Experiencing associative interference facilitates excitatory predictive learning about a conditioned inhibitor in humans

Résumé

The purpose of this doctoral thesis is to explore the impact of increasing prediction error on subsequent new learning. To achieve this goal, we have carried out two human predictive learning studies to evaluate the impact of a sudden increase in prediction error generated by different association interference experience on subsequent learning. Assuming that the increase in prediction error leads to a general increase in attention, this will facilitate subsequent learning. Both, theoretical review and empirical results, show that the manipulation of prediction error can increased attention, which in turn translates to improvements in subsequent learning, regardless of whether the increases in the prediction error involves extinction of a previously trained cue or pairing a cue with the outcome when it was previously followed by the absence of it.