Publication
Wolfram Schultz. Dopamine reward prediction error coding. Dialogues in Clinical Neuroscience 18:23–32 (2016). [19]
Site summary — not the original abstract
Schultz’s review describes reward-prediction-error signaling: many dopamine neurons respond more to better-than-expected rewards, remain near baseline for fully predicted rewards, and reduce activity when rewards are worse than expected, drawing on human, monkey, and rodent research. [19]
In this account, a fully predicted reward need not produce a new prediction-error signal, while an already learned behavior can continue; this is not the claim that expected rewards cannot support behavior. [19]
The article explains a learning mechanism rather than testing a dog-training protocol, and it does not establish the site’s former claims about obedience, training efficiency, or a stronger human–dog bond. [19]
Responses