Wellcome Centre for Human Neuroimaging, UCL, London, UK.
Institute of Neuroscience, Newcastle University, Newcastle upon Tyne, UK.
Sci Rep. 2019 Nov 14;9(1):16771. doi: 10.1038/s41598-019-53353-5.
Understanding speech when background noise is present is a critical everyday task that varies widely among people. A key challenge is to understand why some people struggle with speech-in-noise perception, despite having clinically normal hearing. Here, we developed new figure-ground tests that require participants to extract a coherent tone pattern from a stochastic background of tones. These tests dissociated variability in speech-in-noise perception related to mechanisms for detecting static (same-frequency) patterns and those for tracking patterns that change frequency over time. In addition, elevated hearing thresholds that are widely considered to be 'normal' explained significant variance in speech-in-noise perception, independent of figure-ground perception. Overall, our results demonstrate that successful speech-in-noise perception is related to audiometric thresholds, fundamental grouping of static acoustic patterns, and tracking of acoustic sources that change in frequency. Crucially, speech-in-noise deficits are better assessed by measuring central (grouping) processes alongside audiometric thresholds.
在存在背景噪声的情况下理解言语是一项至关重要的日常任务，不同人之间存在很大差异。一个关键的挑战是理解为什么有些人尽管听力临床正常，但在语音感知方面仍存在困难。在这里，我们开发了新的图形-背景测试，要求参与者从随机背景音中提取出连贯的音调模式。这些测试区分了与检测静态（同频）模式和跟踪随时间变化频率的模式相关的语音感知变异性。此外，被广泛认为是“正常”的听力阈值升高，可独立于图形-背景感知，解释语音感知方面的显著差异。总的来说，我们的研究结果表明，成功的语音感知与听阈、静态声模式的基本分组以及随频率变化的声源跟踪有关。至关重要的是，通过同时测量中心（分组）过程和听阈，可以更好地评估语音感知缺陷。
