Statistically optimal filtering of averaged ABRS

The transient Auditory evoked Brainstem Response (ABR) requires the averaging of 2000 or more trials. A set of ABRs obtained at various stimulus intensity levels, for hearing threshold assessments can require 30-45 minutes. In an effort to reduce this acquisition time, the number of responses in the averaging is reduced resulting in noisy ABRs. Here, the authors use a damped sinusoid model of the ABRs and present a method of reducing the noise in the averaged ABRs. The new method is a two step procedure that first estimates some of the ABR model parameters which are used to form a statistically optimal line-enhancer. An ABR example is presented to show the effectiveness of this approach.