Home | About | Journals | Submit | Contact Us | Français |

**|**Front Comput Neurosci**|**v.4; 2010**|**PMC2870944

Formats

Article sections

Authors

Related links

Front Comput Neurosci. 2010; 4: 9.

PMCID: PMC2870944

Edited by: Philipp Berens, Baylor College of Medicine, USA Max Planck Institute for Biological Cybernetics, Germany

Reviewed by: Nestor Parga, Columbia University, USA; John A. Hertz, Niels Bohr Institute, Denmark; Arvind Kumar, University of Freiburg, Germany

*Correspondence: Robert J. Rosenbaum, University of Houston, Department of Mathematics, Houston, TX 77204-3008, USA. e-mail: ude.hu.htam@rtrebor

Received 2009 November 26; Accepted 2010 March 24.

Copyright © 2010 Rosenbaum, Trousdale and Josić.

This is an open-access article subject to an exclusive license agreement between the authors and the Frontiers Research Foundation, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are credited.

This article has been cited by other articles in PMC.

Correlations between spike trains can strongly modulate neuronal activity and affect the ability of neurons to encode information. Neurons integrate inputs from thousands of afferents. Similarly, a number of experimental techniques are designed to record pooled cell activity. We review and generalize a number of previous results that show how correlations between cells in a population can be amplified and distorted in signals that reflect their collective activity. The structure of the underlying neuronal response can significantly impact correlations between such pooled signals. Therefore care needs to be taken when interpreting pooled recordings, or modeling networks of cells that receive inputs from large presynaptic populations. We also show that the frequently observed runaway synchrony in feedforward chains is primarily due to the pooling of correlated inputs.

Cortical neurons integrate inputs from thousands of afferents. Similarly, a variety of experimental techniques record the pooled activity of large populations of cells. It is therefore important to understand how the structured response of a neuronal network is reflected in the pooled activity of cell groups.

It is known that weak dependencies between the response of cell pairs in a population can have a significant impact on the variability and signal-to-noise ratio of the pooled signal (Shadlen and Newsome, 1998; Salinas and Sejnowski, 2000; Moreno-Bote et al., 2008). It has also been observed that weak correlations between cells in two populations can cause much stronger correlations between the pooled activity of the populations (Bedenbaugh and Gerstein, 1997; Chen et al., 2006; Gutnisky and Josić, 2010; Renart et al., 2010). We give a simple example of this effect in Figure Figure1C:1C: Weak correlations were introduced between the spiking activity of cells in two non-overlapping presynaptic pools each providing input to a postsynaptic cell (see diagram in Figure Figure1B).1B). The activity between pairs of excitatory, and pairs of inhibitory cells was correlated, but excitatory–inhibitory pairs were uncorrelated. Even without shared inputs and with background noise, pooling resulted in strong correlations in postsynaptic membrane voltages. The connectivity in the presynaptic network was irrelevant – it only mattered that the inputs to the downstream neurons reflected the pooled activity of the afferent populations. A similar effect can cause large correlations between recordings of multiunit activity (MUA) or recordings of voltage sensitive dyes (VSD), even when correlations between cells in the recorded populations are small (Bedenbaugh and Gerstein, 1997; Chen et al., 2006; Stark et al., 2008). The effect is the same, but in this case pooling occurs at the level of a recording device rather than a downstream neuron (compare Figures Figures11A,B).

We present a systematic overview, as well as extensions and applications of a number of previous observations related to this phenomenon. Using a linear model, we start by examining the potential effects of pooling on recordings from large populations obtained using VSD or MUA recording techniques. These techniques are believed to reflect the pooled postsynpatic activity of groups of cells. We extend earlier models introduced to examine the impact of pooling on correlations (Bedenbaugh and Gerstein, 1997; Chen et al., 2006; Nunez and Srinivasan, 2006), and show that heterogeneities in the presynaptic pools can have subtle effects on correlations between pooled signals.

Since neurons respond to input from large presynaptic populations, pooling also impacts the activity of single cells and cell pairs. As observed in Figure Figure1C,1C, pooling can inflate weak correlations between afferents. However, excitatory–inhibitory correlations (Okun and Lampl, 2008) can counteract this amplification, as shown in Figure Figure1D1D (Hertz, 2010; Renart et al., 2010). We examine these effects analytically by modeling the subthreshold activity of postsynaptic cells as a filtered version of the inputs received (Tetzlaff et al., 2008). The impact of correlated subthreshold activity on the output spiking statistics is a nontrivial question which we address only briefly (Moreno-Bote and Parga, 2006; de la Rocha et al., 2007; Ostojić et al., 2009).

The effects of pooling provide a simple explanation for certain aspects of the dynamics of feedforward chains. Simulations and *in vitro* experiments show that layered feedforward architectures give rise to a robust increase in synchronous spiking from layer to layer (Diesmann et al., 1999; Litvak et al., 2003; Reyes, 2003; Doiron et al., 2006; Kumar et al., 2008). We describe how output correlations in one layer impact correlations between the pooled inputs to the next layer. This approach is used to derive a mapping that describes how correlations develop across layers (Tetzlaff et al., 2003; Renart et al., 2010), and to illustrate that the pooling of correlated inputs is the primary mechanism responsible for the development of synchrony in feedforward chains. Examining how correlations are mapped between layers also helps explain why asynchronous states are rarely observed in feedforward networks in the absence of strong background noise (van Rossum et al., 2002; Vogels and Abbott, 2005). This is in contrast to recurrent networks which can display stable asynchronous states (Hertz, 2010; Renart et al., 2010) similar to those observed *in vivo* (Ecker et al., 2010).

The cross-covariance of a pair of stationary stochastic processes, *x*(*t*) and *y*(*t*), is *C _{xy}*(

$${\gamma}_{xy}={\displaystyle \underset{-\infty}{\overset{\infty}{\int}}{C}_{xy}(t)\text{\hspace{0.17em}}dt},\text{\hspace{1em}}{\sigma}_{x}^{2}={\gamma}_{xx},\text{\hspace{1em}}{\rho}_{xy}=\frac{{\gamma}_{xy}}{{\sigma}_{x}{\sigma}_{y}}.$$

(1)

While the asymptotic correlation, ρ* _{xy,}*, measures correlations between

Given two collections of correlated random variables ${\left\{{x}_{i}\right\}}_{i=1}^{{n}_{x}}$ and ${\left\{{y}_{j}\right\}}_{j=1}^{{n}_{y}},$ define the pooled variables, $X={\Sigma}_{i}{x}_{i}$ and $Y={\Sigma}_{i}{y}_{i}.$ Since covariance is bilinear $(cov({\Sigma}_{i}{x}_{i},{\Sigma}_{j}{y}_{j})={\Sigma}_{ij}cov({x}_{i},{y}_{j}))$ the variance and covariance of the pooled variables are

$${\sigma}_{X}^{2}={\displaystyle \sum _{i=1}^{{n}_{x}}{\displaystyle \sum _{\begin{array}{l}j=1\\ j\ne i\end{array}}^{{n}_{x}}{\sigma}_{{x}_{i}}{\sigma}_{{x}_{j}}{\rho}_{{x}_{i}{x}_{j}}}+{\displaystyle \sum _{i=1}^{{n}_{x}}{\sigma}_{{x}_{i}}^{2}},\text{\hspace{1em}and\hspace{1em}}}{\gamma}_{XY}={\displaystyle \sum _{i=1}^{nx}{\displaystyle \sum _{j=1}^{ny}{\sigma}_{{x}_{i}}{\sigma}_{{y}_{j}}{\rho}_{{x}_{i}{y}_{j}}}},$$

and similarly for ${\sigma}_{Y}^{2}.$

Using these expressions along with some algebraic manipulation, the correlation coefficient, ${\rho}_{XY}={\gamma}_{XY}/\sqrt{{\sigma}_{X}{\sigma}_{Y}},$ between the pooled variables can be written as

$${\rho}_{XY}=\frac{\overline{\rho}xy}{\sqrt{\left[{w}_{x}{\overline{\rho}}_{xx}+\frac{1}{{n}_{x}}\left(\frac{{\overline{v}}_{x}}{\overline{{\sigma}_{x}{\sigma}_{y}}}-{w}_{x}{\overline{\rho}}_{xx}\right)\right]\left[{w}_{y}{\overline{\rho}}_{yy}+\frac{1}{{n}_{y}}\left(\frac{{\overline{v}}_{y}}{\overline{{\sigma}_{x}{\sigma}_{y}}}-{w}_{y}{\overline{\rho}}_{yy}\right)\right]}}$$

(2)

$$=\frac{{\overline{\rho}}_{xy}}{\sqrt{{\overline{\rho}}_{xx}{\overline{\rho}}_{yy}}}+\mathcal{O}\left(\frac{1}{\sqrt{{n}_{x}{n}_{y}}}\right),$$

(3)

where

$$\begin{array}{l}{w}_{x}=\frac{\overline{{\sigma}_{x}{\sigma}_{x}}}{\overline{{\sigma}_{x}{\sigma}_{y}}},\text{\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}}\overline{{v}_{x}}=\frac{1}{{n}_{x}}{\displaystyle \sum _{i=1}^{{n}_{x}}{\sigma}_{{x}_{i}}^{2}},\\ \overline{{\sigma}_{x}{\sigma}_{y}}=\frac{1}{{n}_{x}{n}_{y}}{\displaystyle \sum _{i=1}^{{n}_{x}}{\displaystyle \sum _{j=1}^{{n}_{y}}{\sigma}_{{x}_{i}}{\sigma}_{{y}_{i}}}},\text{\hspace{1em}}\overline{{\sigma}_{x}{\sigma}_{x}}=\frac{1}{{n}_{x}\left({n}_{x}-1\right)}{\displaystyle \sum _{i=1}^{{n}_{x}}{\displaystyle \sum _{\begin{array}{l}j=1\\ j\ne 1\end{array}}^{{n}_{x}}{\sigma}_{{x}_{i}}{\sigma}_{{x}_{j}}}},\\ {\overline{\rho}}_{xy}=\frac{1}{{n}_{x}{n}_{y}\overline{{\sigma}_{x}{\sigma}_{y}}}{\displaystyle \sum _{i=1}^{{n}_{x}}{\displaystyle \sum _{j=1}^{{n}_{y}}{\sigma}_{{x}_{i}}{\sigma}_{{y}_{j}}{\rho}_{{x}_{i}{y}_{j}}}},\\ {\overline{\rho}}_{xx}=\frac{1}{{n}_{x}\left({n}_{x}-1\right)\overline{{\sigma}_{x}{\sigma}_{x}}}{\displaystyle \sum _{i=1}^{{n}_{x}}{\displaystyle \sum _{\begin{array}{l}j=1\\ j\ne 1\end{array}}^{{n}_{x}}{\sigma}_{{x}_{i}}{\sigma}_{{y}_{j}}{\rho}_{{x}_{i}{x}_{j}}}}\end{array}$$

and similarly for *w _{y}*,

Each overlined term above is a population average. Notably, ${\overline{\rho}}_{xy}$ represents the average correlation between *x _{i}* and

