Time for thinking modulates learning

Here is a follow-up to Study 1. We use the same standard bandit learning task, and manipulate the ITI to be either short (1s) or long (4s) as used in Study 1, except here, we mean to ablate our main effect of ITI. Study 1's showed that that, after a loss, there is greater lose-shift behavior after a long ITI compared to after a short ITI. We believe this is due to replay of the negative event during the ITI. Per past studies (Gershman et al., J Exp Psych Gen), having a working-memory intensive task can disrupt replay, which should remove the effect we found in Study 1. Thus here in Study 2, we have people carry out a demanding perceptual decision making task during the ITI. In pilot evidence, we show this does not disrupt learning in general, but appears to remove the ITI effect. We thus show that we can remove the ITI effect when disrupting replay, which serves to bolster our main hypothesis, and show evidence against the confounding effect that the ITI effect is due to forgetting (which is passive and should not be interrupted by the new design in Study 2). Initial analysis of pilot data and analyses can be found here: https://github.com/sharplabBIU/iti_learning