Original language | English (US) |
---|---|
Title of host publication | Encyclopedia of Neuroscience |
Publisher | Elsevier Ltd |
Pages | 749-755 |
Number of pages | 7 |
ISBN (Print) | 9780080450469 |
DOIs | |
State | Published - 2009 |
Abstract
Population coding is a way of representing stimuli using large sets of neurons with similar response properties, such as hypercolumns of orientation-selective neurons in primary visual cortex. Population codes are widespread in cortex and their study is a cornerstone of systems neuroscience. Their power lies in the large amount of information that can be encoded even if the neurons are broadly tuned and subject to high variability, as is typically the case. This article reviews the main properties of population codes from two complementary perspectives: encoding and decoding. Encoding is the probabilistic mapping from stimulus to population response. Decoding is the process of extracting a stimulus estimate from a given population response.
Keywords
- Bayesian inference
- Computation
- Decoding
- Encoding
- Fisher information
- Noise
- Optimality
- Perception
- Population codes
- Representation
- Variability
ASJC Scopus subject areas
- General Neuroscience