Assuming that variances are homogeneous within each population, that is ${\sigma}_{{x}_{i}}={\sigma}_{x}$ and ${\sigma}_{{y}_{j}}={\sigma}_{y}$ for *i*=1,…,*n _{x}* and

$${\rho}_{XY}=\frac{{\overline{\rho}}_{xy}}{\sqrt{\left[{\overline{\rho}}_{xx}+\frac{1}{{n}_{x}}\left(1-{\overline{\rho}}_{xx}\right)\right]\hspace{0.17em}\left[{\overline{\rho}}_{yy}+\frac{1}{{n}_{y}}\left(1-{\overline{\rho}}_{yy}\right)\right]}}.$$

(4)

Assuming further that the populations are symmetric, σ* _{x}*=σ

$${\rho}_{XY}=\frac{{\rho}^{b}}{{\rho}^{w}+\frac{1}{n}\left(1-{\rho}^{w}\right)}=\frac{{\rho}^{b}}{{\rho}^{w}}+\mathcal{O}\left(\frac{1}{n}\right),$$

(5)

where ${\rho}^{b}={\overline{\rho}}_{xy}$ is the average pairwise correlation between the two populations and ${\rho}^{w}={\overline{\rho}}_{xx}={\overline{\rho}}_{yy}$ is the average pairwise correlation within each population. Eq. (5) was derived in Bedenbaugh and Gerstein (1997) in an examination of correlations between multiunit recordings. In Chen et al. (2006), a version of Eq. (5) with ρ* ^{w}*=ρ

Note that the results above hold for correlations computed over arbitrary time windows. We concentrate on infinite windows, and discuss extensions in the Appendix.

In the second part of the presentation we consider two excitatory and two inhibitory input populations projecting to two postsynaptic cells. The *j*^{th} excitatory input to cell *k* is labeled e* _{j,k}*(

Each of the excitatory and inhibitory inputs to cell *k*, are stationary spike trains modeled by point processes, ${\text{e}}_{j,k}(t)={\Sigma}_{i}\text{\hspace{0.17em}}\delta \text{\hspace{0.17em}}(t-{t}_{j,k}^{i})$ and ${\text{i}}_{j,k}(t)={\Sigma}_{i}\text{\hspace{0.17em}}\delta \text{\hspace{0.17em}}(t-{s}_{j,k}^{i})$ where $\left\{{t}_{j,k}^{i}\right\}$ and $\left\{{s}_{j,k}^{i}\right\}$ are input spike times. We assume that the spike trains are stationary in a multivariate sense (Stratonovich, 1963). The pooled excitatory and inhibitory inputs to neuron *k* are ${E}_{k}(t)={\Sigma}_{j=1}^{{n}_{{e}_{k}}}{\text{e}}_{j,k}(t),$ and ${I}_{k}(t)={\Sigma}_{j=1}^{{n}_{{\text{i}}_{k}}}{\text{i}}_{j,k}(t).$

To generate correlated inputs to cells, we used the multiple interaction process (MIP) method (Kuhn et al., 2003), then jittered each spike time independently by a random value drawn from an exponential distribution with mean 5ms. The resulting processes are Poisson with cross-covariance functions proportional to a double exponential, *C _{xy}*(

While the dynamics of the afferent population were not modeled explicitly, the response of the two downstream neurons was obtained using a conductance-based IF model. The membrane potentials of the neurons were described by

$${C}_{m}\frac{d{V}_{k}}{dt}=-{g}_{L}({V}_{k}-{V}_{L})-{g}_{{E}_{k}}(t)({V}_{k}-{V}_{E})-{g}_{{I}_{k}}(t)({V}_{k}-{V}_{I}),$$

(6)

with excitatory and inhibitory conductances determined by ${g}_{{E}_{k}}(t)=({\text{E}}_{k}\ast {\alpha}_{\text{e}})(t)$ and ${g}_{{\text{I}}_{k}}(t)=({\text{I}}_{k}\ast {\alpha}_{\text{i}})(t)$ where * denotes convolution. We used synaptic responses of the form ${\alpha}_{\text{e}}(t)=\mathcal{E}{\tau}_{\text{e}}^{-2}t{\text{e}}^{-t/{\tau}_{\text{e}}}\Theta (t)$ and ${\alpha}_{\text{i}}(t)=\mathcal{I}{\tau}_{\text{i}}^{-2}t{e}^{-t/{\tau}_{\text{i}}}\Theta (t),$ where Θ(*t*) is the Heaviside function. The *area* of a single excitatory or inhibitory postsynaptic conductance (EPSC or IPSC) is therefore equal to the synaptic weight, or , with units nS·ms. This analysis can easily be extended to situations where each input, e* _{j,k}* or i

When examining spiking activity, we assume that when *V _{k}* crosses a threshold voltage,

As a measure of balance between excitation and inhibition we used (Troyer and Miller, 1997; Salinas and Sejnowski, 2000)

$$\beta =\frac{\left|{V}_{L}-{V}_{\text{E}}\right|\mathcal{E}{\nu}_{\text{e}}{n}_{\text{e}}}{\left|{V}_{L}-{V}_{\text{I}}\right|\mathcal{I}{\nu}_{i}{n}_{i}}.$$

When β=1, the net excitation and inhibition are balanced and the mean free membrane potential equals *V _{L}*. In simulations, we set

The conductance-based IF neuron behaves as a nonlinear filter in the sense that membrane potentials cannot be written as a linear transformation of the inputs. However, following Kuhn et al. (2004) and Coombes et al. (2007), we derive a linear approximation to the conductance based model. Let *U*=*V _{k}*−

$${C}_{m}\frac{dU}{dt}=\left(-{g}_{L}-{g}_{\text{E}}(t)-{g}_{\text{I}}(t)\right)U-{g}_{\text{E}}(t)\left({V}_{L}-{V}_{\text{E}}\right)-{g}_{\text{I}}(t)\left({V}_{L}-{V}_{\text{I}}\right).$$

Define the effective membrane time constant,τ_{eff}C_{m}/(*E*[*g _{L}*+g

$$\frac{dU}{dt}=-\frac{1}{{\tau}_{\text{eff}}}U+{J}_{k}(t),$$

(7)

where ${J}_{k}(t)=(-{g}_{{\text{E}}_{k}}(t)({V}_{L}-{V}_{\text{E}})-{g}_{{\text{I}}_{k}}(t)({V}_{L}-{V}_{\text{I}}))/{C}_{m}$ is the total input current to cell *k*. Solving and reverting to the original variables gives the linear approximation *V _{k}*(

The pooling of signals from groups of neurons can impact both recordings of population activity and the structure of inputs to postsynaptic cells. We start by discussing correlations in pooled recordings using a simple linear model. A similar model is then used to examine the impact of pooling on the statistics of inputs to cells. For simplicity we assume that all spike trains are stationary. However, non-stationary results can be obtained using similar methods as outlined in the Section “Discussion.” Though all parameters are defined in the Meterials and Methods, Tables Tables11 and and22 in the Appendix contain brief descriptions of parameters for quick reference. Also, Tables Tables33 and and44 summarize the values of parameters used for simulations throughout the article.

Pooling can impact correlations between recordings of population activity obtained from voltage sensitive dyes (VSDs), multi-unit recordings and other techniques. Such signals might each represent the summed activity of hundreds or thousands of neurons. Let two recorded signals, *X*_{1}(*t*) and *X*_{2}(*t*), represent the weighted activity of cells in two populations (see diagram in Figure Figure1A).1A). If we assume homogeneity in the input variances and equal size of the recorded populations, Eq. (4) gives the correlation between the recorded signals

$${\rho}_{{X}_{1}{X}_{2}}=\frac{{\overline{\rho}}_{12}}{\sqrt{\left[{\overline{\rho}}_{11}+\frac{1}{n}\left(1-{\overline{\rho}}_{11}\right)\right]\hspace{0.17em}\left[{\overline{\rho}}_{22}+\frac{1}{n}\left(1-{\overline{\rho}}_{22}\right)\right]}}=\frac{{\overline{\rho}}_{12}}{\sqrt{{\overline{\rho}}_{11}{\overline{\rho}}_{22}}}+\mathcal{O}\left(\frac{1}{n}\right).$$

(8)

Here *n* represents the number of neurons recorded, ${\overline{\rho}}_{kk},$ *k*=1,2 represents the average correlation between cells contributing to signal *X _{k}*(

When the two recording sites are nearby, so that ${\overline{\rho}}_{12}\approx {\overline{\rho}}_{11}\approx {\overline{\rho}}_{22},$ even small correlations between individual cells are amplified by pooling so that the correlations between the recorded signals can be close to 1. This effect was observed in experiments and explained in similar terms in Stark et al. (2008).

A significant stimulus-dependent change in correlations between individual cells might be reflected only weakly in the correlation between the pooled signals. This can occur, for instance, in recordings of large populations when ${\overline{\text{\rho}}}_{12},$${\overline{\text{\rho}}}_{11},$ and ${\overline{\text{\rho}}}_{22}$ are increased by the same factor when a stimulus is presented. Similarly, an increase in correlations between cells can actually lead to a *decrease* in correlations between recorded signals when ${\overline{\text{\rho}}}_{11}$ and ${\overline{\text{\rho}}}_{22}$ increase by a larger factor than${\overline{\text{\rho}}}_{12}$.

To illustrate these effects, we construct a simple model of stimulus dependent correlations motivated by the experiments in Chen et al. (2006), in which VSDs were used to record the population response in visual area V1 during an attention task. In their experiments, the imaged area is divided into 64 pixels, each 0.25mm× .25mm in size. The signal recorded from each pixel represents the pooled activity of *n*≈1.25×10^{4} neurons.

