¹⁰There is a subtle point about perception of the reward loss, which is that the system may first compute percepts of the two quantities (expected and actually obtained rewards) and then compute the difference, or it can directly attempt to compute the percept of the difference. In other words, we can have perception of the difference, or the difference of the perceptions. (In the case of the expectation, its “perception” might rather be called the “estimation” of the expectation.) Intelligent systems may use either of these two approaches. I shall not venture into speculating which might be the case in the human brain.