Modified variance ratio for objective detection of transient evoked potentials in bottlenose dolphins (Tursiops truncatus).

Evoked potential studies have often relied on one or more human observers to visually assess the averaged waveforms and decide if a response is present. Although simple and easy to implement, response detection strategies based on human observers are inherently subjective and depend on the observers' experience and biases. To avoid these shortcomings, some recent marine animal studies utilizing auditory steady-state responses have applied frequency-domain, statistically based objective detection methods; however, statistically based objective methods have not yet been applied to marine animal tests involving transient evoked responses, which are normally analyzed in the time domain. The present study applied a modified version of the variance ratio F(SP) to determine the presence or absence of evoked responses in two bottlenose dolphins (Tursiops truncatus) stimulated with tone pips. The appropriate degrees of freedom for the statistical tests were empirically determined in four dolphins. The modified variance ratio was found to be a useful tool and to provide an objective statistical approach for the detection of transient evoked potentials.