中文摘要：

背景：以往借助脑成像技术研究双语的脑机制,但存在时间分辨率不足的缺陷,而利用高时间分辨率的相关电位技术可根据需要观察到任一时段的大脑活动情况。目的：利用事件相关电位技术探讨双语者加工母语及第二语言句子时的脑机制。设计、时间及地点：两因素重复测量试验,于2007-01在江苏省语言科学与神经认知工程重点实验室完成。参试者：选择年龄在21～23岁的18名在校英语本科大三学生志愿者参加。方法：选择120个简单的主谓宾句子为试验语料,包括60个汉语句子和60个英语句子,句子中每个单词呈现时间为300ms,刺激间隔为800ms,两个句子间的刺激间隔为2000ms。首先要求被试认真阅读屏幕上出现的句子,记录句子末尾词语的脑电;在经过3个（或4个、5个）句子后,出现一个符号“＋”作为提示,再出现一个句子,此时让被试做出反应,判断该句子是否是前面句子中出现过的,并以左右手作出“是”或“不是”的按键反应,记录反映时和准确率。主要观察指标：采用NeuroScan公司32导电极帽10/20系统记录脑电,离线分析反映人类内隐加工机制的N400和与外显加工机制有关的LPC成分。结果：在N400成分上,第二语言句子的平均波幅高于母语句子;而在LPC成分上,母语句子的平均波幅高于第二语言句子。结论：两种句子在加工时,都运用了内隐的加工机制和外显的加工机制,母语句子加工时运用了较多的内隐加工机制、较少的外显加工机制,而第二语言句子加工时运用了较少的内隐加工机制、较多的外显加工机制。两种句子加工时,内隐加工和外显加工呈现出互补的态势。

英文摘要：

BACKGROUND： Cerebral imaging is used to study bilingual processing, but it lacks time resolution. Event-related potentials （ERP） with high time resolution can be used to observe brain activities at any time period. OBJECTIVE： To investigate the neural mechanism of first and second language processing by bilinguals using ERP measures. DESIGN, TIME AND SETTING： Double repetitive measurement. The experiment was performed at the Key Laboratory of Language Sciences and Neurocognitive Engineering of Jiangsu Province in January 2007. PARTICIPANTS： Eighteen Junior students majoring in English, aged 21-23 years, were selected. METHODS： 120 simple sentences were used including 60 simple sentences in both Chinese and English. Each word/character was presented for 300 ms with 800 ms interval. And the interval between two sentences was 2 000 ms. Subjects were required to read each word/character presented on the screen. The ERP of the last word/character in each sentence was recorded. Every 3-5 sentences there was a plus sign indicating a sentence for subjects to decide whether this sentence had been presented. The subjects responded by pushing buttons with left hand for ＂yes＂ and right hand for ＂no＂. The reaction time and accuracy were recorded. MAIN OUTCOME MEASURES： The brain electricity was recorded by American Neuroscan electric conduction evoked potential workstation. N400 was used as an index of implicit processing mechanism, and P600 was used as an index of explicit processing mechanism. RESULTS： Two conditions elicited different N400 and LPC. In specific, the L1 （mother language） elicited N400 with lower magnitude than L2 （second language）, while for LPC, the magnitude for L1 was significantly higher than that for L2. CONCLUSION： Both implicit and explicit processing mechanisms are used for processing L1 and L2 sentences. The processing of L1 sentences relied more on implicit processing mechanism but less on explicit processing mechanism, while the opposite pattern is