The Effects of Reward Omission and Change on Proactive Interference in Spatial Alternation. William Scott Terry

Whether reward omission on the sample runs of a delayed alternation task
would reduce proactive interference was tested in two experiments. On the first
run of a trial, the animal entered one arm of a maze to obtain a food reward;
on the second run the reward was available in the opposite arm. Alternation
declined across a block of massed trials, but there was less of a decrease when
the reward was omitted in the final trial or when another maze was substituted.
In a task that required retention of three mazes at a time, alternation again declined
across massed trials but was reinstated either by omitting the reward on
the final trial or by substituting a different reinforcer. The results suggest that
altering the reward outcome can improve memory discriminability of recently
visited maze locations and thereby overcome trace confusion in memory.