Specificity of memory measures in an adjusting delayed nonmatching-to-position task for rats

While the delayed nonmatching-to-position (DNMTP) behavioral paradigm has often been used by neuroscientists to assess working memory in rats, its measure of working memory is compromised by floor and ceiling effects. Specifically, these floor and ceiling effects undermine the ability to detect a significant interaction in a two-factor repeated-measures ANOVA, which is required in order to conclude that impaired performance has resulted from disrupted working memory and not from a change in another psychological process (i.e. that the impairment is 'specific' to working memory). The present study was conducted to evaluate if these limitations could be overcome in a DNMTP by adjusting the length of time that the rat was required to remember (the 'delay') so as to avoid the floor and ceiling. The general procedure for two experiments presented subjects with trials where there was either a minimal (1 s) delay or a longer delay of varying length, with the goal of maintaining nonmatch-to-position accuracy near 75%. The procedure was such, that if the average accuracy was at or above 75%, then the next trial would be a long-delay trial. If it were below 75%, then the next trial would be a 1 s delay trial. In the first experiment, the subjects were presented with trials where the value of the longer delay was systematically varied between-sessions. This was done to simulate the faster rate of forgetting found in persons with amnesia. DNMTP accuracy diminished at the longer delay but not the short delay as the second interval was lengthened. However, other measures, including accuracy at the 1 s delay, discrimination accuracy during the sample phase, and the number of trials completed per session, did not change. This experiment suggested that this DNMTP could precisely measure small changes in the rate of forgetting. In a second experiment, a potential non-mnemonic confound, the level of motivation, was directly manipulated by giving the water-restricted subjects access to water immediately prior to the start of the session. The number of trials completed per session diminished, but the accuracy at both the 1 s and the longer delays did not change. These results indicated that the measures of working memory in this DNMTP were insensitive to changes in motivation. Together, these experiments confirmed that adjusting the delay in the DNMTP improves the validity of the measures in this task.