Mathematical models may explain the properties of EFP.

(A) In computer simulations, both MUA and EFP showed shorter average response latency and more precise timing than single neurons. The average response latency of single neurons (20 ms, red dash-dot line) itself does not affect the temporal precision of either the MUA or EFP and was chosen arbitrarily for illustrative purposes. (B) In computer simulations, MUAs and EFPs became larger in amplitude and more precise in timing with an increasing number of neurons firing synchronously. With an increasing number of neurons, EFPs showed better temporal precision than MUA for the high-synchrony scenario (left column). Red dashed line shows the predicted effect of the number of firing neurons on the EFP through least squares fitting of the simulation data based on the mathematical models presented in figure panel C (see S1 Text for details). The gof represents the adjusted r-square of least squares fitting. (C) Equation 1 shows that the amplitude of the EFP (L(t)) increases linearly with the number of firing neurons (n). Equation 2 shows that the SD of EFP latency () is proportional to the SD of single neurons (σ) and decreases with increasing number of neurons (n), following a square root law. (D) Biophysical simulations with 100 neurons firing at varying degrees of synchrony, based on a cell model from Neocortical Microcircuitry [44]. The first peaks of both MUA and EFP had better temporal precision and shorter response latency than the single neurons. Importantly, the EFP was better in temporal precision than the MUA when neurons fired in high synchrony. EFP, extracellular field potential; gof, goodness of fit; MUA, multiunit activity.