We model correlations between the signals, *X*_{1}(*t*) and *X*_{2}(*t*), recorded from two pixels in the presence or absence of a stimulus (see Figure Figure2B),2B), using a simplified model of stimulus dependent rates and correlations. The firing rate of a cell located at distance *d* from the center of the retinotopic image of a stimulus is

$$r(d)=\{\begin{array}{cc}B+\frac{(1-B){(1+\mathrm{cos}(d\pi ))}^{\lambda}}{2}& \text{stimulus\hspace{0.17em}present}\\ B\text{\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{1em}\hspace{0.17em}\hspace{0.05em}}& \text{stimulus\hspace{0.17em}absent.}\end{array}$$

(9)

Here, *B* [0,1] represents baseline activity and λ≥ 1 controls the rate at which activity decays with *d*. Both *d* and *r* were scaled so that their maximum value is 1 (see Figure Figure22A).

We assume that the correlation between the responses of two neurons is proportional to the geometric mean of their firing rates (de la Rocha et al., 2007; Shea-Brown et al., 2008), and that correlations decay exponentially with cell distance (Smith and Kohn, 2008; see however Poort and Roelfsema, 2009; Ecker et al., 2010). We therefore model the correlation between two cells as ${\text{\rho}}_{\mathrm{jk}}=S\sqrt{r\left({d}_{j}\right)r\left({d}_{k}\right)}{\text{e}}^{-\alpha {D}_{j,k}}$ where *d _{j}* and

If pixels are small compared to the scales at which correlations are assumed to decay, then the average correlation between cells within the same pixel are ${\overline{\text{\rho}}}_{11}=Sr({d}_{1})$ and ${\overline{\text{\rho}}}_{22}=Sr({d}_{2})$. The average correlation between cells in different pixels is ${\overline{\text{\rho}}}_{12}=S\sqrt{r\left({d}_{1}\right)r\left({d}_{2}\right)}{\text{e}}^{-\alpha {D}_{1,2}}.$

In this case, whether a stimulus is present or not, the correlation between the pooled signals is of the form ${\text{\rho}}_{{X}_{1}{X}_{2}}={\text{e}}^{-\alpha {D}_{1,2}}+\mathcal{O}(1/n).$ Thus, even significant stimulus dependent changes in correlations would be invisible in the recorded signals. This overall trend is consistent with the results in Chen et al. (2006) (compare Figure Figure2C2C to their Figure 2f). In such settings, it is difficult to conclude whether pairwise correlations are stimulus dependent or not from the pooled data.

However, in Supplementary Figure 3 of Chen et al. (2006) the presence of a stimulus apparently results in a slight decrease in correlations between more distant pixels. In Figure Figure2D2D this effect was reproduced using the alternative model described above, with the additional assumption that baseline activity, *B*, decreases in the presence of a stimulus (Mitchell et al., 2009). The effect can also be reproduced by assuming that spatial correlation decay, α, increases when a stimulus is present.

As this example shows, care needs to be taken when inferring underlying correlation structures from pooled activity. The statistical structure of the recordings can depend on pairwise correlations between individual cells in a subtle way, and different underlying correlation structures may be difficult to distinguish from the pooled signals. However, downstream neurons may also be insensitive to the precise structure of pairwise correlations, as they are driven by the pooled input from many afferents.

We next examine the effects of pooling by relating the correlations between the activity of downstream cells to the pairwise correlations between cells in the input populations (see Figure Figure1B).1B). The idea that pooling amplifies correlations carries over from the previous section. However, the presence of inhibition and non- instantaneous synaptic responses introduces new issues.

For simplicity, we first consider a homogeneous population model (see Figure Figure3A).3A). Each cell receives *n*_{e} inputs from a homogeneous pool of inputs with pairwise correlation coefficients ρ_{ee} and an additional *q*_{e}*n*_{e} inputs from an outside pool of independent inputs. The two cells share *p*_{e}*n*_{e} of the inputs drawn from the correlated pool. Processes in the independent pool are uncorrelated with all other processes. All excitatory inputs have variance ${\sigma}_{\text{e}}^{2}.$

The correlation between the pooled excitatory inputs is given by (see Appendix)

$${\rho}_{{\text{E}}_{1}{\text{E}}_{2}}=\frac{{\rho}_{\text{ee}}+\frac{{p}_{\text{e}}}{{n}_{\text{e}}}\left(1-{\rho}_{\text{ee}}\right)}{{\rho}_{\text{ee}}+\frac{1}{{n}_{\text{e}}}\left(1-{\rho}_{\text{ee}}+{q}_{\text{e}}\right)}.$$

(10)

A form of this equation, with *p*_{e}=0 and *q*_{e}=0, is derived in Chen et al. (2006). In the absence of correlations between processes in the input pools, ρ_{ee}=0, the correlation between the pooled signals is just the proportion of shared inputs, ${\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}}={p}_{\text{e}}.$ When ρ_{ee}>0 and *n*_{e} is large, pooled excitatory inputs are highly correlated, even when pairwise correlations in the presynaptic pool, ρ_{ee}, are small, and the neurons do not share inputs (*p*_{e}=0). Even when most inputs to the downstream cells are independent (*q*_{e}>1), correlations between the pooled signals will be nearly 1 for sufficiently large input pools (see Figure Figure44A).

Under analogous homogeneity assumptions for the inhibitory pools, the correlation, ${\text{\rho}}_{{\text{I}}_{1}{\text{I}}_{2}},$ between the pooled inhibitory inputs is given by an equation identical to Eq. (10), and the correlation between the pooled excitatory and inhibitory inputs is given by

$${\rho}_{{\text{E}}_{1}{\text{I}}_{2}}={\rho}_{{\text{I}}_{1}{\text{E}}_{2}}=\frac{{\rho}_{\text{ei}}}{\sqrt{\left({\rho}_{\text{ee}}+\frac{1}{{n}_{e}}\left(1-{\rho}_{ee}+{q}_{e}\right)\right)\left({\rho}_{\text{ii}}+\frac{1}{{n}_{\text{i}}}\left(1-{\rho}_{\text{ii}}+{q}_{\text{i}}\right)\right)}}.$$

(11)

Interestingly, since $\left|{\text{\rho}}_{{\text{E}}_{1}{\text{I}}_{2}}\right|\text{\hspace{0.17em}}\le 1,$ pairwise excitatory– inhibitory correlations obey the bound $\left|{\text{\rho}}_{\text{ei}}\right|\le \sqrt{{\text{\rho}}_{\text{ee}}{\text{\rho}}_{\text{ii}}}+\mathcal{O}(1/\sqrt{{n}_{\text{e}}{n}_{\text{i}}}).$ Combining this inequality with Eq. (10) and the analogous equation for ${\text{\rho}}_{{\text{I}}_{1}{\text{I}}_{2}},$ it follows that $\left|{\text{\rho}}_{{\text{E}}_{1}{\text{I}}_{2}}\right|\le \sqrt{{\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}}{\text{\rho}}_{{\text{I}}_{1}{\text{I}}_{2}}}+\mathcal{O}\left(1/\sqrt{{n}_{\text{e}}{n}_{\text{i}}}\right)$ for homogeneous populations. These are a result of the non-negative definiteness of covariance matrices.

We next discuss how heterogeneity can dampen the amplification of correlations due to pooling. In the absence of any homogeneity assumptions on the excitatory input population (see the population model in the Materials and Methods), Eq. (3) gives the pooled excitatory signals, ${\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}}={\overline{\text{\rho}}}_{{\text{e}}_{1}{\text{e}}_{2}}/\sqrt{{\overline{\text{\rho}}}_{{\text{e}}_{1}{\text{e}}_{2}}}+\mathcal{O}\left(1/\sqrt{{n}_{{\text{e}}_{1}}{n}_{{\text{e}}_{2}}}\right).$ The term ${\overline{\text{\rho}}}_{{\text{e}}_{1}{\text{e}}_{2}}$ is a weighted average of the correlation coefficients *between* the two excitatory populations, and ${\overline{\text{\rho}}}_{{\text{e}}_{1}{\text{e}}_{1}}$ and ${\overline{\text{\rho}}}_{{\text{e}}_{2}{\text{e}}_{2}}$ are weighted averages of the correlations *within* each excitatory input population.

To illuminate this result, we assume symmetry between the populations: Let ${n}_{{\text{e}}_{k}}={n}_{\text{e}}$ and ${\text{\sigma}}_{{\text{e}}_{k}^{j}}={\sigma}_{\text{e}}$ for *k*=1,2 and *j*=1,2,…,*n*_{e}, and assume ${\overline{\text{\rho}}}_{{\text{e}}_{1}{\text{e}}_{1}}={\overline{\text{\rho}}}_{{\text{e}}_{2}{\text{e}}_{2}}.$ The average “within” and “between” correlations, are ${\text{\rho}}_{\text{ee}}^{\text{w}}={\overline{\text{\rho}}}_{{\text{e}}_{1}{\text{e}}_{1}}={\overline{\text{\rho}}}_{{\text{e}}_{2}{\text{e}}_{2}}$ and ${\text{\rho}}_{\text{ee}}^{\text{b}}={\overline{\text{\rho}}}_{{\text{e}}_{1}{\text{e}}_{2}}$ respectively (see Figure Figure3B).3B). Under these assumptions, Eq. (5) can be applied to obtain (See also Bedenbaugh and Gerstein, 1997)

$${\rho}_{{\text{E}}_{1}{\text{E}}_{2}}=\frac{{\rho}_{\text{ee}}^{\text{b}}}{{\rho}_{\text{ee}}^{\text{w}}+\frac{1}{{n}_{\text{e}}}\left(1-{\rho}_{\text{ee}}^{\text{w}}\right)}=\frac{{\rho}_{\text{ee}}^{\text{b}}}{{\rho}_{\text{ee}}^{\text{w}}}+\mathcal{O}\left(\frac{1}{{n}_{\text{e}}}\right)$$

(12)

which is plotted in Figure Figure4A4A (green line) and Figure Figure4B.4B. For large *n*_{e}, the correlation between the pooled signals is the ratio of “between” and “within” correlations.

This observation has implications for a situation ubiquitous in the cortex. A neuron is likely to receive afferents from cells that are physically close. The activity of nearby cells may be more strongly correlated than the activity of more distant cells (Chen et al., 2006; Smith and Kohn, 2008). We therefore expect that pairwise correlations *within* each input pool are on average larger than correlations *between* two input pools, that is, ${\text{\rho}}_{\text{ee}}^{\text{w}}>{\text{\rho}}_{\text{ee}}^{\text{b}}.$ This reduces the correlation between the inputs, regardless of the input population size.

