Maze-arm length affects a choice criterion in the radial-arm maze

Male rats were tested in a 12-arm radial maze with 6 arms that were standard in length and 6 arms that were half the standard length. As previously reported by Brown (1990), revisits to short arms were more likely than revisits to long arms. Two explanations of this effect of mazearm length on choice accuracy were experimentally contrasted. The first attributes the effect to diminished discriminability of visited and unvisited arms when the arms are short. The second attributes the effect to a relatively lax choice criterion being applied to short arms. An analysis of the microstructure of choices, applying the logic of signal detection theory, provided evidence for the latter explanation.