This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, reproduction and adaptation in any medium and for any purpose provided that it is properly attributed. For attribution, the original author(s), title, publication source (PeerJ Preprints) and either DOI or URL of the article must be cited.
Cite this article
Bidelman GM, McElwain C.2017. Objective detection of auditory steady-state responses based on mutual information: Receiver operating characteristics and validation across modulation rates and levels. PeerJ Preprints5:e3399v1https://doi.org/10.7287/peerj.preprints.3399v1
Auditory steady-state responses (ASSRs) are sustained potentials used to assess the physiological integrity of the auditory pathway and objectively estimate hearing thresholds. ASSRs are typically analyzed using statistical procedures in order to remove the subjective bias of human operators. Knowing when to terminate signal averaging in ASSR testing is also critical for making efficient clinical decisions and obtaining high-quality data in empirical research. Here, we investigated a new detection metric for ASSRs based on mutual information (MI) [Bidelman, G. M. (2014). Objective information-theoretic algorithm for detecting brainstem evoked responses to complex stimuli. J. Am. Acad. Audiol., 25(8), 711-722], previously bench tested using only a single suprathreshold stimulus. ASSRs were measured in n=10 normal hearing listeners to various stimuli varying in modulation rate (40, 80 Hz) and level (80 – 20 dB SPL). MI-based classifiers applied to ASSRs recordings showed that accuracy of ASSR detection ranged from ~75 - 99% and was better for 40 compared to 80 Hz responses and for higher compared to lower stimulus levels. Detailed receiver operating characteristics (ROC) were used to establish normative ranges for MI for reliable ASSR detection across levels and rates (MI=0.9-1.6). Relative to current statistics for ASSR identification (F-test), MI was found to be a more efficient metric for determining the stopping criterion for signal averaging. Our new results confirm that MI can be applied across a broad range of ASSR stimuli and might offer improvements to conventional objective techniques for ASSR detection.