboratory evidence confirms that the environment an animal was raised in can affect whether it uses the shape of the room or the features of the room to navigate.

Interestingly, all the fish could learn to use the shape of the room to navigate; the only difference in the two groups is that when given a choice between shape or features of the room for navigation, the fish raised in a rectangle tank were much more likely to use shape.

Brown, A.A., Spetch, M.L., Hurd, P.L. (2007). Growing in Circles: Rearing Environment Alters Spatial Navigation in Fish. Psychological Science, 18(7), 569-573. DOI: 10.1111/j.1467-9280.2007.01941.x