A method to improve the latency estimation of the frequency-following potential (FFP)

If the latency of a noisy frequency-following potential (FFP) is estimated by determining the shift of the (periodical) cross-correlation function (CCF) between the stimulus and the FFP, the result may be unambiguous only within +/-1 or +/-2 periods of the CCF, because the absolute maximum and adjacent local maxima may not be significantly different. Here we present a method to amplify this difference by applying amplitude modulated stimuli. Using this method we first illustrate the effect of the method by a simulation and then demonstrate its usefulness by measuring real FFPs and estimating their latencies.