Fig. 2.

Design of Experiment 1. An initial Familiarisation phase informed participants about which scene category was associated with positive or negative valence through passive ‘Watch’ and active ‘Predict’ blocks. In the subsequent Training phase, new scenes from the same categories were paired with specific words (adjectives) that possessed the same valence as that given in the Familiarisation phase. In the critical Study phase, new scenes were again paired with new words of the same valence as Training, but on the third time a scene category was presented, the valence of the paired word was either switched from that experienced during Training/Familiarisation (inconsistent: High PE condition) on one half of trials, or remained (consistent: Low PE condition) on the other half. Because these “critical” Study trials were intermixed with “filler” trials of first and second presentations of a scene category in the Study phase (for which the valence of the paired word was always consistent with Training/Familiarisation), the overall proportion of inconsistent trials was only 1/6th of trials in the Study phase. Memory for the pairings in the critical trials of the Study phase was assessed in the final Test Phase, which used 3AFC in which the target word was offered together with two foil words that were of the same valence and encountered in other trials of the same condition during the critical Study Phase.