An increase in correlations in the presynaptic pool can also decorrelate the pooled signals. If correlations *within* each input pool increase by a greater amount than correlations *between* the two pools, then the variance in the input to each cell will increased by a larger amount than the covariance between the inputs. As a consequence the correlations between the pooled inputs will be reduced. Modulations in correlation have been observed as a consequence of attention in V4 (Cohen and Maunsell, 2009; Mitchell et al., 2009; but apparently not in V1, Roelfsema et al., 2004). Such changes may be, in part, a consequence of small changes in “within” correlations between neurons in V1.

Equation 12 implies that correlations between large populations cannot be significantly larger than the correlations within each population. Since $\left|{\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}}\right|\text{\hspace{0.17em}}\le 1,$ it follows that $\left|{\text{\rho}}_{\text{ee}}^{\text{b}}\right|\text{\hspace{0.17em}}\le \text{\hspace{0.17em}}\left|{\text{\rho}}_{\text{ee}}^{\text{w}}\right|+\mathcal{O}(1/{n}_{\text{e}}).$

The correlation, ${\text{\rho}}_{{\text{I}}_{1}{\text{I}}_{2}},$ between the pooled inhibitory inputs is given by an identical equation to Eq. (12) and the correlation between the pooled excitatory and inhibitory inputs is given by

$$\begin{array}{c}{\rho}_{{\text{E}}_{1}{\text{I}}_{2}}={\rho}_{{\text{I}}_{1}{\text{E}}_{2}}=\frac{{\rho}_{\text{ei}}^{\text{b}}}{\sqrt{\left({\rho}_{\text{ee}}^{\text{w}}+\frac{1}{{n}_{e}}\left(1-{\rho}_{\text{ee}}^{\text{w}}\right)\right)\left({\rho}_{\text{ii}}^{\text{w}}+\frac{1}{{n}_{\text{i}}}\left(1-{\rho}_{\text{ii}}^{\text{w}}\right)\right)}}\\ =\frac{{\rho}_{\text{ei}}^{\text{b}}}{\sqrt{{\rho}_{\text{ee}}^{\text{w}}{\rho}_{\text{ii}}^{\text{w}}}}+\mathcal{O}\left(\frac{\text{1}}{\sqrt{{n}_{\text{e}}{n}_{\text{i}}}}\right).\end{array}$$

We now look at the correlation between the free membrane potentials of two downstream neurons. The *free* membrane potentials are obtained by assuming an absence of threshold or spiking activity. For simplicity we assume symmetry in the statistics of the inputs to the postsynaptic cells: ${\text{\sigma}}_{{\text{E}}_{k}}\text{\hspace{0.17em}}={\sigma}_{\text{E}},$${\text{\sigma}}_{{\text{I}}_{k}}\text{\hspace{0.17em}}={\sigma}_{\text{I}},$${\text{\rho}}_{{\text{E}}_{1}{\text{I}}_{2}}\text{\hspace{0.17em}}={\text{\rho}}_{{\text{E}}_{2}{\text{I}}_{1}},$${\text{\rho}}_{{\text{E}}_{1}{\text{I}}_{1}}\text{\hspace{0.17em}}={\text{\rho}}_{{\text{E}}_{2}{\text{I}}_{2}},$${\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{1}}\text{\hspace{0.17em}}={\text{\rho}}_{{\text{E}}_{2}{\text{E}}_{2}}$ and ${\text{\rho}}_{{\text{I}}_{1}{\text{I}}_{1}}\text{\hspace{0.17em}}={\text{\rho}}_{{\text{I}}_{2}{\text{I}}_{2}}.$ The analysis is similar in the asymmetric case.

In the Section “Materials and Methods”, we derive a linear approximation of the free membrane potentials,

$${V}_{k}(t)=\left({J}_{k}\ast K\right)(t)+{V}_{L},$$

where ${J}_{k}(t)=-({g}_{{\text{E}}_{k}}(t)({V}_{L}-{V}_{\text{E}})+{g}_{{\text{I}}_{k}}(t)({V}_{L}-{V}_{\text{I}}))/{C}_{m}$ are the total input currents and $K(t)=\Theta (t){\text{e}}^{-t/{\tau}_{\text{eff}}}$ for *k*=1,2. Under this approximation, the correlation, ${\text{\rho}}_{{V}_{1}{V}_{2}},$ between the membrane potentials is equal to the correlation, ${\text{\rho}}_{\text{in}}={\text{\rho}}_{{J}_{1}{J}_{2}},$ between the total input currents and can be written as a weighted average of the pooled excitatory and inhibitory spike train correlations (see Appendix),

$${\rho}_{{V}_{1}{V}_{2}}\approx {\rho}_{\text{in}}=\frac{{W}_{\text{E}}^{2}{\rho}_{{\text{E}}_{1}{\text{E}}_{2}}+{W}_{\text{I}}^{2}{\rho}_{{\text{I}}_{1}{\text{I}}_{2}}-2{W}_{\text{E}}{W}_{\text{I}}{\rho}_{{\text{E}}_{1}{\text{I}}_{2}}}{{W}_{\text{E}}^{2}+{W}_{\text{I}}^{2}-2{W}_{\text{E}}{W}_{\text{I}}{\rho}_{{\text{E}}_{1}{\text{I}}_{1}}}$$

(13)

where ${\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}},\text{\hspace{0.17em}}{\text{\rho}}_{{\text{E}}_{1}{\text{I}}_{2}},\text{\hspace{0.17em}}{\text{\rho}}_{{\text{I}}_{1}{\text{I}}_{2}}$ and ${\text{\rho}}_{{\text{E}}_{1}{\text{I}}_{1}}$ are derived above, and *W*_{E}=|*V*_{E}−*V*_{L}|σ_{E} and *W*_{I}=|*V*_{I}−*V _{L}*|σ

The correlation between the membrane potentials has *positive* contributions from the correlation between the excitatory inputs $({\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}}),$ and between the inhibitory inputs $({\text{\rho}}_{{\text{I}}_{1}{\text{I}}_{2}}).$ Contributions coming from excitatory–inhibitory correlations (${\text{\rho}}_{{\text{E}}_{1}{\text{I}}_{2}}$ and ${\text{\rho}}_{{\text{E}}_{2}{\text{I}}_{1}}$) are negative, and can thus decorrelate the activity of downstream cells. This “cancellation” of correlations is observed in Figures Figures1D1D and and5,5, and can lead to asynchrony in recurrent networks (Hertz, 2010; Renart et al., 2010).

Feedforward chains, like that depicted in Figure Figure6A,6A, have been studied extensively (Diesmann et al., 1999; van Rossum et al., 2002; Litvak et al., 2003; Reyes, 2003; Tetzlaff et al., 2003; Câteau and Reyes, 2006; Doiron et al., 2006; Kumar et al., 2008). In such networks, cells in a layer necessarily share some of their inputs, leading to correlations in their spiking activity (Shadlen and Newsome, 1998). Frequently, spiking in deeper layers is highly synchronous (Reyes, 2003; Tetzlaff et al., 2003). However, in the presence of background noise, correlations can remain negligible (van Rossum et al., 2002; Vogels and Abbott, 2005).

Feedforward chains amplify correlations as follows: When inputs to the network are independent, small correlations are introduced in the second layer by overlapping inputs. The inputs to each subsequent layer are pooled from the previous layer. The amplification of correlations by pooling is the primary mechanism for the development of synchrony (Compare solid and dotted blue lines in Figure Figure4A).4A). Overlapping inputs serve primarily to “seed” synchrony in early layers. The internal dynamics of the neurons and background noise can decorrelate the output of a layer, and compete with the correlation amplification due to pooling.

We develop this explanation by considering a feedforward network with each layer containing *N*_{e} excitatory and *N*_{i} inhibitory cells. Each cell in layer *k*+1 receives *n*_{e} excitatory and *n*_{i} inhibitory inputs selected randomly from layer *k*. For simplicity we assume that all excitatory and inhibitory cells are dynamically identical and |*V*_{E}−*V _{L}*|=|

To explain the development of correlations, we consider a simplified model of correlation propagation (See also Renart et al., 2010 for a recurrent version). In the model, any two cells in a layer share the expected proportion *p*_{e}=*n*_{e}/*N*_{e} of their excitatory inputs and *p*_{i}=*n*_{i}/*N*_{i} of their inhibitory inputs (the expected proportions are taken with respect to random connectivity). We also assume that inputs are statistically identical across a layer.

For a pair of cells in layer *k*≥ 1, let ${\text{\rho}}_{k}^{\text{in}}$ and ${\text{\rho}}_{k}^{\text{out}}$ represent the correlation coefficient between the total input currents and output spike trains respectively. The outputs from layer *k* are pooled (with overlap) to obtain the inputs to layer *k*+1. Using the results developed above, ${\text{\rho}}_{\text{in}}^{1}=P({\text{\rho}}_{0})$ and ${\text{\rho}}_{k+1}^{\text{in}}=P({\text{\rho}}_{k}^{\text{out}}),$ for *k*≥ 1, where (see Appendix and Tetzlaff et al., 2003 for a similar derivation)

$$P(\rho )=\frac{\rho {(\beta -1)}^{2}+\frac{1}{{n}_{\text{i}}}(1-\rho )\left(\beta {p}_{\text{e}}+{p}_{\text{i}}\right)}{\rho {(\beta -1)}^{2}+\frac{1}{{n}_{\text{i}}}(1-\rho )(1+\beta )}.$$

(14)

Here β measures the balance between excitation and inhibition (see Materials and Methods). From our assumptions, β=*n*_{e}/*n*_{i}. With imbalance (β≠1) and a large number of cells in a layer, pooling amplifies small correlations, *P*(ρ)>ρ, as discussed earlier.

To complete the description of correlation transfer from layer to layer, we relate the correlations between inputs to a pair of cells, ${\text{\rho}}_{k}^{\text{in}},$ to correlations in their output spike trains, ${\text{\rho}}_{k}^{\text{out}}.$ We assume that there is a transfer function, *S*, so that ${\text{\rho}}_{k}^{\text{out}}=S({\text{\rho}}_{k}^{\text{in}})$ at each layer *k*. We additionally assume that *S*(0)=0 and *S*(1)=1, that is uncorrelated (perfectly correlated) inputs result in uncorrelated (perfectly correlated) outputs. We also assume that the cells are decorrelating, |ρ|>|*S*(ρ)|>0 for ρ≠0,1 (Shea-Brown et al., 2008). This is an idealized model of correlation transfer, as output correlations depend on cell dynamics and higher order statistics of the inputs (Moreno-Bote and Parga, 2006; de la Rocha et al., 2007; Barreiro et al., 2009; Ostojić et al., 2009).

