Somatic PSPs and spikes are uncorrelated because PSPs are attenuated by dendritic tree (compare with Figure 1a).

<p>(<b>a</b>) Simulation of a bar moving across the dendritic tree with a weak presynaptic+post synaptic mechanism, where the preferred direction was set to be 0 degrees (see <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1000899#pcbi-1000899-g014" target="_blank">Figure 14</a>). Peak somatic depolarization is greater for Null (black) than Pref (gray). In the Pref direction, distal dendrites were depolarized enough to spike, and the dendritic spike propagated to the soma and initiated a somatic spike, despite the somatic PSP being below spike threshold. In the Null direction, distal dendritic PSPs were insufficient to elicit a dendritic spike, and the somatic PSP did not reach spike threshold, so no spikes were generated. (<b>b</b>) The same simulation, but with the preferred direction changed to 180 degrees, showing the effect of the asymmetrical dendritic tree on the PSP rise time, shape and spiking. The apparent difference in initial spike threshold between (a) and (b) reflects the underlying PSPs, i.e. in (a) the preferred direction the underlying PSP rose more quickly. The simulation was run on the same morphology as <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1000899#pcbi-1000899-g002" target="_blank">Figure 2c</a> (DS060825), with bar velocity 1000 um/s, excitatory conductance 75pS/synapse, inhibitory conductance 95 pS/synapse, presynaptic mechanism (difference in conductances) reduced to 25% of the mechanism from <a href="http://www.ploscompbiol.org/article/info:doi/10.1371/journal.pcbi.1000899#pcbi-1000899-g008" target="_blank">Figure 8</a>.</p>