The ‘sensitive period’ for phonetic learning posits that between 6 and 12 months of age, infants’ discrimination of native and nonnative speech sounds diverge. Individual differences in this dynamic processing of speech have been shown to predict later language acquisition up to 30 months of age, using parental surveys. Yet, it is unclear whether infant speech discrimination could predict longer-term language outcome and risk for developmental speech-language disorders, which affect up to 16 % of the population. The current study reports a prospective prediction of speech-language skills at a much later age—6 years-old—from the same children’s nonnative speech discrimination at 11 months-old, indexed by MEG mismatch responses. Children’s speech-language skills at 6 were comprehensively evaluated by a speech-language pathologist in two ways: individual differences in spoken grammar, and the presence versus absence of speech-language disorders. Results showed that the prefrontal MEG mismatch response at 11 months not only significantly predicted individual differences in spoken grammar skills at 6 years, but also accurately identified the presence versus absence of speech-language disorders, using a machine-learning classification. These results represent new evidence that advance our theoretical understanding of the neurodevelopmental trajectory of language acquisition and early risk factors for developmental speech-language disorders.

语音学习的“敏感期”假设在 6 到 12 个月大的婴儿对母语和非母语语音的辨别力存在差异。使用父母调查显示，这种动态语音处理的个体差异可以预测 30 个月大的以后的语言习得。然而，尚不清楚婴儿言语歧视是否可以预测长期语言结果和发展性言语语言障碍的风险，这会影响多达 16% 的人口。目前的研究报告了对相同儿童在 11 个月大时的非母语语言歧视的未来年龄（6 岁）的语音语言技能的前瞻性预测，由 MEG 不匹配反应索引。语言病理学家通过两种方式综合评估 6 岁儿童的语言技能：口语语法的个体差异，以及是否存在语言障碍。结果表明，11 个月时的前额叶 MEG 不匹配反应不仅显着预测了 6 岁时口语语法技能的个体差异，而且还使用机器学习分类准确地识别了语音语言障碍的存在与否。这些结果代表了新的证据，可以促进我们对语言习得的神经发育轨迹和发育性语言障碍的早期危险因素的理论理解。结果表明，11 个月时的前额叶 MEG 不匹配反应不仅显着预测了 6 岁时口语语法技能的个体差异，而且还使用机器学习分类准确地识别了语音语言障碍的存在与否。这些结果代表了新的证据，可以促进我们对语言习得的神经发育轨迹和发育性语言障碍的早期危险因素的理论理解。结果表明，11 个月时的前额叶 MEG 不匹配反应不仅显着预测了 6 岁时口语语法技能的个体差异，而且还使用机器学习分类准确地识别了语音语言障碍的存在与否。这些结果代表了新的证据，可以促进我们对语言习得的神经发育轨迹和发育性语言障碍的早期危险因素的理论理解。