The state-of-the-art article by van Hell provided an excellent overview of the current state of the science in the neural and neurocognitive basis of second language (L2) processing and learning. While the target article devoted much effort to reviewing studies related to the syntactic and semantic components of language and to a lesser extent to the lexicon, it is important to also consider the phonetic and phonological components of language in L2 research. I have highlighted some of the findings in this area of research and discussed some potential new directions.

Successful (spoken) L2 learning includes extracting phonetic and phonological information from the speech stream. The issues raised by van Hell such as the critical period hypothesis, age of acquisition, proficiency, and individual differences have also been studied in the context of these components (e.g., Golestani & Zatorre, 2004). This line of research often focused on individual differences and demonstrated that pretraining neural differences may forecast learning success at the group level (e.g., Sheppard et al., 2012). Future studies can explore how individual differences in neural speech tracking of different chunk sizes (e.g., Ding et al., 2015) may lead to differences in L2 learning outcomes.

To investigate individual differences, research must augment analytics that are designed for observing group-level performance by also using methods that are precise enough for making individual-level predictions. In research on first language acquisition (Wong et al., 2021), machine learning techniques have been adopted to make predictions about individual learners’ learning outcomes with very promising prediction performance. The use of such techniques has begun in L2 learning as well (Feng et al., 2021). In addition to forecasting learning success, future research can also predict differences in response to different types of interventions, so that training can be altered before it even begins in order to optimize learning for every learner.

In addition to investigating learner-internal individual difference variables, as reviewed by van Hell (see Wong et al., 2022, for potential genetic variables), L2 research has also examined how different learner-external variables (e.g., training methods such as explicit training) lead to better or worse outcomes as discussed in the target article. To inform pedagogical practice, research must also consider how subject-internal and subject-external variables interact. Some of this learner-by-training research has been conducted in phonetic and phonological learning as well. Different training methods can lead to different brain activities in foreign speech learning (Deng et al., 2018). Methods that allow for more precise individual-level prediction such as machine learning, coupled with studies that investigate how different types of training should be prescribed to different learners, would have the best chance of enabling personalized learning (Wong et al., 2017).

In her article, van Hell also discussed the importance of conducting ecologically valid research in L2 learning. From the neurocognitive perspective, one new avenue of research could include an understanding of how brains of learners and teachers interact in natural interactions, including studies of brain synchronies in a conversation involving a L2 as well as L2 learning in a classroom. There is a small but growing literature on using hyperscanning techniques to study first language acquisition. Those studies have typically investigated parent–child dyads communicating, with parents speaking in child-directed speech. Independently, research has also begun to examine student-to-student and student-to-teacher brain correlations in a classroom, with as many as 12 brains being examined at the same time (Dikker et al., 2017). Lessons from these studies in terms of both technical and neural theoretical contributions can propel L2 research to be more ecologically valid in the learning of sound and other components of language.

Beside its translational impact, research concerning phonetics and phonology in L2 learning can shed light on the basic mechanisms of learning and processing. An important property of phonetic and phonological features in spoken language is that they are acoustic in nature, which necessitates processing by the neural auditory system. Therefore, the study of the neural basis of speech can inform researchers about the fundamentals of how acoustically and functionally complex sounds are processed in the central nervous system. In fact, studies have investigated the interaction of music and speech in L2, both of which are acoustically and functionally complex. While much research on L2 has focused on cortical structures as reviewed by van Hell's target article, studies of music and speech additionally can allow researchers to investigate functions of subcortical neural centers such as the inferior colliculus that may not be engaged during the processing of other language components.

The field of L2 learning has expanded in unprecedented scope in the past decades to include not only behavioral research across different disciplines but also neuroscience fields. Researchers now know much more about how the nervous system processes and learns two or more languages. The goal of this commentary was not to provide a comprehensive review but to highlight just some of the research studies that have been conducted in phonetics and phonology and to discuss areas of potential future direction. By conducting research that targets individual-level prediction, learner-by-training interaction, and brain-to-brain correlations, and by studying the language system holistically, researchers will reach newer heights in understanding the basic mechanisms behind learning. Furthermore, this research will put researchers in a much stronger position for developing pedagogical strategies to make learning most effective for all learners.

