Abstract
The delayed matching-to-sample (DMS) task is widely employed to assess memory in a range of non-human animals. On the standard “common outcomes” (CO) DMS task, correct performance following either sample stimulus results in reinforcement. In contrast, on a “differential outcomes” (DO) DMS task, the outcome following either sample stimulus is different. One of the most consistent findings in the comparative literature is that performance under a DO condition is superior to that under a CO condition. The superior performance is attributed to the fact the DO condition enhances memory for the sample stimulus by tagging each sample with a discrete reward. Here, we investigate an alternative possibility, that pigeons use positional mediation during the delay under DO, but not CO, conditions. To test this, we tracked the head position of pigeons performing a DO (n = 4) or CO (n = 4) task. Consistent with the positional mediation account, all subjects in the DO condition displayed evidence of positional mediation. Surprisingly, positional mediation was not unique to subjects in the DO condition, with subjects in the CO condition also displaying evidence of mediation.