Correlations between the spiking activity of cells in layers *k*+1 are related to correlations in layer *k* by the layer-to-layer transfer function, *T*=*S*○*P*. The development of correlations across layers is modeled by the dynamical system, ${\text{\rho}}_{k+1}^{\text{out}}=T({\text{\rho}}_{k}^{\text{out}}),$ with ${\text{\rho}}_{1}^{\text{out}}=S({\text{\rho}}_{0}).$

When the network is not balanced (β≠1), pooling amplifies correlations at each layer and the activity between cells in deeper layers can become highly correlated (see Figure Figure6C).6C). The output of the first layer is uncorrelated if the individual inputs are independent (ρ_{0}=0). In this case all of the correlations between the total inputs to the second layer come from shared inputs,

$${\rho}_{2}^{\text{in}}=P(0)=\frac{{n}_{\text{e}}{p}_{\text{e}}+{n}_{\text{i}}{p}_{\text{i}}}{{n}_{\text{e}}+{n}_{\text{i}}}.$$

These correlations are then reduced by the second layer of cells, ${\text{\rho}}_{2}^{\text{out}}=S({\text{\rho}}_{2}^{\text{in}})=T(0)>0,$ and subsequently amplified by pooling and input sharing before being received by layer 3, ${\rho}_{\text{3}}^{\text{in}}=P({\rho}_{\text{2}}^{\text{out}}).$ This process continues in subsequent layers. If the correlating effects of pooling and input sharing dominate the decorrelating effects of internal cell dynamics, correlations will increase from layer to layer (see Figure Figure66C).

When ρ_{0}=0, overlapping inputs increase the input correlation to layer 2, but have a negligible effect on the mapping once correlations have developed since the effects of pooling dominate [see Eq. (14) and the dashed blue line in Figure Figure4A4A which shows that the effects of input overlaps are small when *n*_{e} is large, ρ>0 and β≠1]. Therefore, shared inputs seed correlated activity at the first layer, and pooling drives the development of larger correlations. When ρ_{0}=0, we cannot expect large correlations before layer 3, but when ρ_{0}>0 large correlations can develop by layer 2.

To verify this conclusion, we constructed a two-layer feedforward network with no overlap between inputs (*P*_{e}=*P*_{i}=0). In Figure Figure7A,7A, the inputs to layer 1 were independent (ρ_{0}=0), and the firing of cells in layer 2 was uncorrelated. In Figure Figure7B,7B, we introduced small correlations (ρ_{0}=0.05) between inputs to layer 1. These correlations were amplified by pooling so that strong synchrony is observed between cells in layer 2. We compared these results with a standard feedforward network with overlap in cell inputs (Figure (Figure7C,7C, where *P*_{e}=*P*_{i}=0.05). Inputs to layer 1 were independent (ρ_{0}=0), and hence outputs from layer 1 uncorrelated. Dependencies between inputs to layer 2 were weak and due to overlap alone, ${\rho}_{\text{2}}^{\text{in}}=P(0)=\mathrm{0.05.}$ Cells in layer 3 received pooled inputs from layer 2, and their output was highly correlated.

These results predict that correlations between spike trains develop in deeper layers, but they do not not directly address the timescale of the correlated behavior. In simulations, spiking becomes tightly synchronized in deeper layers (see for instance Litvak et al., 2003; Reyes, 2003; and Figure Figure7).7). This can be understood using results in Maršálek et al. (1997) and Diesmann et al. (1999) where it is shown that the response of cells to volleys of spikes is tighter than the volley itself. The firing of individual cells in the network becomes bursty in deeper layers and large correlations are manifested in tightly synchronized spiking events. Alternatively, one can predict the emergence of synchrony by observing that pooling increases correlations over finite time windows (see next section and Appendix) and therefore the analysis developed above can be adapted to correlations over small windows.

In the simplified feedforward model above, when excitation balances inhibition, that is β≈1, correlations between the pooled inputs to a layer are due to overlap alone, ${\rho}_{\text{k}}^{\text{in}}=P({\text{\rho}}_{k-1}^{\text{out}})\approx ({p}_{\text{e}}+{p}_{\text{i}})/2$ for all *k*. The correlating effects of this map are weak, and this would seem to imply that cells in balanced feedforward chains remain asynchronous. Indeed, our model of correlation propagation displays a stable fixed point at low values of ρ when β≈1 (see Figure Figure6D).6D). However, in practice, synchrony is difficult to avoid without careful fine-tuning (Tetzlaff et al., 2003), and almost always develops in feedforward chains (Litvak et al., 2003). We provide some reasons for this discrepancy.

Our focus so far has been on correlations over infinitely large time windows (see Materials and Methods where we define ρ* _{xy}*). Even when the membrane potentials are nearly uncorrelated over large time windows, differences between the excitatory and inhibitory synaptic time constants can cause larger correlations over smaller time windows (Renart et al., 2010). This can, in turn, lead to significant correlations between the output spike trains. We discuss this effect further in the Appendix and give an example in Figure Figure8.8. In this example, the correlations between the membrane potentials over long windows are nearly zero due to cancellation (see Figure Figure8A8A where ρ

Another source of discrepancies between the idealized model and simulations of feedforward networks are inhomogeneities, which become important when balance is exact. Note that Eq. (14) is an approximation obtained by ignoring fluctuations in connectivity from layer to layer. In a random network, inhomogeneities will be introduced by variability in input population overlaps. To fully describe the development of correlations in a feedforward network, it is necessary to include such fluctuations in a model of correlation propagation. The asynchronous fixed point that appears in the balanced case has a small basin of attraction and fluctuations induced by input inhomogeneities could destroy its stability (see Figure Figure6D).6D). Other sources of heterogeneity can further destabilize the asynchronous state (see Appendix).

It has been shown that asynchronous states can be stabilized through the decorrelating effects of background noise (van Rossum et al., 2002; Vogels and Abbott, 2005). To emulate these effects, a third transfer function, *N*, can be added to our model. The correlation transfer map then becomes *T*(ρ)=*S**N**P*(ρ). Sufficiently strong background noise can increase decorrelation from input to output of a layer, and stabilize the asynchronous fixed point.

We have illustrated how pooling and shared inputs can impact correlations between the inputs and free membrane voltages of postsynaptic cells in a feedforward setting. The increase in correlation due to pooling was discussed in a simpler setting in (Bedenbaugh and Gerstein, 1997; Super and Roelfsema, 2005; Chen et al., 2006; Stark et al., 2008), and similar ideas were also developed for the variance alone in (Salinas and Sejnowski, 2000; Moreno-Bote et al., 2008). The saturation of the signal-to-noise ratio with increasing population size observed in (Zohary et al., 1994) has a similar origin. Our aim was to present a unified discussion of these results, with several generalizations.

Other mechanisms, such as recurrent connectivity between cells receiving the inputs, can modulate correlated activity (Schneider et al., 2006; Ostojić et al., 2009). Importantly, the cancellation of correlations may be a dynamic phenomenon in recurrent networks, as observed in (Hertz, 2010; Renart et al., 2010). On the other hand, neurons may become entrained to network oscillations, resulting in more synchronous firing (Womelsdorf et al., 2007). A full understanding of the statistics of population activity in neuronal networks will require an understanding of how these mechanisms interact to shape the spatiotemporal properties of the neural response.

The results we presented relied on the assumption of linearity at the different levels of input integration. These assumptions can be expected to hold at least approximately. For instance, there is evidence that membrane conductances are tuned to produce a linear response in the subthreshold regime (Morel and Levy, 2009). The assumptions we make are likely to break down at the level of single dendrites where nonlinear effects may be much stronger (Johnston and Narayanan, 2008). The effects of correlated inputs to a single dendritic branch deserve further theoretical study (Gasparini and Magee, 2006; Li and Ascoli, 2006).

We demonstrated that the structure of correlations in a population may be difficult to infer from pooled activity. For instance, a change in pairwise correlations between individual cells in two populations causes a much smaller change in the correlation between the pooled signals. With a large number of inputs, the change in correlations between the pooled signals might not be detectable even when the change in the pairwise correlations is significant.

While we discussed the growth of second order correlations only, higher order correlations also saturate with increasing population size. For example, in a 3-variable generalization of the homogeneous model from Figure Figure3A,3A, it can be shown that ${\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}{\text{E}}_{3}}=1-\mathcal{O}(1/{n}_{\text{e}})$ where *n*_{e} is the size of each population and ${\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}{\text{E}}_{3}}$ is the triple correlation coefficient (Stratonovich, 1963) between the pooled signals E_{1}, E_{2}, and E_{3}. The reason that higher order correlations also saturate follows from the generalization of the following observation at second order: Pooling amplifies correlations because the variance and covariance grow asymptotically with the same rate in *n*_{e}. In particular ${\sigma}_{\text{E}}^{2}$ and ${\text{\gamma}}_{{\text{E}}_{1}{\text{E}}_{2}}$ both behave asymptotically like ${n}_{\text{e}}^{2}{\text{\rho}}_{\text{ee}}{\sigma}_{\text{e}}^{2}+\mathcal{O}({n}_{\text{e}}),$ and their ratio, ${\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}}={\text{\gamma}}_{{\text{E}}_{1}{\text{E}}_{2}}/{\sigma}_{\text{E}}^{2},$ approaches unity (Bedenbaugh and Gerstein, 1997; Salinas and Sejnowski, 2000; Moreno-Bote et al., 2008).

We concentrated on correlations over infinitely long time windows (see Materials and Methods where we define ρ* _{xy}*). However, pooling amplifies correlations over finite time windows in exactly the same way as correlations over large time windows. Due to the filtering properties of the cells, the timescale of correlations between downstream membrane potentials may not reflect that of the inputs. We discuss this further in the Appendix where the auto- and cross-covariance functions between the membrane potentials are derived.

To simplify the presentation, we have so far assumed stationary. However, since Eq. (2) applies to the Pearson correlation between any pooled data, all of the results on pooling can easily be extended to the non-stationary case. In the non-stationary setting, the cross- covariance function has the form *R _{xy}*(

Correlated neural activity has been observed in a variety of neural populations (Gawne and Richmond, 1993; Zohary et al., 1994; Vaadia et al., 1995), and has been implicated in the propagation and processing of information (Oram et al., 1998; Maynard et al., 1999; Romo et al., 2003; Tiesinga et al., 2004; Womelsdorf et al., 2007; Stark et al., 2008), and attention (Steinmetz et al., 2000; Mitchell et al., 2009). However, correlations can also introduce redundancy and decrease the efficiency with which networks of neurons represent information (Zohary et al., 1994; Gutnisky and Dragoi, 2008; Goard and Dan, 2009). Since the joint response of cells and recorded signals can reflect the activity of large neuronal populations, it will be important to understand the effects of pooling to understand the neural code (Chen et al., 2006).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

We thank Jaime de la Rocha, Brent Doiron and Eric Shea-Brown for helpful discussions. We also thank the reviewers and the handling editor for numerous useful suggestions. This work was supported by NSF Grants DMS-0604429 and DMS-0817649 and a Texas ARP/ATP award.

