Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China.
Department of Biomedical Engineering, School of Medicine, Tsinghua University, Beijing, China; Research Center of Biomedical Engineering, Graduate School at Shenzhen, Tsinghua University, Shenzhen, Guangdong Province, China.
Neuroscience. 2018 Mar 1;373:60-71. doi: 10.1016/j.neuroscience.2018.01.009. Epub 2018 Jan 11.
Since sound perception takes place against a background with a certain amount of noise, both speech and non-speech processing involve extraction of target signals and suppression of background noise. Previous works on early processing of speech phonemes largely neglected how background noise is encoded and suppressed. This study aimed to fill in this gap. We adopted an oddball paradigm where speech (vowels) or non-speech stimuli (complex tones) were presented with or without a background of amplitude-modulated noise and analyzed cortical responses related to foreground stimulus processing, including mismatch negativity (MMN), N2b, and P300, as well as neural representations of the background noise, that is, auditory steady-state response (ASSR). We found that speech deviants elicited later and weaker MMN, later N2b, and later P300 than non-speech ones, but N2b and P300 had similar strength, suggesting more complex processing of certain acoustic features in speech. Only for vowels, background noise enhanced N2b strength relative to silence, suggesting an attention-related speech-specific process to improve perception of foreground targets. In addition, noise suppression in speech contexts, quantified by ASSR amplitude reduction after stimulus onset, was lateralized towards the left hemisphere. The left-lateralized suppression following N2b was associated with the N2b enhancement in noise for speech, indicating that foreground processing may interact with background suppression, particularly during speech processing. Together, our findings indicate that the differences between perception of speech and non-speech sounds involve not only the processing of target information in the foreground but also the suppression of irrelevant aspects in the background.
由于声音感知是在一定背景噪声下进行的，因此语音和非语音处理都涉及到目标信号的提取和背景噪声的抑制。以前关于语音音位早期处理的研究在很大程度上忽略了背景噪声是如何编码和抑制的。本研究旨在填补这一空白。我们采用了一种oddball 范式，在有或没有背景调幅噪声的情况下呈现语音（元音）或非语音刺激（复合音），并分析与前景刺激处理相关的皮质反应，包括失匹配负波（MMN）、N2b 和 P300，以及背景噪声的神经表示，即听觉稳态反应（ASSR）。我们发现，与非语音刺激相比，语音刺激诱发的 MMN 潜伏期更长、强度更弱，N2b 和 P300 的潜伏期也更长，但 N2b 和 P300 的强度相似，这表明语音中某些声学特征的处理更为复杂。只有在元音中，背景噪声增强了 N2b 的强度，而不是在沉默中，这表明了一种与注意力相关的语音特异性过程，以提高对前景目标的感知。此外，在刺激开始后，通过 ASSR 振幅减小来量化语音背景噪声的抑制，这种抑制偏向于左半球。N2b 之后的左侧抑制与语音中 N2b 增强后的噪声抑制有关，表明前处理可能与背景抑制相互作用，尤其是在语音处理中。总的来说，我们的研究结果表明，语音和非语音感知之间的差异不仅涉及到前导目标信息的处理，还涉及到背景中不相关方面的抑制。
