If the synaptic input to the neurons in the vicinity of a recording electrode always had been uncorrelated, we could have reported the following simple rule of thumb: almost all of the LFP signal measured by an electrode comes from neurons within a lateral distance of about 200 μm. This estimate is in accordance with recent results by Katzner et al. (2009) and Xing et al. (2009). The independence of the spatial reach, i.e., the size of the region generating the LFP, from the morphology of the neurons in the population and the spatial distribution Trichostatin A of the synapses

may be at odds with common thinking on the origin of the LFP emphasizing the distinction between open-field (pyramidal) and closed-field (stellate) neurons ( Lorente de No, 1947 and Johnston and Wu, 1995), and this highlights the importance of a thorough quantitative investigation of the origin of LFP. The situation when the synaptic input to the neuronal population is correlated is, however, more in line with common thinking regarding the dominant contributions from pyramidal neurons, but only when the input is spatially asymmetric, i.e., solely onto either the basal or apical dendritic branches. In this case correlated http://www.selleckchem.com/Wnt.html synaptic inputs

were found to give correlated neuronal LFP sources and consequently an amplified LFP signal. With homogeneous inputs onto pyramidal neurons, this correlation transfer is observed to be very weak, resulting in very little such correlation amplification. For the stellate layer 4 neurons until with very symmetric dendritic branching, the LFP contributions from individual neurons were found to be essentially uncorrelated, independent of the level of synaptic input correlations. With spike-train correlations present in the synaptic input, as in our laminar network example in Figure 6, one might thus expect pyramidal neurons to give larger LFP contributions than the stellate neurons. Given the observed strong dependence on input correlations and spatial distribution

of the synaptic inputs, our model study thus suggests several possible explanations for the significant variation for the reach of the LFP seen in various experimental studies (Kreiman et al., 2006, Liu and Newsome, 2006, Berens et al., 2008a, Katzner et al., 2009 and Xing et al., 2009). As the level of synaptic input correlations depends on the state of cortical network, it follows that the LFP reach in general will not be a static fixed quantity, even for a particular fixed electrode in a particular experiment. We also find the population LFP to depend strongly on the depth position of the recording electrode. With the electrode placed above or below the dendrites of the generating population, as, e.g., for recordings done in L4, L5 or L6 with an active L3 population depicted in Figure 3, the reach is much larger than for recordings done in the soma layer (L2/3). However, the LFP amplitudes recorded in these lower layers are tiny in comparison.