Rationale and Implementation of a Bimodal Fitting Formula

Introduction<br>The effectiveness of cochlear implants has increased markedly over the past years. As a consequence, candidacy for implantation has been relaxed and more CI users have functionally useful contralateral hearing. Such bimodal users often report improved speech understanding, particularly in noisy situations, better tonal quality and improved lateralization of sound sources. The individual bimodal benefit, however, varies considerably. A potential reason for this is that conventional hearing aid fitting does not account for certain specific characteristics of bimodal listening. Of particular importance are: low-frequency audibility, spectral overlap of electric and acoustic stimulation, loudness balance, inter-aural tonotopy and (dynamic) synchronization of adaptive signal processing. While many of these requisites are highly individual and thus, may have to be addressed by individual fine-tuning, others might be accounted for, at least partially, by a prescriptive fitting formula specifically designed for bimodal listeners. Method<br>In order to improve the efficiency of bimodal fitting, a bimodal fitting formula is being developed that aims at accounting for the specific characteristics of bimodal listening. This formula can use any established fitting formula (e.g. DSL v5, NAL-RP) as a starting point. In the current implementation, Phonak’s proprietary formula (Adaptive Phonak Digital) is used. The importance of low-frequency audibility is addressed by maximizing effective audibility [1] based on the individual audiogram. Edge frequencies of Dead Regions can be either entered manually or are estimated automatically from the audiogram. As a second step, the static and dynamic behavior of the CI’s AGC has been approximated with the hearing aid’s compression system to account for loudness balance and dynamic synchronization between electric and acoustic hearing. This is accomplished by a special parameterization of channel-coupling, time constants, compression knee-points and compression ratios in the hearing aid. Results<br>Compared to conventional hearing fitting aid formulae, the novel bimodal fitting formula typically results in decreased gain at high frequencies and increased gain at low frequencies. Technical measurements using percentile analysis show that the dynamic behavior of the CI’s AGC can be approximated with the HA’s AGC nicely for speech signals. First results from studies using the bimodal fitting formula indicate that this formula can improve speech understanding in noise and provides a good starting point for individual fine-tuning.引言 人工耳蜗（Cochlear Implants, CI）的临床效果在近年来得到了显著提升。随之而来的是，人工耳蜗植入的适应症标准有所放宽，更多人工耳蜗使用者保留了具备实用功能的对侧听力。此类双模式聆听者通常报告称，其言语理解能力得到改善——尤其是在嘈杂环境中，且音质更佳、声源定位能力也有所提升。然而，个体间的双模式聆听获益差异显著。造成这一现象的潜在原因之一是，传统助听器验配公式并未考虑双模式聆听的特定特征。其中尤为关键的要点包括：低频可听度、电刺激与声刺激的频谱重叠度、响度平衡、耳间音调拓扑以及自适应信号处理的（动态）同步性。尽管上述诸多必备条件具有高度的个体差异性，需通过个性化微调予以适配，但其中部分条件至少可通过专为双模式聆听者设计的标准化验配公式得到部分满足。 方法 为提升双模式验配的效率，本研究开发了一款适配双模式聆听特征的验配公式。该公式可依托任意已成熟的验配公式（如DSL v5、NAL-RP）作为初始基准。在当前的实现版本中，采用了峰力（Phonak）的专有验配公式（Adaptive Phonak Digital）。针对低频可听度的重要性，本公式基于个体听力图最大化有效可听范围[1]。听阈死区（Dead Regions）的边界频率可手动输入，或通过听力图自动估算得到。第二步，本公式通过将助听器（Hearing Aid, HA）的压缩系统与人工耳蜗的自动增益控制（Automatic Gain Control, AGC）的静态与动态特性进行匹配，以实现电听觉与声听觉间的响度平衡与动态同步。这一匹配过程通过对助听器的通道耦合、时间常数、压缩拐点以及压缩比进行特殊参数化设置得以实现。 结果 与传统助听器验配公式相比，这款新型双模式验配公式通常可在高频段降低增益，在低频段提升增益。采用百分位分析的技术测试结果表明，针对言语信号，人工耳蜗的自动增益控制动态特性可通过助听器的自动增益控制实现良好匹配。采用该双模式验配公式的初步研究结果显示，该公式可改善嘈杂环境下的言语理解能力，并可为后续的个性化微调提供良好的初始基准。