Influence of coding strategies in electric-acoustic hearing: A simulation dedicated to EAS cochlear implant, in the presence of noise

• Published in: Applied acoustics 2014-02, Vol.76, p.300-309
• Main Authors: Seldran, Fabien, Gallego, Stéphane, Thai-Van, Hung, Berger-Vachon, Christian
• Format: Article
• Language: English
• Subjects: Biological and medical sciences; Cochlear implant; Electric-acoustic stimulation; Head and neck surgery. Maxillofacial surgery. Dental surgery. Orthodontics; Medical sciences; N-of-M; Speech-coding strategies; Surgery (general aspects). Transplantations, organ and tissue grafts. Graft diseases; Surgery of the ear, the auditive nerve and the facial nerve; Vocoder simulation
• ISSN: 0003-682X; 1872-910X
• DOI: 10.1016/j.apacoust.2013.08.003

•N-of-M coding, stimulation rate and pre-emphasis affect speech recognition with CI.•Pre-emphasis has an influence on performance in electric-acoustic stimulation.•In EAS, reducing the information by a factor 20 can be compensated by pre-emphasis. This work deals with electric-acoustic stimulation (EAS), which keeps the low frequency acoustic information and electrically codes the high frequencies of the signal. One of the goals of the coding strategies is to limit the phenomenon of channel interaction, which can occur in CIs. The “N-of-M” strategy, where only a subset of electrode channels is stimulated, may be of advantage. Generally, this processing is associated with a pre-emphasis filter. Two important parameters for the N-of-M strategy are the number of active channels (N) and the updating rate; the latter corresponds to the stimulation rate. M is the number of electrical channels. The goal of this study was to investigate the influence of these parameters on speech intelligibility in EAS. The signal was presented, in simulation, to normal-hearing (NH) subjects in acoustic (A), electric (E) and electric-acoustic conditions. Recognition performance was measured in quiet and in the presence of background noise (cafeteria noise). Signal-to-noise ratios (SNRs) ranged from 0 to +12dB. Fifteen listeners participated in the experiment. The N values ranged from 2 to 10 (out of 10); M was 10. The frame updating rate was 250updatespersecond(ups) and 1000ups. Results showed that increasing N from 2 to 10 improved speech intelligibility, especially in the presence of the background noise, under E and EAS conditions. In noisy situations, 2/10 coupled with a high-pass pre-emphasis filter led to results similar to the 10/10 condition. Changing the frame rate from 250ups to 1000ups did not modify the performance. Future investigations on patients using EAS are now needed to validate the performance seen with NH listeners. Above all, in the strategy 2 out of 10, the number of pulses per second can be divided by 20, and when the pre-emphasis is used only a slight decrease in performance is expected; this is of interest when interaction between the electrodes corrupts the performance.