Equation (10) can be derived from Eq. (2). However, we find that it is more easily derived directly. We will calculate the variance, ${\text{\sigma}}_{{\text{E}}_{1}}^{\text{2}}={\text{\sigma}}_{{\text{E}}_{2}}^{\text{2}},$ and covariance ${\text{\gamma}}_{{\text{E}}_{1}{\text{E}}_{2}}$ between the pooled signals.

The covariance is given by the sum of all pairwise covariances between the populations, ${\text{\gamma}}_{{\text{E}}_{1}{\text{E}}_{2}}={\Sigma}_{{\text{e}}_{1}\in {\text{E}}_{1}{\text{,e}}_{\text{2}}\in {\text{E}}_{2}}{\text{\sigma}}_{{\text{e}}_{1}}{\text{\sigma}}_{{\text{e}}_{2}}{\text{\rho}}_{{\text{e}}_{1}{\text{e}}_{2}}.$ Each cell receives ${n}_{\text{e}}+{q}_{\text{e}}{n}_{\text{e}}$ inputs so that there are (*n*_{e}+*q*_{e}*n*_{e})^{2} terms that appear in this sum. However, the *q*_{e}*n*_{e} “independent” inputs from each pool are uncorrelated with all other inputs and therefore don't contribute to the sum. Of the remaining ${n}_{\text{e}}^{2}$ pairs, *n*_{e}*p*_{e} are shared and therefore have correlation ${\text{\rho}}_{{\text{e}}_{1}{\text{e}}_{2}}=1.$ These shared processes therefore collectively contribute ${n}_{\text{e}}{p}_{\text{e}}{\sigma}_{\text{e}}^{2}$ to ${\text{\gamma}}_{{\text{E}}_{1}{\text{E}}_{2}}.$ The remaining ${n}_{\text{e}}^{2}-{n}_{\text{e}}{p}_{\text{e}}$ processes are correlated with coefficient ρ_{ee} and collectively contribute $({n}_{\text{e}}^{2}-{n}_{\text{e}}{p}_{\text{e}}){\text{\rho}}_{\text{ee}}{\sigma}_{\text{e}}^{2}.$ The pooled covariance is thus

$${\gamma}_{{\text{E}}_{1}{\text{E}}_{2}}=\underset{\text{Correlated}}{\underbrace{({n}_{\text{e}}^{2}-{p}_{\text{e}}{n}_{\text{e}}){\rho}_{\text{ee}}{\sigma}_{\text{e}}^{2}}}+\underset{\text{Shared}}{\underbrace{{n}_{\text{e}}{p}_{\text{e}}{\sigma}_{\text{E}}^{2}}}.$$

The variance is given by the sum of all pairwise covariances within a population, ${\text{\sigma}}_{{\text{E}}_{1}}^{\text{2}}={\Sigma}_{{\text{e}}_{1}\in {\text{E}}_{1},{\text{e}}_{2}\in {\text{E}}_{1}}{\text{\sigma}}_{{\text{e}}_{1}}{\text{\sigma}}_{{\text{e}}_{2}}{\text{\rho}}_{{\text{e}}_{1}{\text{e}}_{2}}.$ As above, there are *n*_{e}+*q*_{e}*n*_{e} neurons in the population, so that the sum has (*n*_{e}+*q*_{e}*n*_{e})^{2} terms. Of these, *n*_{e}+*q*_{e}*n*_{e} are “diagonal” terms (e_{1}=e_{2}), each contributing ${\sigma}_{\text{e}}^{2},$ for a total contribution of $({n}_{\text{e}}+{q}_{\text{e}}{n}_{\text{e}}){\sigma}_{\text{e}}^{2}$ to ${\text{\sigma}}_{{\text{E}}_{1}}^{2}.$ The processes from the independent pool do not contribute any additional terms. This leaves *n*_{e}(*n*_{e}−1) correlated pairs which each contribute ${\sigma}_{\text{e}}^{2}{\text{\rho}}_{\text{ee}}$ for a collective contribution of ${n}_{\text{e}}({n}_{\text{e}}-1){\sigma}_{\text{e}}^{2}{\text{\rho}}_{\text{ee}},$ giving

$${\sigma}_{{\text{E}}_{1}}={\sigma}_{{\text{E}}_{2}}=\underset{\text{Correlated}}{\underbrace{{n}_{\text{e}}({n}_{\text{e}}-1){\sigma}_{\text{e}}^{2}{\rho}_{\text{ee}}}}+\underset{\text{Diagonal}}{\underbrace{({n}_{\text{e}}+{n}_{\text{e}}{q}_{\text{e}}){\sigma}_{\text{e}}^{2}}}.$$

Now, ${\text{\rho}}_{{\text{E}}_{1}{\text{E}}_{2}}={\text{\gamma}}_{{\text{E}}_{1}{\text{E}}_{2}}/{\text{\sigma}}_{{\text{E}}_{1}}$ can be simplified to give Eq. (10). Equations for ${\text{\rho}}_{{\text{I}}_{1}{\text{I}}_{2}}$ and ${\text{\rho}}_{{\text{E}}_{1}{\text{I}}_{2}}={\text{\rho}}_{{\text{I}}_{1}{\text{E}}_{2}}$ can be derived identically.

Throughout the text, we concentrated on correlations over large time windows. However, the effects of pooling described by Eq. (2) apply to the correlation, ρ* _{xy}*(

However, the cell filters the pooled inputs to obtain the membrane potentials and, as a result, the correlations between membrane potentials is “spread out” in time (Tetzlaff et al., 2008). To quantify this effect, we derive an approximation to the auto- and cross-covariance functions between the membrane potentials.

The pooled input spike trains are obtained from from a weighted sum of the individual excitatory and inhibitory spike trains (see Materials and Methods). As a result cross-covariance functions between the pooled spike trains are just sums of the individual cross-covariance functions, ${C}_{XY}(t)={\Sigma}_{x\in X,y\in Y}{C}_{xy}(t)$ for *X*, *Y*=E_{1}, E_{2}, I_{1}, I_{2} and *x*, *y*=e, i accordingly. Thus only the magnitude of the cross- covariance functions is affected by pooling. The change in magnitude is quadratic in *n*_{e} or *n*_{i}. This is consistent with the observation that pooling amplifies correlations equally over all timescales.

The conductances are obtained by convolving the total inputs with the synaptic filter kernels,

$${g}_{{\text{E}}_{k}}(t)=({\text{E}}_{k}\ast {\alpha}_{\text{e}})(t),\text{\hspace{1em}and\hspace{1em}}{g}_{{\text{I}}_{k}}(t)=({\text{I}}_{k}\ast {\alpha}_{i})(t);\text{\hspace{1em}}k=1,2.$$

The cross-covariance between the conductances can therefore be written as a convolution of the cross-covariance function between the input signals and the deterministic cross-covariance between the synaptic kernels (Tetzlaff et al., 2008). In particular,

$${C}_{{g}_{X}{g}_{Y}}(t)=({C}_{XY}\ast ({\alpha}_{x}\star {\alpha}_{y}))(t)$$

(15)

for *X*, *Y*=E_{1}, E_{2}, I_{1}, I_{2} and *x*, *y*=e, i accordingly, where $({\alpha}_{x}\star \text{\hspace{0.17em}}{\alpha}_{y})(t)={\int}_{-\infty}^{\infty}{\alpha}_{x}(s){\alpha}_{y}(t+s)\text{\hspace{0.17em}d}s$ is the deterministic cross- covariance between the synaptic filters, α* _{x}* and α

The total input currents, ${J}_{K}\left(t\right)=-\left({g}_{{E}_{k}}\left(t\right)\left({V}_{L}-{V}_{I}\right)\right)/{C}_{m},$ obtained from the linearization of the conductance-based model described in the Section “Materials and Methods” are simply linear combinations of the individual conductances. The cross-covariance function between the input currents is therefore a linear combination of those between the conductances,

$$\begin{array}{l}{C}_{{J}_{h}{J}_{k}}(t)\text{\hspace{0.17em}}=\text{\hspace{0.17em}}|{V}_{\text{E}}-{V}_{L}{|}^{2}{C}_{{g}_{{\text{E}}_{h}}{g}_{{\text{E}}_{k}}}(t)\text{\hspace{0.17em}}+\text{\hspace{0.17em}}|{V}_{\text{I}}-{V}_{L}{|}^{2}{C}_{{g}_{{\text{I}}_{h}}{g}_{{\text{I}}_{k}}}(t)\\ \text{\hspace{1em}\hspace{1em}\hspace{1em}\hspace{0.17em}\hspace{0.17em}}-2|{V}_{\text{E}}-{V}_{L}||{V}_{\text{I}}-{V}_{L}|{C}_{{g}_{{\text{I}}_{h}}{g}_{{\text{E}}_{k}}}(t).\end{array}$$

Combining this result with Eq. (1), yields the correlation, ${\text{\rho}}_{\text{in}}={\text{\rho}}_{{J}_{1}{J}_{2}},$ between the total input currents given in Eq. (13).

Using the solution of the linearized equations described in the Section “Materials and Methods”, we obtain a linear approximation to the cross-covariance functions,

$${C}_{{V}_{h}{V}_{k}}(t)\approx \frac{1}{2{\tau}_{\text{eff}}}({C}_{{J}_{h}{J}_{k}}\ast (K\star K))(t)$$

(16)

for *h*,*k*=1,2 where $(K\star K)(t)=\frac{{\tau}_{\text{eff}}}{2}{e}^{-\left|t\right|/{\tau}_{\text{eff}}}$ is the cross-covariance between the linear kernel, *K*, and itself. The convolution with (*KK*)(*t*) scales the area of both the auto- and cross-covariance functions by a factor of ${\tau}_{\text{eff}}^{2},$ and therefore leaves the ratio of the areas, ${\text{\rho}}_{{V}_{1}{V}_{2}}$ unchanged. Thus, the linear approximation predicts that ${\text{\rho}}_{{V}_{1}{V}_{2}}\approx {\text{\rho}}_{\text{in}}.$

When the total inputs are strong, τ_{eff} is small and we can simplify Eq. (16) by approximating (*KK*)(*t*) with a delta function with mass ${\tau}_{\text{eff}}^{2}$ so that ${C}_{{V}_{1}{V}_{2}}(t)\approx {\tau}_{\text{eff}}{C}_{{J}_{1}{J}_{2}}(t)/2$ and similarly for ${C}_{{V}_{1}{V}_{1}}(t).$ This approximation is valid when the synaptic time constants are significantly larger than τ_{eff}, which is likely to hold in high conductance states. We compare this approximation to cross-covariance functions obtained from simulations in Figure Figure88.

