Beyond Rescorla–Wagner: the Ups and Downs of Learning

We check the robustness of a recently proposed dynamical model associative Pavlovian learning that extends Rescorla-Wagner (RW) in natural way and predicts progressively damped oscillations response subjects. Using data two experiments, we compare oscillatory (DOM) with an made superposition RW curve oscillations. Not only do clearly show pattern, but they also favor DOM over added oscillation model, thus pointing out these are manifestation process. The latter is interpreted as fact subjects make predictions on trial outcomes more extended time than uncertainty.