| The Morris water maze is routinely used to explore neural mechanisms of working memory. Humans can often acquire working memory relevant to performing a task without having to actually perform the task. In the last two decades, little recent research has been done on rats' ability to acquire information in the water maze by a latent learning procedure, learning by sheer spatial observation as opposed to active trial and error. Much of the older research dealt with reference memory obtained by direct placement of the subject on the escape platform. While there were some indications that direct placement improved working memory, it was claimed that the resulting performance was not efficient (I.Q. Whishaw, Quarterly J. Exper. Psych. 43: 83-103, 1991). The present study differed from older research in using a different control procedure and in using different procedural training prior to working memory assessment.;We attempted to induce latent spatial working memory by direct placement of the rat on the platform, so that it observed its location relative to landmarks surrounding the tub. Prior to working memory assessment, n = 24 rats received three days of procedural training, swimming progressively longer distances to the platform. During working memory testing, the experimental group received direct placement on the hidden escape platform in the water maze testing room while the control group received direct platform placement in a separate pool in an irrelevant environment. Each day for six days, all rats were required to swim to differing platform locations, with either 5, 10, or 30 minute time intervals between placement and testing. Path lengths and escape latencies to the platform were determined. A mixed-model ANOVA of path length revealed a significant, p = .008, main effect where prior exposure to the platform location reduced swim path length to the platform. Direct placement also showed a similar reduction in latency to find the platform, p = .022. In both cases, there was no significant effect of placement-testing interval and no significant interaction effect. There was some evidence of efficient performance at the shorter retention intervals, since the latencies averaged less than a third of the corresponding latencies reported by Whishaw (1991). Also, the median path length at the 10 min. retention interval was only 15% greater than the direct path for the experimental rats, compared to 167% greater for control rats. It will be of interest to explore how such memory for latent learning is affected in models of normal and abnormal aging. |