In all examples considered, the cross-covariance functions have exponentially decaying tails. We define the correlation time constant, ${\tau}_{xy}=\underset{t\to \infty}{\mathrm{lim}}-t/\mathrm{ln}({C}_{xy}(t)),$ as a measure of the decay rate of the exponential tail. If *t*τ* _{xy}*, then

The time constant of a convolution between two exponentially decaying functions is just the maximum time constant of the two functions. Thus, from the results above, the correlation time constant between the membrane potentials is the maximum of the correlation time constants between the inputs, the synaptic time constants, and the effective membrane time constant ${\tau}_{{V}_{1}{V}_{2}}=\mathrm{max}\{{\tau}_{{\text{E}}_{1}{\text{E}}_{2}},{\tau}_{{\text{I}}_{1}{\text{I}}_{2}},{\tau}_{{\text{E}}_{1}{\text{I}}_{2}},{\tau}_{{\text{E}}_{2}{\text{I}}_{1}},{\tau}_{\text{e}},{\tau}_{\text{i}},{\tau}_{\text{eff}}\}$ where ${\tau}_{{\text{E}}_{1}{\text{E}}_{2}},$${\tau}_{{\text{I}}_{1}{\text{I}}_{2}},$${\tau}_{{\text{E}}_{1}{\text{I}}_{2}},$ and ${\tau}_{{\text{E}}_{2}{\text{I}}_{1}}$ are the time constants of the input spike trains and τ_{e} and τ_{i} are synaptic time constants. Thus the cross-covariances functions between the membrane potentials are generally broader than the cross-covariance functions between the spike train inputs.

Consider a feedforward network where each layer consists of *N*_{e} excitatory cells and *N*_{i} inhibitory cells; each cell in layer *k* receives *n*_{e} excitatory and *n*_{i} inhibitory inputs from layer (*k*−1), and these connections are chosen randomly and independently across neurons in layer *k*. Then the degree of overlap in the excitatory and inhibitory inputs to a pair of cells in layer *k* is a random variable. Following the derivation in Derivation of Eq. (10) in Appendix,

$${\rho}_{{\text{E}}_{1}{\text{E}}_{2}}=\frac{{\gamma}_{{\text{E}}_{1}{\text{E}}_{2}}}{{\sigma}_{\text{E}}^{2}}=\frac{{s}_{\text{e}}{\sigma}_{\text{e}}^{2}+({n}_{\text{e}}^{2}-{s}_{\text{e}}){\rho}_{\text{ee}}{\sigma}_{\text{e}}^{2}}{{n}_{\text{e}}{\sigma}_{\text{e}}^{2}+({n}_{\text{e}}^{2}-{n}_{\text{e}}){\rho}_{\text{ee}}{\sigma}_{\text{e}}^{2}},$$

where *s*_{e} denotes the number of common excitatory inputs between the two cells. To understand the origin of *s*_{e}, suppose the *n*_{e} excitatory inputs to cell 1 have been selected. Then the selection of the *n*_{e} excitatory inputs to cell 2 involves choosing, without replacement, from two pools: the first, of size *n*_{e}, projects to cell 1, and the second, of size (*N*_{e}−*n*_{e}), does not. Therefore, *s*_{e} is follows a hyper-geometric distribution with parameters (*N*_{e}, *n*_{e}, *n*_{e}), and has mean ${n}_{\text{e}}^{2}/{N}_{\text{e}}={n}_{\text{e}}{p}_{\text{e}}.$ In addition, this random variable is independently selected amongst each pair in layer *k*. Using the mean value of *s*_{e}, we obtain Eq. (10).

For simplicity, we assume that |*V*_{E}−*V _{L}*|=|

$${\rho}_{\text{in}}=\frac{|{V}_{\text{E}}-{V}_{L}{|}^{2}{\gamma}_{{\text{E}}_{1}{\text{E}}_{2}}+|{V}_{\text{I}}-{V}_{L}{|}^{2}{\gamma}_{{\text{I}}_{1}{\text{I}}_{2}}-2|{V}_{\text{E}}-{V}_{L}||{V}_{\text{I}}-{V}_{L}|{\gamma}_{{E}_{1}{I}_{2}}}{|{V}_{\text{E}}-{V}_{L}{|}^{2}{\sigma}_{\text{E}}^{2}+|{V}_{\text{I}}-{V}_{L}{|}^{2}{\sigma}_{\text{I}}^{2}-2|{V}_{\text{E}}-{V}_{L}||{V}_{\text{I}}-{V}_{L}|{\gamma}_{{\text{E}}_{1}{\text{I}}_{1}}}.$$

Substituting the values of the covariances and variances, and dividing the numerator and denominator by (|*V*_{E}−*V*_{L}|σ_{e})^{2}, we get

$${\rho}_{\text{in}}=\text{}\frac{{s}_{\text{e}}+({n}_{\text{e}}^{2}-{s}_{\text{e}})\rho +{s}_{\text{i}}+({n}_{\text{i}}^{2}-{s}_{\text{i}})\rho -2{n}_{\text{e}}{n}_{\text{i}}\rho}{{n}_{\text{e}}+({n}_{\text{e}}^{2}-{n}_{\text{e}})\rho +{n}_{\text{i}}+({n}_{\text{i}}^{2}-{n}_{\text{i}})\rho -2{n}_{\text{e}}{n}_{\text{i}}\rho \text{}}.$$

Rearranging terms and dividing numerator and denominator by ${n}_{\text{i}}^{2},$ along with the substitution β=*n*_{e}/*n*_{i}, we have

$${\rho}_{\text{in}}=\frac{\rho {(1-\beta )}^{2}+\frac{1}{{n}_{\text{i}}^{2}}\left({s}_{\text{e}}+{s}_{\text{i}}\right)(1-\rho )}{\rho {(1-\beta )}^{2}+\frac{1}{{n}_{\text{i}}^{2}}({n}_{\text{e}}+{n}_{\text{i}})(1-\rho )}$$

(17)

This equation takes into account the variations in overlap due to finite size effects since *s*_{e} and *s*_{i} are random variables. Eq. (14) in the text represents the expected value P(ρ)=ρ_{in} which can be obtained by replacing the variables *s*_{e} and *s*_{i} in Eq. (17) with their respective means, *s*_{e}=*n*_{e}*p*_{e} and *s*_{i}=*n*_{i}*p*_{i}. The expectation above is taken over realizations of the random connectivity of the feedforward network.

To calculate the standard deviation for the inset in Figure Figure6D,6D, we ran Monte Carlo simulations, drawing *s*_{e} and *s*_{i} from a hypergeometric distribution and calculating the resulting transfer, $S({\text{\rho}}_{\text{in}})={\text{\rho}}_{\text{in}}^{\text{2}}$ using Eq. (17). Note, however, that Eq. (17) and the inset in Figure Figure6D,6D, do not account for all of the effects of randomness which may destabilize the balanced network. In deriving Eq. (17), we assumed that the statistics in the second layer were uniform. However, variations in the degree of overlap in one layer will cause inhomogeneities in the variances and rates at the next layer. In a feedforward setting, these inhomogeneities are compounded at each layer to destabilize the asynchronous fixed point.

