Event-related fMRI study of recognition of simulated CI speech

Abstract

Objective To identify potential differences in brain activation when normal-hearing subjects listen to natural speech compared to speech signals with distortions characteristic of a cochlear implant (CI speech). Study Design Children and adults ( n=6) listened to different categories of speech stimuli or rested across trials during silent intervals of the functional magnetic resonance imaging (fMRI) scans. Stimuli including natural speech and speech signals processed to simulate a CI with eight spectral channels were presented. FMRI was performed on a 3.0-T MRI scanner using an acquisition method referred to as Hemodynamics Unrelated to Scanner Hardware (HUSH)-fMRI here. Auditory stimulation during silent intervals of the fMRI scans alternated with MR image acquisition on each trial. Trials from different experimental conditions were grouped, averaged, and analysed. Results Natural speech signals evoked robust cortical activation bilaterally relative to rest in all subjects. This brain activation pattern overlapped substantially with that evoked by CI speech, such that CI speech stimuli that were intelligible evoked brain activation patterns that were similar to natural speech. CI-speech intelligibility during fMRI acquisition was comparable to that observed outside of the scanner, suggesting minimal interference due to acoustic noise from the scanner. Conclusions HUSH fMRI on a 3-T MRI scanner is a time-efficient and sensitive method to detect cortical activation due to auditory processing of simulated CI speech.