Novel speech processing strategies in cochlear implants: real improvements competing with learning effects

Abstract

The aim of this thesis was to develop new speech coding strategies for cochlear implants (CIs) that improve the quality of the incoming sound. To implement a new speech coding strategy, often multiple components are adjusted without knowing the effect of these individual factors. Our research showed that the replacement of bandpass filters for more efficient filters, based on Fast Fourier transformation, has no effect on the performance of CI users. The FFT filters can therefore be implemented without problems.

In this PhD research, the Dynamic Current Focusing (DCF) strategy was developed that leads to a narrower intra-cochlear electric field, especially at low loudness. This causes an improved spectral resolution and possibly also a better speech understanding in the long term. The studies in this dissertation showed that psychophysical tests, which are frequently used to evaluate new strategies, show learning effects, complicating the interpretation of results.

Phantom stimulation is a strategy that results in a shift of the stimulated area. When applied to the most apical electrode contact it can be used to transmit lower tones than currently possible. This dissertation shows that the perceived pitch can be shifted 0.92 electrodes on average.

Cochlear ImplantDynamic Current FocusingPhantom StimulationSpeech Coding Strategies