The mesolimbic pathways appear to be involved in signaling both factual and counterfactual reward prediction errors (regret and relief). In which ways are these signals believed to affect neurons in the cortex? It seems to me there are several possibilities, for example:
- The mesolimbic pathways "know" which cortical areas are responsible for recent actions and directly send signals to them to change,
- or the cortical areas themselves "know" when they are responsible for a prediction error, and they know which neurons to blame and how to change them. All the mesolimbic pathways do is just encouraging the cortex to repeat recent activations in a counterfactual manner a couple of times so that the error becomes smaller each time. (This is how relief and regret feel to me, if I'm not fooled by some preconceptions.)