Much of computational neuroscience begins and ends with the responses of individual neurons. The field itself sprang from work in the 1950s aimed at uncovering the biophysical mechanisms underlying spike generation (Hodgkin and Huxley 1952), and other classic studies focus on the computational capabilities of dendritic trees and on how neural activity encodes sensory stimuli or statistical information about the world, to mention a couple of well-known examples (Dayan and Abbott 2001). A different point of view, however, is one in which neurons are the intermediaries between a subject and its environment. As the engines of behavior, neurons need to be computationally powerful for the express purpose of giving the subject an advantage, and hence their efficiency or performance should be measured with respect to the subject’s success. So the computational neuroscience of decision making is computational neuroscience in this context; it is the quest to understand how...
- Dayan P, Abbott LF (2001) Theoretical neuroscience. MIT Press, Cambridge, MAGoogle Scholar