Research project: Temporal information in cochlear implants - Dormant

Cochlear implant (CI) users have limited access to spectral information because they use a small number of electrode channels and also because of cross-electrode interaction. Consequently, temporal information in the incoming signal is disproportionately important for the perceptual abilities in CI users. Two areas relevant to temporal coding in CI devices that have received attention in the literature are pulsatile stimulation rate and modulation depth perception in cochlear implant users. Research evidence regarding the benefits or otherwise of increased pulsatile stimulation rates in CI users is contradictory and variable, although as a general rule benefits to stimulation rates higher than about 1000 pulses per second per channel (pps/ch) have not been shown except in the odd individual. However, there is some evidence (Fu 2002; Liu et al. 2008) that modulation depth perception is related to speech recognition abilities in CI users. A limitation of previous work is the lack of consideration of CI filter bank effects, poor control of CI processor characteristics, uncertainty over the link between stimulation rate, modulation perception and speech perception, and limited understanding of how temporal information transmission impacts on specific acoustic characteristics of speech in quiet or noise. The broad aim of the present project is (i) to understand the effect of cochlear implant processing, in particular the filter bank analysis methods used, on temporal information in CI devices; (ii) the link between (i) and temporal resolution in cochlear implant users; (iii) to determine the impact of the previous factors on speech perception in CI users, including perception of different types of acoustic speech cue in quiet and noise. A further aim is to determine what filter bank characteristics might be needed to optimise temporal information. A published paper (Verschuur, 2005) looked at stimulation rate and found weak effects of this parameter when filter bank parameters were not co-varied with stimulation rate. Further work (not yet published) has estimated temporal information transmission from a CI system objectively and determined measurements of temporal modulation detection and speech recognition as a function of stimulation rate in CI users. Current and future work seeks to account for the effect of filter bank type and FFT characteristics on temporal information output and temporal resolution in CI users and normal hearing listeners, and to determine the relationship between these processing parameters, temporal modulation perception and perception of different kinds of speech information.