Quality Estimation of Click-evoked Oto-acoustic Emissions

Abstract

The F_sp statistic used by Elberling & Don (1984) as a quality estimator for averaged ABR recordings has been applied to averaged click-evoked oto-acoustic emissions. This statistic is akin to signal-to-noise ratio. The sampling distribution of F_sp under the null hypothesis that no signal is present has been determined by using no-stimulus trials. A determination by multiple replications in three normally-hearing adults demonstrated significant inter-subject variation, leading to a larger experiment in 58 children. From the latter experiment, the sampling distribution across subjects allowed the setting of a criterion value for F_sp of 2.0 which was only exceeded on 1% of trials. Thus, values above this criterion would indicate the presence of a signal in the average with a probability of Type 1 error below 1%. When two consecutive averages both exceeded an F_sp of 1.6 the probability of Type 1 error was also below 1%.