Perception of spectrally shifted speech: Implications for cochlear implants

Objective: Present study used the normal hearing subjects to check the effect of spectral shifts in monoaural and binaural cochlear implant simulations. Cochlear implant signal processing was simulated using eight channel sine wave vocoders. Materials and methods: The spectrally shifted sentences (HINT) were created to simulate 20 mm & 25 mm insertion depth of cochlear implants. Speech recognition scores were assessed in three mono aural conditions (No shift, 25 mm & 20 mm insertion) and two binaural conditions [asymmetrical spectral shift (20 & 25 mm insertions in each ear) & no spectral shift in both ears]. Greenwood (1990) function was used to decide the corner frequencies for the simulation of spectral shifts. Results: It reveled that partial shift (20 mm) resulted in poor scores in monaural than 25 mm shift. Spectrally unshifted speech resulted in significantly higher scores when compared to spectrally shifted speech. In binaural asymmetrical condition, scores were equivalent to 25 mm shift in monaural condition. Even in symmetrical condition also binaural scores were equivalent to mono aural scores which implies that at least in initial stages there may not be binaural integration or interference may be present. Conclusion: In binaural cochlear implantation there may not be any change in the speech perception in the initial stages even if the insertion depth varies both ears, at least in quiet condition.