van Hell 撰写的这篇最先进的文章对第二语言 (L2) 处理和学习的神经和神经认知基础的科学现状进行了精彩的概述。虽然目标文章投入了大量精力来回顾与语言的句法和语义成分相关的研究，并在较小程度上回顾了词汇，但在第二语言研究中考虑语言的语音和音韵成分也很重要。我重点介绍了该研究领域的一些发现，并讨论了一些潜在的新方向。

成功的（口语）L2 学习包括从语音流中提取语音和音韵信息。van Hell 提出的问题，如关键期假说、习得年龄、熟练程度和个体差异，也在这些组成部分的背景下进行了研究（例如，Golestani & Zatorre，2004 ）。这一系列研究通常关注个体差异，并证明预训练神经差异可以预测群体层面的学习成功（例如，Sheppard 等人，2012）。未来的研究可以探索不同块大小的神经语音跟踪的个体差异（例如，Ding et al., 2015）如何导致 L2 学习结果的差异。

为了调查个体差异，研究必须通过使用足够精确的方法来进行个体层面的预测，从而增强旨在观察群体层面表现的分析。在第一语言习得的研究中（Wong et al., 2021），机器学习技术已被用来对个体学习者的学习成果进行预测，具有非常有前途的预测性能。此类技术的使用也已开始在 L2 学习中使用（Feng et al., 2021）。除了预测学习成功之外，未来的研究还可以预测对不同类型干预措施的反应差异，以便在培训开始之前就可以改变培训，以优化每个学习者的学习。

除了调查学习者内部个体差异变量（如 van Hell 所评论的那样）（参见 Wong 等人，2022，了解潜在的遗传变量），L2 研究还研究了不同的学习者外部变量（例如，显式训练方法）如何影响学习者的内部个体差异变量。正如目标文章中所讨论的，培训）会导致更好或更差的结果。为了指导教学实践，研究还必须考虑学科内部变量和学科外部变量如何相互作用。一些通过训练学习者的研究也在语音和语音学习中进行。不同的训练方法会导致外语学习中不同的大脑活动（Deng et al., 2018））。允许更精确的个人水平预测的方法（例如机器学习），再加上研究如何为不同的学习者制定不同类型的培训，将最有可能实现个性化学习（Wong et al., 2017 ）。

van Hell 在她的文章中还讨论了在第二语言学习中进行生态有效研究的重要性。从神经认知的角度来看，一种新的研究途径可能包括了解学习者和教师的大脑如何在自然互动中互动，包括研究涉及 L2 的对话以及课堂上的 L2 学习中的大脑同步性。关于使用超扫描技术来研究第一语言习得的文献虽少，但仍在不断增加。这些研究通常调查亲子二人组的沟通，即父母以孩子为主导的语言说话。独立地，研究也开始检查教室中学生与学生以及学生与老师的大脑相关性，同时检查多达 12 个大脑（Dikker 等人，2017 年））。这些研究在技术和神经理论贡献方面的经验教训可以推动第二语言研究在声音和语言其他组成部分的学习中更具生态有效性。

除了翻译影响之外，有关第二语言学习中语音学和音韵学的研究还可以揭示学习和处理的基本机制。口语中语音和音韵特征的一个重要属性是它们本质上是声学的，这需要神经听觉系统的处理。因此，对语音神经基础的研究可以让研究人员了解中枢神经系统如何处理声学和功能上复杂的声音的基础知识。事实上，研究已经调查了第二语言中音乐和语音的相互作用，这两者在声学和功能上都很复杂。虽然正如 van Hell 的目标文章所评论的那样，关于 L2 的许多研究都集中在皮质结构上，

在过去的几十年里，第二语言学习领域得到了前所未有的扩展，不仅包括跨学科的行为研究，还包括神经科学领域。研究人员现在对神经系统如何处理和学习两种或多种语言有了更多的了解。本评论的目的不是提供全面的综述，而是强调语音学和音韵学领域已进行的一些研究，并讨论未来潜在方向的领域。通过开展针对个体水平预测、学习者训练交互和大脑与大脑关联的研究，并通过整体研究语言系统，研究人员将在理解学习背后的基本机制方面达到新的高度。此外，