Quantifying nonmnemonic strategies in two-alternative forced-choice delayed matching and nonmatching-to-position...

Delayed matching-to-position (DMTP) and delayed nonmatching-to-position (DNMTP) tasks are widely used as standard paradigms for evaluating working memory in rodents. In these tasks, which employ a two-alternative forced-choice (2AFC) design, animals initiate each trial by nosepoking a back key, which triggers the illumination of a sample key (either the left or right front key). Upon nosepoking the sample key, a delay begins. During this delay, the back key remains illuminated and animals may nosepoke it, but these responses have no programmed consequence. After the programmed delay, the first nosepoke on the back key activates both front keys. Reinforcement is contingent on choosing the same key as the sample in the DMTP task, or the opposite key in the DNMTP task. Performance is typically measured as the proportion of correct responses across varying delays, which generally decline as delay increases. However, whether the imposed mediating responses used in many protocols effectively prevents the use of nonmnemonic strategies remains insufficiently explored. In this study, we recorded and analyzed the movements of mice during these tasks using automated video tracking. Our findings suggest that mice may rely on nonmnemonic strategies, such as maintaining body position relative to the sample keys, rather than solely on working memory. These results challenge the assumption that DMTP and DNMTP tasks exclusively measure working memory, highlighting the need for cautious interpretation of task performance.