RESUMO
In order to have a tool to empirically test the ideas derived from a theoretical model, we extended a protocol for evaluation of episodic-like memory in rats, based on the triad "what, where, context" for definition of memories. As with the computational model, our intention was for the animal being tested to store a specific number of object-place-context configurations as different memories, which would then be retrievable from cues. The aim of this work was to evaluate the influence of the number of configurations to be memorized on the performance of the task. Sixty-five Wistar male rats were evaluated. In accordance with previous work, for two configurations, the recognition index was indicative of recognition of the element mismatching the original memory (mean = 0.28; SEM = 0.12). The recognition index for three configurations was lower (mean = 0.15; SEM = 0.10), evidencing less recall with increasing requirements. The results also showed a trend toward recognition of novelty for the first and the last memory when evaluating three configurations (a "U" shape in the exploratory preference's curve), showing the primacy and recency effects typical of memory both in humans and animals. Nonetheless, the data presented a high inter-subject variability which makes the test non-robust for small groups. However, if used before and after a treatment for a same subject, we suggest that the protocol presented in this work can be a useful behavioral test for the evaluation of episodic-like memory in rats in terms of a variable task demand.