The case for economics when dealing with biological reinforcement

The function of rewards is to ensure the survival of biological organisms and their gene propagation. The value of a reward depends on the momentary needs of the organism; it is subjective and extends beyond physical and chemical properties. Hence, meaningful neuronal reward signals should code subjective reward value.

      While reinforcement models can be implemented on computers, silicon machines do not distinguish physical, objective reward value from subjective reward value. But the distinction matters for biological organisms: an efficient reinforcement signal should update subjective rather than objective reward value to avoid more demanding subsequent translation into subjective value. Thus, to efficiently participate in neuronal reinforcement mechanisms, the dopamine signal should code subjective rather than objective reward value. 

      Subjective reward value cannot be measured directly but can be inferred from observable choice: the more frequently an organism chooses a reward, the higher is its subjective value for that organism. Organisms failing to do so have been outcompeted in evolution. Thus, to understand how dopamine neurons code subjective reward value, we need to study choice. We benefit from economic choice theories that provide formal concepts and practical assessment methods for subjective reward value.