Spatial memory and stereotypic behaviour of animals in radial arm mazes.

The symmetric radial arm maze, described by Olton in 1976, has developed into an important tool for the study of spatial memory. In a typical test an animal is placed in the centre of the maze, which contains some small piece of food at the end of each arm. The sampling behaviour of the animal is then recorded. In such studies the score (number of choices of arms which still contain food) of the animal is normally compared with the score of an imaginary animal which changes arms entirely at random. In a new method of analysis the non-random score of the animal is split into two parts, one depending on memory and one on stereotypic choice behaviour. Even mild departures from randomness are shown to alter considerably the expected 'random' score in an eight-armed maze. The part of the score claimed to depend on memory was shown to increase when the animals learn to search the maze, the stereotypic part did not. The general effect from stereotypic choice behaviour is shown to result, in most animals, in an increase in the total score. In an eight-armed maze this increase may amount to more than 20% of the total non-random score, even in a well-trained animal. The effect is less pronounced in mazes with 16 arms. It has been proposed that hippocampal lesions produce a stereotypic behaviour. We propose, based on our analysis, that the stereotypic behaviour is not produced but revealed by hippocampal lesions which destroy almost completely the memory-guided behaviour masking the stereotypic behaviour in the intact animal.