Event-Related Potentials of Single Sided Deaf Cochlear Implant Users - Using a Semantic Oddball Paradigm in Noise

Abstract

ObjectiveIn individuals with single sided deafness (SSD), which is characterised by a profound hearing loss in one ear and normal hearing in the contralateral ear, binaural input is no longer present. A cochlear implant (CI) is the only way to restore functional hearing in the profoundly deaf ear, with previous literature demonstrating improvements in speech in noise intelligibility with the provision of a CI. However, we currently have a limited understanding of the neural processes involved (e.g., how the brain integrates the electrical signal produced by the CI with the acoustic signal produced by the normal hearing ear) and how the modulation of these processes with CI contributes to improved speech in noise intelligibility. Using a semantic oddball paradigm presented in the presence of background noise, this study aims to investigate how the provision of CI impacts speech in noise perception of SSD CI users. MethodHigh density electroencephalography (EEG) from twelve SSD-CI participants was recorded whilst they completed a semantic acoustic oddball task. All participants completed the oddball task in three different free field conditions with the speech and noise coming from different speakers. The three tasks were 1) with the CI-On in background noise, 2) with the CI-Off in background noise and 3) with the CI-On without background noise (Control). We examined task-performance (RT, subjective listening effort, and accuracy) and measured N2N4 and P3b event-related brain potentials (ERPs) linked to the discrimination and evaluation of task relevant stimuli. Speech in noise and sound localisation abilities was also measured. ResultsReaction time was significantly different between all tasks with CI-On (M(SE) = 809(39.9) ms) having faster RTs than CI-Off (M(SE) = 845(39.9) ms) and Control (M(SE) = 785(39.9) ms) being the fastest condition. The Control condition exhibited a significantly shorter N2N4 and P3b area latency when compared to the other two conditions. However, despite these differences noticed in RTs and area latency, we observed similar results between all three conditions for N2N4 and P3b difference area. ConclusionThe inconsistency between the behavioural and neural results suggest that EEG may not be a reliable measure of cognitive effort. This rationale is further supported by the different explanations used in past studies to explain N2N4 and P3b effects. Future studies should look to alternative measures of auditory processing (e.g., pupillometry) to get a deeper understanding of the underlying auditory processes that facilitate speech in noise intelligibility.