For all three neuronal populations, 95% of the amplitude of the signal recorded in the soma layer came from neurons within a radius smaller than 200 μm ( Figure 3D). By plotting the LFP amplitude as a function of cortical depth, we further found the largest LFP amplitudes at the soma level ( Figure 3E). We therefore conclude that when the synaptic activity is

uncorrelated, the LFP is rather local, both in terms of horizontal reach and amplitude variation in the vertical direction. Changing the synaptic distributions to either only apical or only basal dendrites for the pyramidal cells gave a different depth dependence for both the reach and the amplitude of the LFP for the L5 population, whereas the results for the L3 population

were largely unaffected (Figures 3D and 3E). For the apically activated L5 population both the LFP amplitude and the spatial reach are similar for the electrode see more contacts positioned in L2/3 and the L5 soma layer (Figures 3D3–3E3). This demonstrates that these qualitative features of the LFP are determined both by the spatial distributions of the synaptic inputs and the neuronal morphology, in particular the depth profile of the total dendritic area (Lindén et al., 2010). We next compared the numerical simulations with predictions of the simplified model: by using the detailed single-cell decay functions f(r) obtained Cyclopamine concentration above ( Figure 2), we numerically integrated the simplified model ( Equation 1).

As seen in Figure 3, the predictions of the simplified model agree excellently with the results of the comprehensive numerical simulations, suggesting that our simplified model indeed captures the salient features of LFP generation from neuronal populations. How do these results change when the synaptic inputs to different cells in the population are correlated? We used the same simulation setup as above with the difference that spike trains to different cells were drawn from a finite pool of presynaptic spike trains (Figure 4A). This induced a mean correlation cξcξ between the synaptic input currents to different cells due to common input. By varying Terminal deoxynucleotidyl transferase the size of the pool of presynaptic spike trains n  pool we could vary the input correlation cξcξ (see Experimental Procedures). As predicted by the simplified model (Equation 1), inducing correlations between single cell LFP contributions changed the total LFP amplitude in three respects: (1) the LFP amplitude σ becomes considerably higher (Figures 4C1–4C3 and 4F1), (2) the reach R∗ of the LFP (as before defined as the population radius where the amplitude had reached 95 % of the value for R = 1,000 μm) generally increases ( Figures 4D1–4D3 and 4E1), and (3) the LFP amplitude σ no longer appears to converge to a fixed value with increasing population radius.