Figure 2.
Set-shift group performance across trial blocks during set 2. Overall performance scores (average of both PA and RA start arm classes) were close to chance level (50%) at the outset of set 2 (trial block 1) and rose significantly across trial blocks. Analysis of performance from the two start arm classes shows that, at the outset of testing, rats performed significantly below chance (50% correct) from PA starts, whereas from RA starts, their performance was significantly above chance. Thus, at the beginning of set 2, rats maintained the cognitive set associated with the set 1 discrimination rule. As set 2 training progressed, rats shifted their behavior to the new discrimination rule; by trial block 10, performance was nearly perfect for both start arm classes.