- Barreiro A., Shea-Brown E., Thilo E. (2009). Timescales of spike-train correlation for neural oscillators with common drive. Phys. Rev. E 81, Arxiv preprint arXiv:0907.3924. [PubMed]
- Bedenbaugh P., Gerstein G. (1997). Multiunit normalized cross correlation differs from the average single-unit normalized correlation. Neural. Comput. 9, 1265–127510.1162/neco.1997.9.6.1265 [PubMed] [Cross Ref]
- Câteau H., Reyes A. (2006). Relation between single neuron and population spiking statistics and effects on network activity. Phys. Rev. Lett. 96, 58101.10.1103/PhysRevLett.96.058101 [PubMed] [Cross Ref]
- Chen Y., Geisler W. S., Seidemann E. (2006). Optimal decoding of correlated neural population responses in the primate visual cortex. Nat. Neurosci. 9, 1412–142010.1038/nn1792 [PMC free article] [PubMed] [Cross Ref]
- Cohen M., Maunsell J. (2009). Attention improves performance primarily by reducing interneuronal correlations. Nat. Neurosci. 12, 1594–160010.1038/nn.2439 [PMC free article] [PubMed] [Cross Ref]
- Coombes S., Timofeeva Y., Svensson C., Lord G., Josić K., Cox S., Colbert C. (2007). Branching dendrites with resonant membrane: A Òsum-over-tripsÓ approach. Biol. Cybern. 97, 137–14910.1007/s00422-007-0161-5 [PubMed] [Cross Ref]
- de la Rocha J., Doiron B., Shea-Brown E., Josić K., Reyes A. (2007). Correlation between neural spike trains increases with firing rate. Nature 448, 802–80610.1038/nature06028 [PubMed] [Cross Ref]
- Diesmann M., Gewaltig M., Aertsen A. (1999). Stable propagation of synchronous spiking in cortical neural networks. Nature 402, 529–53310.1038/990101 [PubMed] [Cross Ref]
- Doiron B., Rinzel J., Reyes A. (2006). Stochastic synchronization in finite size spiking networks. Phys. Rev. E 74, 30903.10.1103/PhysRevE.74.030903 [PubMed] [Cross Ref]
- Ecker A., Berens P., Keliris G., Bethge M., Logothetis N., Tolias A. (2010). Decorrelated neuronal firing in cortical microcircuits. Science 327, 584–58710.1126/science.1179867 [PubMed] [Cross Ref]
- Gasparini S., Magee J. (2006). State-dependent dendritic computation in hippocampal CA1 pyramidal neurons. J. Neurosci. 26, 2088.10.1523/JNEUROSCI.4428-05.2006 [PubMed] [Cross Ref]
- Gawne T., Richmond B. (1993). How independent are the messages carried by adjacent inferior temporal cortical neurons? J. Neurosci. 13, 2758. [PubMed]
- Goard M., Dan Y. (2009). Basal forebrain activation enhances cortical coding of natural scenes. Nat. Neurosci. 12, 1444–144910.1038/nn.2402 [PMC free article] [PubMed] [Cross Ref]
- Gutnisky D., Dragoi V. (2008). Adaptive coding of visual information in neural populations. Nature 452, 220–22410.1038/nature06563 [PubMed] [Cross Ref]
- Gutnisky D. A., Josić K. (2010). Generation of spatio-temporally correlated spike-trains and local-field potentials using a multivariate autoregressive process. J. Neurophys. 10.1152/jn.00518.2009 [PubMed] [Cross Ref]
- Hertz J. (2010). Cross-correlations in high-conductance states of a model cortical network. Neural. Comput. 22, 427–44710.1162/neco.2009.06-08-806 [PubMed] [Cross Ref]
- Johnston D., Narayanan R. (2008).Active dendrites: colorful wings of the mysterious butterflies. Trends Neurosci. 31, 309–31610.1016/j.tins.2008.03.004 [PubMed] [Cross Ref]
- Kuhn A., Aertsen A., Rotter S. (2003). Higher-order statistics of input ensembles and the response of simple model neurons. Neural Comput. 15, 67–10110.1162/089976603321043702 [PubMed] [Cross Ref]
- Kuhn A., Aertsen A., Rotter S. (2004). Neuronal integration of synaptic input in the fluctuation-driven regime. J. Neurosci. 24, 2345.10.1523/JNEUROSCI.3349-03.2004 [PubMed] [Cross Ref]
- Kumar A., Rotter S., Aertsen A. (2008). Conditions for propagating synchronous spiking and asynchronous firing rates in a cortical network model. J. Neurosci. 28, 5268.10.1523/JNEUROSCI.2542-07.2008 [PubMed] [Cross Ref]
- Li X., Ascoli G. A. (2006). Computational simulation of the input–output relationship in hippocampal pyramidal cells. J. Comput. Neurosci. 21, 191–20910.1007/s10827-006-8797-z [PubMed] [Cross Ref]
- Litvak V., Sompolinsky H., Segev I., Abeles M. (2003). On the transmission of rate code in long feedforward networks with excitatory-inhibitory balance. J. Neurosci. 23, 3006. [PubMed]
- Maršálek P., Koch C., Maunsell J. (1997). On the relationship between synaptic input and spike output jitter in individual neurons. Proc. Natl. Acad. Sci. U.S.A 94, 735.10.1073/pnas.94.2.735 [PubMed] [Cross Ref]
- Maynard E. M., Hatsopoulos N. G., Ojakangas C. L., Acuna B. D., Sanes J. N., Normann R. A., Donoghue J. P. (1999). Neuronal interactions improve cortical population coding of movement direction. J. Neurosci. 19, 8083–8093 [PubMed]
- Mitchell J. F., Sundberg K. A., Reynolds J. H. (2009). Spatial attention decorrelates intrinsic activity fluctuations in macaque area v4. Neuron 63, 879–88810.1016/j.neuron.2009.09.013 [PMC free article] [PubMed] [Cross Ref]
- Morel D., Levy W. (2009). The cost of linearization. J. Comput. Neurosci. 27, 259–27510.1007/s10827-009-0141-y [PubMed] [Cross Ref]
- Moreno-Bote R., Parga N. (2006). Auto-and crosscorrelograms for the spike response of leaky integrate-and-fire neurons with slow synapses. Phys. Rev. Lett. 96, 28101.10.1103/PhysRevLett.96.028101 [PubMed] [Cross Ref]
- Moreno-Bote R., Renart A., Parga N. (2008). Theory of input spike auto- and cross-correlations and their effect on the response of spiking neurons. Neural Comput. 20, 1651–170510.1162/neco.2008.03-07-497 [PubMed] [Cross Ref]
- Nunez P., Srinivasan R. (2006). Electric Fields of the Brain: The Neurophysics of EEG. New York, NY: Oxford University Press; 10.1093/acprof:oso/9780195050387.001.0001 [Cross Ref]
- Okun M., Lampl I. (2008). Instantaneous correlation of excitation and inhibition during ongoing and sensory-evoked activities. Nat. Neurosci. 11, 535–53710.1038/nn.2105 [PubMed] [Cross Ref]
- Oram M. W., Földiák P., Perrett D. I., Sengpiel F. (1998). The ’ideal homunculus’: decoding neural population signals. Trends Neurosci. 21, 259–26510.1016/S0166-2236(97)01216-2 [PubMed] [Cross Ref]
- Ostojić S., Brunel N., Hakim V. (2009). How connectivity, background activity, and synaptic properties shape the cross-correlation between spike trains. J. Neurosci. 29, 10234–1025310.1523/JNEUROSCI.1275-09.2009 [PubMed] [Cross Ref]
- Poort J., Roelfsema P. (2009). Noise correlations have little influence on the coding of selective attention in area v1. Cereb. Cortex 19, 543.10.1093/cercor/bhn103 [PMC free article] [PubMed] [Cross Ref]
- Renart A., de la Rocha J., Bartho P., Hollender L., Parga N., Reyes A., Harris K. (2010). The asynchronous state in cortical circuits. Science 327, 587–59010.1126/science.1179850 [PMC free article] [PubMed] [Cross Ref]
- Reyes A. (2003). Synchrony-dependent propagation of firing rate in iteratively constructed networks in vitro. Nat. Neurosci. 6, 593–59910.1038/nn1056 [PubMed] [Cross Ref]
- Roelfsema P., Lamme V., Spekreijse H. (2004). Synchrony and covariation of firing rates in the primary visual cortex during contour grouping. Nat. Neurosci. 7, 982–99110.1038/nn1304 [PubMed] [Cross Ref]
- Romo R., Hernández A., Zainos A., Salinas E. (2003). Correlated neuronal discharges that increase coding efficiency during perceptual discrimination. Neuron 38, 649–65710.1016/S0896-6273(03)00287-3 [PubMed] [Cross Ref]
- Salinas E., Sejnowski T. J. (2000). Impact of correlated synaptic input on output firing rate and variability in simple neuronal models. J. Neurosci. 20, 6193–209 [PubMed]
- Schneider A., Lewis T., Rinzel J. (2006). Effects of correlated input and electrical coupling on synchrony in fast-spiking cell networks. Neurocomputing 69, 1125–112910.1016/j.neucom.2005.12.058 [Cross Ref]
- Shadlen M. N., Newsome W. T. (1998). The variable discharge of cortical neurons: implications for connectivity, computation, and information coding. J. Neurosci. 18, 3870–96 [PubMed]
- Shea-Brown E., Josić K., de La ocha J., Doiron B. (2008). Correlation and synchrony transfer in integrate-and-fire neurons: basic properties and consequences for coding. Phys. Rev. Lett. 100, 108102.10.1103/PhysRevLett.100.108102 [PubMed] [Cross Ref]
- Smith M. A., Kohn A. (2008). Spatial and temporal scales of neuronal correlation in primary visual cortex. J. Neurosci. 28, 12591–1260310.1523/JNEUROSCI.2929-08.2008 [PMC free article] [PubMed] [Cross Ref]
- Stark E., Globerson A., Asher I., Abeles M. (2008). Correlations between groups of premotor neurons carry information about prehension. J. Neurosci. 28, 10618–1063010.1523/JNEUROSCI.3418-08.2008 [PubMed] [Cross Ref]
- Steinmetz P., Roy A., Fitzgerald P., Hsiao S., Johnson K., Niebur E. (2000). Attention modulates synchronized neuronal firing in primate somatosensory cortex. Nature 404, 187–19010.1038/35004588 [PubMed] [Cross Ref]
- Stratonovich R. (1963). Topics in the Theory of Random Noise: General Theory of Random Processes. Nonlinear Transformations of Signals and Noise. New York, NY: Gordon and Breach
- Super H., Roelfsema P. (2005). Chronic Multiunit Recordings in Behaving Animals: Advantages and Limitations. Development, dynamics and pathology of neuronal networks: from molecules to functional circuits: proceedings of the 23rd International Summer School of Brain Research, held at the Royal Netherlands Academy of Arts and Sciences, Amsterdam, from 25–29 August 2003, 263
- Tetzlaff T., Buschermöhle M., Geisel T., Diesmann M. (2003). The spread of rate and correlation in stationary cortical networks. Neurocomputing 52, 949–954
- Tetzlaff T., Rotter S., Stark E., Abeles M., Aertsen A., Diesmann M. (2008). Dependence of neuronal correlations on filter characteristics and marginal spike train statistics. Neural Comput. 20, 2133–218410.1162/neco.2008.05-07-525 [PubMed] [Cross Ref]
- Tiesinga P. H., Fellous J.-M., Salinas E., José J. V., Sejnowski T. J. (2004). Inhibitory synchrony as a mechanism for attentional gain modulation. J. Physiol. (Paris) 98, 296–31410.1016/j.jphysparis.2005.09.002 [PMC free article] [PubMed] [Cross Ref]
- Troyer T., Miller K. (1997). Physiological gain leads to high ISI variability in a simple model of a cortical regular spiking cell. Neural. Comput. 9, 971–98310.1162/neco.1997.9.5.971 [PubMed] [Cross Ref]
- Vaadia E., Haalman I., Abeles M., Bergman H., Prut Y., Slovin H., Aertsen A. (1995). Dynamics of neuronal interactions in monkey cortex in relation to behavioural events. Nature 373, 515–51810.1038/373515a0 [PubMed] [Cross Ref]
- van Rossum M., Turrigiano G., Nelson S. (2002). Fast propagation of firing rates through layered networks of noisy neurons. J. Neurosci. 22, 1956–1966 [PubMed]
- Vogels T. P., Abbott L. F. (2005). Signal propagation and logic gating in networks of integrate-and-fire neurons. J. Neurosci. 25, 10786–1079510.1523/JNEUROSCI.3508-05.2005 [PubMed] [Cross Ref]
- Womelsdorf T., Schoffelen J.-M., Oostenveld R., Singer W., Desimone R., Engel A. K., Fries P., Jun. (2007). Modulation of neuronal interactions through neuronal synchronization. Science 316, 1609–161210.1126/science.1139597 [PubMed] [Cross Ref]
- Zohary E., Shadlen M., Newsome W. (1994). Correlated neuronal discharge rate and its implications for psychophysical performance. Nature 370, 140–14310.1038/370140a0 [PubMed] [Cross Ref]

Articles from Frontiers in Computational Neuroscience are provided here courtesy of **Frontiers Media SA**