中文摘要：

本研究根据音乐和语言的层级结构，从各个加工阶段出发，探讨音乐和语言的关系。具体而言，我们对已有的音乐与语言的元分析结果进行对比，并在此基础上探讨了音乐和语言在神经基础上的异同。结果表明，语音分析与音乐加工的重合发生在听觉-运动环路，与音程分析的重合节点在左侧中央前回，与结构分析的在左侧额盖区；语义分析与音乐加工的重合发生在核心环路，重合节点分别是左侧颞上回和左侧额下回（音程分析），以及右侧额下回（结构分析）；句子分析和音乐知觉的重合发生在认知-情绪环路中的右侧脑岛。此结果暗示了音乐旋律分析与词义加工的关系可能是音乐语言共享神经基础的核心部分。同时，分别计算音乐两层级与语言簇的重合率发现，音程分析和语言加工重合更多，在大脑左侧有50%重合，而结构分析与语言加工重合较少，在大脑双侧分别只有7%和14%重合，暗示了在较为低层级的加工阶段，音乐和语言更可能存在共享的神经基础，而在更为高层级的加工阶段，二者更多地表现出各自独特的特点。

英文摘要：

Music and language are arguably the most characteristic traits of human beings. On one hand, previous functional magnetic resonance imaging （fMRI） studies have identified multiple cortical regions that are involved in the processing of both music and language, suggesting shared neural substrate for music and language. On the other hand, neuropsychological studies on brain-lesioned patients show the double dissociation between music and language, suggesting distinct neural substrates for music and language respectively. Here we used meta-analysis to examine the relation of the neural basis underlying music and language. 〈br〉 First, we adopted the data from three meta-analysis studies on music and language respectively. Specifically, for the neural substrate of music, we focused on two processing levels specific to music processing, which were interval analysis （15 contrasts and 63 peaks） and structure analysis （19 contrasts and 217 peaks） （Lai, Xu, Song,＆amp;Liu, 2013）. For the neural substrate of language, three processing levels specific to language processing were selected, which were phonological analysis （86 contrasts and 344 peaks）, lexico-semantic analysis （111contrasts and 339 peaks）, and sentence analysis （65 contrasts and 218 peaks） （Vigneau et al., 2006;Vigneau et al., 2011）. Second, we projected these peak activation elicited by processing either music or language onto the MNI （Montreal Neurological Institute） space to visualize the distribution of cortical regions involved in music and language with Caret. Finally, to explore the relation of neural substrates underlying music and language, we calculated whether pairs of peak activation were spatially overlapped or dissociated by K-means clustering and multivariate analysis of variance （MANOVA）. The overlapping rate was estimated to quantify the extent to which music and language shared common neural substrates. 〈br〉 We finally found 11 pairs of overlapping clusters. Music and language shared neural sub