鸡尾酒会效应的关键性脑机制

Under a noisy “cocktail-party” environment, listeners can use various perceptual cues to improve their attention to and recognition of the target speech. To further explore the "cocktail-party" problem, this project will use the methods of psychological physics, functional magnetic resonance imaging (fMRI), magnetoencephalography (MEG), and intracranial electroencephalogram (ECoG) to systematically investigate both the cue-specific and cue-nonspecific psychological and neural mechanisms underlying the enhancement of speech recognition under speech information masking conditions. The four typically selected unmasking cues are voice priming stimuli, lip reading priming stimuli, and speech content priming stimuli, and perceived spatial separation between target speech and masking speech. This project will focus on the basic mechanism underlying various unmasking processes associated with the perceptual cues, based on both the binding problem and the feature capture theory. Accordingly, the following hypothesis will be tested: A unmasking perceptual cue, which represents a feature of the target speech object, facilitates both the saliency of the target-speech object and the recognition of target speech content by enhancing the perceptual binding between various features of the target-speech object (particularly between the envelope component and the fine-structure component of the target-speech signals), inhibiting masking signals, and activating the central speech-production (motor) systems. The results of this project will not only reveal the relationship between the cocktail-party problem and the binding problem, but also be valuable in applications in the fields of industry, medicine, and national defense.

在有多人说话的嘈杂环境下，听者为何能利用知觉或认知线索来强化对目标言语的觉察、注意与识别? 为解答这一重要的鸡尾酒会问题，本项目运用行为测查、功能磁共振成像、脑磁图、和颅内脑电记录的方法来深入考察年轻健康被试利用去掩蔽线索来强化对被掩蔽言语的识别的线索特异性脑机制以及非线索依赖性的脑机制。所考察的线索包括，嗓音启动、唇读启动、语义内容启动、和主观空间分离线索。核心假设是，一个与目标言语客体中的某个特征相关联的去掩蔽线索因强化了以下几种脑网络环路的协同活动而产生对目标言语的去掩蔽作用：注意环路、言语表达环路、抑制功能环路、和言语运动环路。基于特征捕捉理论，研究的重点是，去掩蔽线索如何强化目标言语信号的包络成份和精细结构成份在这些环路中的动态性表达及其相互之间的“捆绑”。本项目不但要推动对鸡尾酒会效应脑机制的理解，研究成果也将在人工智能、精神医学、军队特殊人才的选拔和训练等领域有实际应用价值。

在有多人说话的嘈杂环境下，听者为何能利用知觉或认知线索来强化对目标言语的觉察、注意与识别? 为解答这一重要的鸡尾酒会问题，本项目运用行为测查、功能磁共振成像、脑磁图等方法来深入考察被试利用去掩蔽线索来强化对被掩蔽言语的脑机制。所考察的线索包括，嗓音启动、唇读启动、语义内容启动、和主观空间分离线索。所得结果支持了核心假设，一个与目标言语客体中的某个特征相关联的去掩蔽线索因强化了以下几种脑网络环路的协同活动而产生对目标言语的去掩蔽作用：注意环路、言语表达环路、抑制功能环路、和言语运动环路。基于特征捕捉理论，研究的重点发现是，去掩蔽线索强化了目标言语信号的包络成份和精细结构成份在这些环路中的动态性表达及其相互之间的“捆绑”。本项目不但要推动对鸡尾酒会效应脑机制的理解，研究成果也将在人工智能、精神医学、军队特殊人才的选拔和训练等领域有实际应用价值。