Coding and Representation

Neural activity does not in itself contain much information about the stimulus. Spikes elicited by different stimuli look quite the same, but they do occur in different cells. The information is thus contained in the specificity of the cell generating the spike, together with the spiking activity itself. This may be considered a modern account of the notion of the “specific sense energies” formulated by Müller (1837): Stimulation of the eye is perceived as light, color, or motion because it activates visual pathways; similarly, stimulation of the ear is perceived as noise, a pitched tone, or a human utterance, because it activates auditory parts of the brain. The detailed information encoded by each neuron is described by its tuning curve which summarizes the neuron’s response to possible stimuli. Tuning curves of different neurons overlap, leading to population coding where each stimulus is represented by the activities of a group, or population, of neurons. This chapter explores the consequences of population coding for neural information processing. It also discusses localization: that is, the fact that neighboring neurons in the cortex tend to have overlapping receptive fields and tuning curves. Cortical areas therefore often form “maps” of represented parameters such as visual field position, acoustic pitch, or the surface of the body.