This chapter discusses information coding in higher sensory and memory areas. Neurons are vastly simpler than human beings are, but the metaphor is not completely silly because it illustrates the volatility of the notion of neural codes. Information theory has been developed precisely to quantify communication and is quintessential to an appraisal of neural codes. Applying information theory to neural activity (rather than to the synthetic communication systems for which it was developed) is however riddled with practical problems and subtleties, which must be clarified before reporting experimental results. The chapter considers other means of neuronal communication than the emission of action potentials or spikes and regards them as self-similar all-or-none events whose only distinctive features are the time of emission and the identity of the emitting neuron. The extent to which the firing rates of a population of neurons may or may not carry most of the information represented in the complete list of spike emission times is a question to be addressed experimentally in any given situation.
