The object delayed non-matching to sample task in rats does not depend on working memory

Four rats were trained on an object delayed non-matching to sample task to measure working memory (WM). On separate trials a sample object was presented to the rat, which was presented again along with another object after a 3 s delay rats were rewarded with food if they chose the novel object. In this procedure the choice objects are positioned in a fixed order (the novel object was positioned at first, followed by the sample object). When the order of placement of the choice objects was reversed performance dropped from 87% to 28% correct, showing that the rats always chose the object that was first positioned into the choice area. When the order of positioning of the choice objects was randomized during training, rats were not able to learn the task. It is suggested that performance is based on a discrimination between choice objects instead of WM.