Date of Award

May 2020

Degree Type


Degree Name

Doctor of Philosophy (PhD)


Communication Sciences and Disorders


Kathy R. Vander Werff


Auditory Evoked Potentials, Auditory Neuroscience, Cortical Auditory Processing, Informational Masking

Subject Categories

Social and Behavioral Sciences


Purpose: Understanding speech in a background of other people talking is one of the most difficult listening challenges for hearing-impaired individuals, and even for those with normal hearing. Speech-on-speech masking, is known to contribute to increased perceptual difficulty over non-speech background noise because of informational masking provided over and above the energetic masking effect. While informational masking research has identified factors of similarity and uncertainty between target and masker that contribute to reduced behavioral performance in speech background noise, critical gaps in knowledge including the underlying neural-perceptual processes remain. By systematically manipulating aspects of similarity and uncertainty in the same auditory paradigm, the current study proposed to examine the time course and objectively quantify these informational masking effects at both early and late stages of auditory processing using auditory evoked potentials (AEPs) in a two-factor repeated measures paradigm.

Method: Thirty participants were included in this cross sectional repeated measures design. Target-masker similarity between target and masker were manipulated by varying the linguistic/phonetic similarity (i.e. language) of the talkers in the noise maskers. Specifically, four levels representing hypothesized increasing levels of informational masking were implemented: (1) No masker (quiet), (2) Mandarin (linguistically and phonetically dissimilar), (3) Dutch (linguistically dissimilar, but phonetically similar), and (4) English (linguistically and phonetically similar). Stimulus uncertainty was manipulated by task complexity, specifically target-to-target interval (TTI) of an auditory paradigm. Participants had to discriminate between English word stimuli (/bæt/ and /pæt/) presented in an oddball paradigm in each masker condition at +3 dB SNR by pressing buttons to either the target or standard stimulus (pseudo-randomized between /bæt/ and /pæt/ for all participants). Responses were recorded simultaneously for P1-N1-P2 (standard waveform) and P3 (target waveform). This design allowed for simultaneous recording of multiple AEP peaks, including analysis of amplitude, area, and latency characteristics, as well as accuracy, reaction time, and d’ behavioral discrimination to button press responses. Finally, AEP measurers were compared to performance on a behavioral word recognition task (NU-6 25-word lists) in the proposed language maskers and at multiple signal-to-noise ratios (SNRs) to further explore if AEP components of amplitude/area and latency are correlated to behavioral outcomes across proposed maskers.

Results: Several trends in AEP and behavioral outcomes were consistent with the hypothesized hierarchy of increasing linguistic/phonetic similarity from Mandarin to Dutch to English, but not all differences were significant. The most supported findings for this factor were that all babble maskers significantly affected outcomes compared to quiet, and that the native language English masker had the largest effect on outcomes in the AEP paradigm, including N1 amplitude, P3 amplitude and area, as well as decreased reaction time, accuracy, and d’ behavioral discrimination to target word responses. AEP outcomes for the Mandarin and Dutch maskers, however, were not significantly different across all measured components. Outcomes for AEP latencies for both N1 and P3 also supported an effect of stimulus uncertainty, consistent with a hypothesized increase in processing time related to increased task complexity when target stimulus timing was randomized. In addition, this effect was stronger, as evidenced by larger effect sizes, at the P3 level of auditory processing compared to the N1. An unanticipated result was the absence of the expected additive effect between linguistic/phonetic similarity and stimulus uncertainty. Finally, trends in behavioral word recognition performance were generally consistent with those observed for AEP component measures such that no differences between Dutch and Mandarin maskers were found, but the English masker yielded the lowest percent correct scores. Furthermore, correlations between behavioral word recognition and AEP component measures yielded some moderate correlations, but no common AEP components accounted for a majority of variance for behavioral word recognition.

Conclusions: The results of this study add to our understanding of auditory perception in informational masking in four ways. First, observable effects of both similarity and uncertainty were evidenced at both early and late levels of auditory cortical processing. This supports the use of AEPs to better understand the informational masking deficit by providing a window into the auditory pathway. Second, stronger effects were found for P3 response, an active, top-down level of auditory processing providing some suggestion that while informational masking degradation happens at lower levels, higher level active auditory processing is more sensitive to informational masking deficits. Third, the lack of interaction of main effects leads us to a linear interpretation of the interaction of similarity and uncertainty with an equal effect across listening conditions. Fourth, even though there were few and only moderate correlations to behavioral word recognition, AEP and behavioral performance data followed the same trends as AEP measures across similarity. Through both auditory neural and behavioral testing, language maskers degraded AEPs and reduced word recognition, but particularly using a native-language masker. The behavioral and objective results from this study provide a foundation for further investigation of how the linguistic content of target and masker and task difficulty contribute to difficulty understanding speech in noise.


Open Access

Available for download on Sunday, August 15, 2021