Advances in cochlear implant speech processing [Abstract]

Abstract

Our early research emphasized there was a restriction on the amount of speech and other acoustic information that could be transmitted to the nervous system by electrical stimulation of the auditory nerve. It also showed the need to use multiple-channel stimulation, and minimize channel interaction. As a result our research over the last 30 years has been directed towards optimizing the information presented to the auditory nervous system. This has involved extracting the energy of the first and second formants (FO/F2-WSP II; FO/FI/F2-WSP III; Multipeak-MSP) as well as the outputs of high band pass fixed filters (Multipeak - MSP) and coding these outputs as cochlear place of stimulation. The voicing frequency was coded as rate of stimulation. Our most recent speech processing strategy (SPEAK) extracts a specified number of .maximal outputs from a series of band pass filters, rather than selecting the peaks of energy which was the case with the other strategies. The voltages from the maximal outputs are used to stimulate appropriate electrodes on a place coding basis. The stimuli are presented at a constant stimulus rate to reduce channel interaction. Voicing is conveyed as amplitude variations.