School of Psychological Sciences, University of Manchester, Brunswick Street, Manchester, UK.
Neuroimage. 2010 Jan 1;49(1):1124-32. doi: 10.1016/j.neuroimage.2009.07.032. Epub 2009 Jul 24.
Listeners show remarkable flexibility in processing variation in speech signal. One striking example is the ease with which they adapt to novel speech distortions such as listening to someone with a foreign accent. Behavioural studies suggest that significant improvements in comprehension occur rapidly--often within 10-20 sentences. In the present experiment, we investigate the neural changes underlying on-line adaptation to distorted speech using time-compressed speech. Listeners performed a sentence verification task on normal-speed and time-compressed sentences while their neural responses were recorded using fMRI. The results showed that rapid learning of the time-compressed speech occurred during presentation of the first block of 16 sentences and was associated with increased activation in left and right auditory association cortices and in left ventral premotor cortex. These findings suggest that the ability to adapt to a distorted speech signal may, in part, rely on mapping novel acoustic patterns onto existing articulatory motor plans, consistent with the idea that speech perception involves integrating multi-modal information including auditory and motoric cues.
听众在处理语音信号变化方面表现出惊人的灵活性。一个引人注目的例子是,他们很容易适应新的语音失真,例如听带有外国口音的人说话。行为研究表明,理解能力会迅速提高——通常在 10-20 个句子内。在本实验中,我们使用时间压缩语音来研究在线适应失真语音的神经变化。听众在正常语速和时间压缩句子上执行句子验证任务,同时使用 fMRI 记录他们的神经反应。结果表明,时间压缩语音的快速学习发生在呈现前 16 个句子的第一块期间,与左右听觉联合皮层和左腹侧运动前皮层的激活增加有关。这些发现表明,适应失真语音信号的能力可能部分依赖于将新的声学模式映射到现有的发音运动计划上,这与语音感知涉及整合多模态信息(包括听觉和运动线索)的观点一致。