Investigating semantic networks used for music and language Processing: Is a cross-domain transfer of skills possible?

多年来，音乐和语言处理之间的关系一直吸引着科学家，许多研究人员认为音乐专业水平和语言表现之间存在积极联系。例如，专业音乐家在学习新单词（Dittinger et al., 2016）和人工歌唱语言（franois & Schon, 2011）方面似乎比非音乐家更快。此外，Elmer及其同事（2014）比较了专业音乐家、同声传译员和对照组之间的语音、音乐和中性声音的分类。正如预期的那样，音乐家在音乐分类方面表现得更好，而同声传译者在语言分类方面表现得更好。然而，同声传译者在音乐任务中的表现也比对照组好，这表明学习在这两个领域之间迁移。因此，音乐和语言专长似乎具有双向转移效应。但是，音乐和语言之间的技能转移究竟是如何发生的呢？一些研究表明，音乐和语言都传达了感知、语义和句法意义，因此音乐练习可以改善语言处理，部分原因是共享神经网络的训练。然而，之前的研究使用了包含语言和音乐信息的刺激，并没有试图将两者分开（例如Steinbeis & Koelsch， 2011）。因此，观察到的音乐和语言处理之间的相互作用可能是由于同时激活的网络之间的干扰，而不是一个网络内部的干扰。事实上，有来自脑损伤患者的证据表明，音乐和语言处理是完全可分离的（例如Wilson， Parsons & Reutens, 2006; Ayotte, Peretz & Hyde, 2002）。为了解决这个问题，将语言和音乐材料在视觉和听觉领域提供的信息和反应分开是很重要的。因此，这个博士项目的目的是使用行为和电生理（ERP）措施来研究参与处理音乐和语言材料中语义信息的大脑网络。有三个主要的研究问题：(1)在音乐和语言材料的处理过程中，是相同的语义网络，还是它们可以在功能上分开？(2)专业知识（音乐训练或双语）是否只会带来特定领域的处理好处（例如，音乐训练会增强对音乐材料的处理），还是也会转移到其他领域（例如，音乐训练会增强对语言材料的处理）。如果观察到远转移效应，促进这种效应的机制是什么？(3)短期的音乐训练是否对语言处理困难（如阅读障碍）的人群有益？重要的是，通过研究有学习困难（如阅读障碍）的人在处理音乐和语言材料时所涉及的语义网络，可以获得很大的见解。阅读障碍是一种识别和阅读单词的困难，被认为源于感知、词汇和/或语义缺陷（Torkildsen et al., 2007）。在一些使用脑电图的语言学研究中，阅读障碍患者在处理语义和句法违规时表现出与对照组不同的大脑反应（Horowitz-Kraus & Breznitz, 2008; Sabisch et al., 2006；孟，Tiay, Jian & Zhou, 2007）。有趣的是，到目前为止，还没有研究包括音乐刺激，并探索在阅读障碍中发现的加工缺陷是否可以推广到音乐材料。这一发现不仅支持共享音乐语言处理网络的观点，而且还表明，旨在减轻阅读障碍症状的干预措施可能包括音乐技能的训练。

The relationship between music and language processing has fascinated scientists for many years and a number of researchers have suggested a positive link between levels of musical expertise and language performance. For example, professional musicians seem to be faster at learning novel words (Dittinger et al., 2016) and artificial sung languages (François & Schon, 2011) than non-musicians. Furthermore, Elmer and colleagues (2014) compared the categorisation of speech, music and neutral sounds between professional musicians, simultaneous interpreters and controls. As might be expected, musicians performed better at musical categorisation and simultaneous interpreters were superior at speech categorisation. However, simultaneous interpreters also performed better at in the music task than controls, suggesting learning transfer across these two domains. Therefore, music and language expertise appear to have bidirectional transfer effects. But how exactly does skill transfer between music and language occur? Several studies propose that both music and language convey perceptual, semantic and syntactic meaning, and thus musical practice improves language processing due, in part, to training of shared neural networks. However, previous studies have used stimuli that contain both linguistic and musical information, with no attempt to separate the two (e.g. Steinbeis & Koelsch, 2011). Therefore, observed interactions between music and language processing could be due to interference between simultaneously activated networks rather than interference within one network. In fact, there is evidence from brain-damaged patients suggesting that music and language processing to be fully dissociable (e.g. Wilson, Parsons & Reutens, 2006; Ayotte, Peretz & Hyde, 2002). In order to address this paradigm, it is important to separate information and responses provided by linguistic and musical material in both the visual and auditory domainsThe aim of this PhD project is therefore to use behavioural and electrophysiological (ERP) measures to investigate brain networks involved in processing semantic information in musical and linguistic material. There are three main research questions: (1) Are the same semantic networks engaged in the processing of musical and linguistic material or can they be separated functionally? (2) Does expertise (musical training or bilingualism) only results in domain-specific processing benefits (e.g., musical training leads to enhanced processing of musical material) or does it also transfer to other domains (e.g., musical training enhances processing of linguistic material). If far-transfer effects are observed, what are the mechanisms facilitating such an effect? (3) Could a short period of musical training benefit populations with language processing difficulties (e.g. dyslexia)? Importantly, great insight could be gained from investigating semantic networks involved in the processing of both musical and linguistic material in people with learning difficulties, such as dyslexia. Dyslexia is a difficulty in recognizing and reading of words that is believed to stem from perceptual, lexical and/or semantic deficits (Torkildsen et al., 2007). In several linguistic studies using EEG, people with dyslexia show different brain responses compared to controls when processing semantic and syntactic violations (Horowitz-Kraus & Breznitz, 2008; Sabisch et al., 2006; Meng, Tiay, Jian & Zhou, 2007). Interestingly, so far no study has included musical stimuli and explored whether processing deficits found in dyslexia can generalize to musical material. Such finding would not only support the idea of shared music-language processing network but would also suggest that interventions aiming to reduce symptoms of dyslexia could include the training of musical skills.