An Evaluation of the Temporal Integrator Processing Strategy for Cochlear Implants in Comparison to the Clinical Strategy

Abstract

Objectives: Speech perception remains challenging for cochlear-implant recipients in conditions containing background noise. The sound processing strategy of the cochlear implant transforms the acoustic signal into a pattern of electrical stimulation pulses which are transmitted by the electrode array. Modifications to the sound processing strategy could improve the transmission of information by the cochlear implant, alleviating speech perception difficulties in noise. One such modification, the temporal integrator processing strategy (TIPS), uses a model of temporal masking to identify and remove stimulation pulses that are unlikely to be perceived. In a previous investigation, TIPS significantly improved the perception of Continuous Interleaved Sampling-processed speech in the presence of speech-shaped noise while reducing the power required for stimulation (Lamping et al., 2020). This study extends the previous investigation to include conditions that better reflect everyday listening for cochlear-implant recipients. Design: Two pre-registered double-blind within-participant experiments measured the speech-recognition performance of twelve cochlear implant listeners for sentences before and after TIPS processing. Experiments 1 and 2 investigated TIPS performance against the Continuous Interleaved Sampling or Advanced Combination Encoder baseline speech processing strategies, respectively. Speech reception thresholds were measured in a speech-shaped noise (in Experiment 1) and a multi-talker noise (in Experiments 1 and 2) and mean-opinion score speech quality ratings were obtained in quiet (in Experiments 1 and 2). In both experiments, all strategies employed the subjects’ clinical pulse rate, inter-phase gap, phase duration, and ground electrodes, and participants were acclimatised for 10 minutes before testing each strategy. Results: In Experiments 1 and 2, the median speech reception thresholds obtained were comparable across the baseline and TIPS strategies, with no statistical differences noted in either speech-shaped or multi-talker noise. Quality ratings were lower for TIPS when compared to the Continuous Interleaved Sampling strategy (𝑝 = 0.044) but not when compared to the Advanced Combination Encoder strategy, which all participants used as their everyday strategy. In all conditions, TIPS could reduce device power consumption by between 21% and 42%. Conclusions: TIPS could result in substantial power savings, without compromising speech quality or speech intelligibility in noise. Although some individuals demonstrated improved speech reception thresholds with TIPS, the significant group-level improvement previously found by Lamping et al. (2020) was not observed in this study.