Sorensen David O, Sugai Jenna A, Parthsarathy Aravindakshan, Hancock Kenneth E, Polley Daniel B
bioRxiv. 2025 May 5:2025.05.05.652235. doi: 10.1101/2025.05.05.652235.
Among all the sounds occurring at any given time, people are often interested in listening to just one. Some competing sounds are merely background noise, whereas others distract attention from target sounds and are less easily suppressed. During active listening, the central auditory pathway unmixes target and distractor sounds based on temporal differences that vary across three orders of magnitude - from millisecond differences in acoustic temporal fine structure to slower perceptual grouping factors that stretch out to multiple seconds. Here, we developed an approach to directly measure central auditory encoding of multiplexed target and distractor sound features in human listeners to determine which timescales are most impacted by the presence of distracting sounds. Target sounds contained nested features along four timescales, including temporal fine structure (∼500 Hz), temporal envelope (∼25-80 Hz), envelope changes (∼5 Hz), and slower changes in embedded context reflecting whether target stimuli were randomly arranged or formed a repeating pattern (∼0.5 Hz). Targets were presented with competing sounds that provided variable levels of distraction: either a highly distracting melody or a less distracting noise. Neural synchronization to each timescale was simultaneously and independently measured for target and distractor sounds from electroencephalogram (EEG) recordings during a listening task. Sustained shifts from random to regular arrangements of temporal sequences were reliably perceived, yet did not evoke a pattern recognition potential, nor neural synchronization changes at any timescale. Synchronization to relatively slow changes in envelope transitions (<10Hz) of the target sound deteriorated with the addition of a more distracting sound while synchronization to more rapid fluctuations in the fine structure or envelope modulation rate were unaffected by varying levels of distraction. Categorizing trials according to task performance revealed a conjunction of enhanced entrainment to slower temporal features in the distractor sound and reduced synchronization to the target sound on error trials. By designing a stimulus paradigm that leveraged the remarkable temporal processing capabilities of the auditory nervous system, we were able to simultaneously quantify multiple target and distractor sound features reproduced in the EEG. This paradigm identified synchronization processes in the 7-10 Hz alpha range that has been linked to distractor suppression, which may prove valuable for research on clinical populations who report difficulty suppressing awareness of distracting sounds.
在任何给定时间出现的所有声音中,人们通常只对其中一种声音感兴趣。一些竞争声音仅仅是背景噪音,而其他声音则会分散对目标声音的注意力,并且更难被抑制。在主动聆听过程中,中枢听觉通路根据时间差异将目标声音和干扰声音区分开来,这些时间差异跨越三个数量级——从声学时间精细结构中的毫秒差异到延伸至数秒的较慢的感知分组因素。在这里,我们开发了一种方法,直接测量人类听众对多路复用的目标声音和干扰声音特征的中枢听觉编码,以确定哪些时间尺度受干扰声音的存在影响最大。目标声音包含沿四个时间尺度嵌套的特征,包括时间精细结构(约500赫兹)、时间包络(约25 - 80赫兹)、包络变化(约5赫兹)以及反映目标刺激是随机排列还是形成重复模式的嵌入语境中的较慢变化(约0.5赫兹)。目标声音与提供不同程度干扰的竞争声音同时呈现:要么是高度干扰的旋律,要么是干扰较小的噪音。在聆听任务期间,从脑电图(EEG)记录中同时独立测量目标声音和干扰声音与每个时间尺度的神经同步。从随机到规则的时间序列排列的持续变化能够被可靠地感知到,但不会引发模式识别电位,也不会在任何时间尺度上引起神经同步变化。随着添加更具干扰性的声音,与目标声音包络过渡(<10赫兹)中相对较慢变化的同步性变差,而与精细结构或包络调制率中更快波动的同步性不受干扰程度变化的影响。根据任务表现对试验进行分类显示,在错误试验中,对干扰声音中较慢时间特征的同步增强与对目标声音的同步减少同时出现。通过设计一种利用听觉神经系统卓越时间处理能力的刺激范式,我们能够同时量化脑电图中再现的多个目标声音和干扰声音特征。这种范式识别出了与干扰抑制相关的7 - 10赫兹阿尔法范围内的同步过程,这对于那些报告难以抑制对干扰声音意识的临床人群的研究可能具有重要价值。