The acoustic change complex as a diagnostic tool for cochlear dead regions evaluated in normally hearing adults.

The loss of inner hair cells and/or neurons in the cochlea leads to cochlear dead regions (DRs). One of the consequences of DRs is noisy transmission of information from the cochlea to the brain, which results in poorer than expected perception of speech in noise. DRs can be detected using behavioural masking techniques such as the threshold-equalizing noise (TEN) test. However, the TEN test, although fast and easy to perform, requires participants' active cooperation and thus is not suitable for use with infants, young children and adults who cannot provide behavioural responses. Recently, Kang et al. (2018) proposed the electrophysiological Acoustic Change Complex (ACC) as an objective test for diagnosing cochlear DRs, but with limited evidence supporting their proposed approach. Thus, the first step towards addressing this gap is to develop and assess an objective ACC detection method. Here, we evaluated three methods of detecting the ACC based on: i) the root mean square of N1-P2 amplitude as reported in Kang et al. (2018); ii) Signal-to-Noise Ratio (SNR) as recommended by the British Society of Audiology; and iii) bootstrap by multiple resampling of the original data in a random manner and calculating a measure of variance in the N1-P2 amplitude to noise floor ratio from the resampled data. We have also examined the relationship between ACC amplitude, threshold, and stimulus frequency/intensity. Twenty-three normally hearing adults were tested. The ACC was evoked by either a 1-kHz or 4-kHz pure tone presented simultaneously with the TEN at SNRs between 0 and 15 dB, in 3 dB steps relative to the TEN level. We found that the bootstrap method was most efficient at detecting ACC and was two times faster to administer than the commonly used British Society of Audiology method. We also found that the threshold of ACC detection is frequency dependent, i.e., the higher the frequency, the higher the signal level needed to evoke ACC. Thus, frequency-dependent ACC norms need to be established before this method can be used for detection of cochlear DRs in children and adults who cannot provide behavioural responses.