Neural correlations, population coding and computation

@article{Averbeck2006NeuralCP,
  title={Neural correlations, population coding and computation},
  author={Bruno B. Averbeck and Peter E. Latham and Alexandre Pouget},
  journal={Nature Reviews Neuroscience},
  year={2006},
  volume={7},
  pages={358-366}
}
How the brain encodes information in population activity, and how it combines and manipulates that activity as it carries out computations, are questions that lie at the heart of systems neuroscience. During the past decade, with the advent of multi-electrode recording and improved theoretical models, these questions have begun to yield answers. However, a complete understanding of neuronal variability, and, in particular, how it affects population codes, is missing. This is because variability… 
The Geometry of Information Coding in Correlated Neural Populations.
TLDR
This review provides an overview of the theoretical developments on the influence of noise correlations on the properties of neural coding and emphasizes the relevance of the fine structure of noise correlation.
Correlation susceptibility and single neuron computation
TLDR
This work has theoretically derived that a ratio of an output and input correlation, termed correlation susceptibility, can be written in terms of the firing rate and its gain under some assumptions, and revisited the question of how rate and correlation based coding would work in a neural population.
Neural Population Codes
TLDR
It is argued that the combinatorial number of potential activity patterns and correlations implies that understanding the code of large populations must rely on identifying simplifying principles of the design of neural codes, and an information theory–based framework is presented for assessing the role and impact of correlations and noise between cells and how it shapes the nature of the code.
When does recurrent connectivity improve neural population coding?
TLDR
Preliminary results indicate that, in some cases – similar to those discussed in [9] – recurrent coupling can combat the cell-intrinsic variabililty enough that the population’s coding performance is improved even though the noise correlations the are generated by that coupling do themselves hinder that coding performance.
Correlations and Neuronal Population Information.
TLDR
It is argued that this is a critical lesson for those interested in neuronal population responses more generally: Descriptions of population responses should be motivated by and linked to well-specified function.
Measuring and interpreting neuronal correlations
TLDR
These guidelines for interpreting correlation data are offered and a discussion of how best to evaluate the effect of correlations on cortical processing is discussed.
Do high order cross-correlations among neurons convey useful information?
TLDR
The application of the method proposed here will unravel the role of high order correlations among neurons in sensory coding, thus giving an insight into the complexity of the coding.
Effects of Neuronal Correlations on Population Decoding and Encoding Models
TLDR
This thesis analyzes the effect of the correlations in neural activity on the information that is encoded in and can be decoded from a population of neurons using models that take into account the pairwise correlations.
...
...

References

SHOWING 1-10 OF 92 REFERENCES
Coding and transmission of information by neural ensembles
Nonlinear Population Codes
TLDR
This letter studies correlated neuronal populations whose higher-order statistics, and in particular response variances, are also modulated by the stimulus, and proposes a bilinear readout scheme that involves spatial decorrelation, quadratic nonlinearity, and population vector summation.
Synergy, Redundancy, and Independence in Population Codes, Revisited
TLDR
It is shown that synergy and ΔIshuffled are confounded measures: they can be zero when correlations are clearly important for decoding and positive when they are not; in contrast, ΔI is not confounded, and has an information theoretic interpretation.
Synergy, Redundancy, and Independence in Population Codes
TLDR
This work distinguishes questions about how information is encoded by a population of neurons from how that information can be decoded, and shows that these measures form an interrelated framework for evaluating contributions of signal and noise correlations to the joint information conveyed about the stimulus.
Effects of noise correlations on information encoding and decoding.
TLDR
d prime and related information measures were used to examine how studies of noise correlations from these two perspectives are related and showed that d prime provides an accurate measure of information encoding and decoding in the population of neurons.
Population coding in neuronal systems with correlated noise.
TLDR
The analysis provides an estimate of the effective number of statistically independent degrees of freedom, denoted N(eff), that a large correlated system can have, and shows that positive correlations decrease the estimation capability of the network, relative to the uncorrelated population.
Correlations, feature‐binding and population coding in primary visual cortex
TLDR
It is suggested that firing rates, rather than correlations, are the main element of the population code for feature binding in primary visual cortex.
Correlations and the encoding of information in the nervous system
TLDR
This analysis provides fundamental constraints upon the nature of information coding, showing that over short time–scales, correlations cannot dominate information representation, and that stimulus–independent correlations may lead to synergy, but that only certain combinations of the different sources of correlation result in significant synergy.
Reading population codes: a neural implementation of ideal observers
TLDR
Using simulations and analysis, it is shown that a close approximation to maximum likelihood can be implemented in a biologically plausible model of cortical circuitry, suggesting that cortical areas may, in general, function as ideal observers of activity in preceding areas.
Coding of visual information by precisely correlated spikes in the lateral geniculate nucleus
TLDR
An information-theoretic analysis is used to examine the role of correlations among neurons in the lateral geniculate nucleus with overlapping receptive fields in visual coding and finds that precise temporal correlation could be used as an additional information channel from thalamus to visual cortex.
...
...