人工耳蜗中的双耳线索表征：信号处理和耳间联动通道选择的影响。

文献信息

中文摘要

背景： 双侧人工耳蜗植入是重建重度至极重度听力损失患者双耳听觉的主要临床干预手段。尽管许多双耳听觉功能关键依赖于耳间线索，现有的声音编码策略并未完全保留这些线索。方法：基于模拟，本研究系统考察了人工耳蜗信号处理如何在关键处理阶段影响耳间时间差和耳间水平差的编码。同时研究了联动通道选择和最大值数量对双耳线索编码及噪声中言语可懂度的影响。结果：非线性压缩被确定为耳间水平失真的主要来源，引入高达10.4 ± 0.5 dB的误差。而包络提取和通道选择是耳间时间退化的主要贡献者，误差高达101 ± 17 µs。在左右处理器之间联动通道选择可在强头影条件下减少水平失真（从1.8 ± 0.2 dB降至0.2 ± 0.1 dB），然而这种改善在后续处理阶段并不持续，并且还引入了显著的时间失真（高达338 ± 14 µs）。将最大值从8增加到12可提供适度益处，而进一步增加则带来很少额外改善，这同样在后续处理阶段不持续。言语可懂度模拟分析显示，联动通道选择在信噪比较差的一侧具有一致益处，将最大值从8增加到12至16时也是如此。结论：这些发现强调了耳间线索保留方面持续存在的局限性，并凸显了需要明确保留这些线索的编码策略。

英文摘要

Bilateral cochlear implantation is a primary clinical intervention to re-establish access to binaural hearing in individuals with severe-to-profound hearing loss. While many binaural auditory functions depend critically on interaural cues, existing sound coding strategies do not fully preserve these cues. Based on simulations, this study systematically examines how cochlear implant signal processing affects the encoding of interaural time and level differences across key processing stages. The impact of linked channel selection and number of maxima is also investigated on binaural cue encoding and speech intelligibility in noise. Non-linear compressions emerge as the dominant source of interaural level distortion, introducing errors up to 10.4 ± 0.5 dB. While envelope extraction and channel selection are the primary contributors to interaural timing degradation, with errors up to 101 ± 17 µs. Linking channel selection between left and right processors reduces level distortions under strong head-shadow conditions (from 1.8 ± 0.2 dB to 0.2 ± 0.1 dB), however, this improvement does not persist at later processing stages and also introduces substantial timing distortions (up to 338 ± 14 µs). Increasing maxima from 8 to 12 provides modest benefits, while further increases lead to little extra improvement, which also does not persist at later processing stages. Speech intelligibility simulation analyses show consistent benefits of linked channel selection on the side with poorer signal-to-noise ratio, also when increasing maxima from 8 to 12 to 16. These findings underscore persistent limitations in interaural cue preservation and highlight the need for coding strategies that explicitly preserve these cues.