Cognitive maps and spatial inference in animals: Rats fail to take a novel shortcut, but can take a previously experienced one

Previous work has shown that children are able to make a spatial inference about adjacent locations that have only been experienced indirectly (Hazen, Lockman, & Pick, 1978). We sought to replicate this finding in rats, on a conceptually analogous task. In a first experiment, rats (n = 8) were given 110 training trials on a task in which they entered a series of four square environments via connecting alleyways. Following training, we conducted a probe session in which the original training route was blocked and three novel routes were introduced, one of which led directly to the food reward. Surprisingly, rats failed to choose this shortcut route over the alternative routes. In a second experiment, following additional training with a series of platforms that were visible from one another, rats again failed to take a shortcut when given the opportunity to do so. In a third experiment with naive rats (n = 11), a shortcut was chosen, but only by rats that were given unrewarded preexposure to the shortcut route. These tests suggest that, despite their dedicated neural representations of location and direction, rats lack the capacity for a novel spatial inference. For rats, the use of a shortcut requires learning