N400 was first reported by Kutas et al,1 who demonstrated that if the end word in a sentence was semantically anomalous, a negative-going potential would be recorded on the scalp about 400 ms after the onset of the word; it was called N400. Since this report, people recorded this negative-going potential during several different tasks; word/picture naming tasks, lexical decision tasks, or when misspelling existed.2 Kiang et al3 proved later that N400 would be elicited only if only the target word or sentence is non-predicted. N400 was also shown to be elicited in an auditory sentence paradigm. Currently, the tasks most often used for eliciting N400 are semantic categorization task, naming task, and lexical decision task.4-7
In published works, the stimuli words or sentences used in N400 researches were mainly in English.2,8-12 Seldom have authors used Chinese as stimuli materials in N400 researches. Unlike phonetic words in English, Chinese is ideographic. Chinese is completely different to Indo-European languages. Chinese characters (i.e. single Chinese character, which may not meaningful) are pictographic characters,8,9,12 more complicated than English words, in that every single Chinese character has its own shape, phoneme and meaning. The characters are made up of different strokes, and the combination of these strokes (shape) has particular pronunciation (phoneme) and meaning. In Chinese, although some of the single characters are meaningful Chinese words, others must be combined to make meaningful words. Most of the Chinese words are composed of two characters, expressing one or more concepts. Typically, a Chinese idiom is composed of four characters, which can express more meaning than a common word. Each Chinese character has one of the four tones, the level tone, the rising tone, the falling-rising tone, and the falling tone, which are also not seen in English words. Based on Chinese grammars, several Chinese characters, words or idioms can be joined together to express people's thoughts and feeling. So, the mechanism of language processing in Chinese may be totally different from English.9,12
Chinese has been used by some authors as stimuli to elicit N400,13-18 and they report a similar N400 effect using Chinese stimuli. However, the stimuli used in these studies are mostly Chinese sentences and Chinese words. There are few report of the N400 elicited by Chinese idioms, although N400 in response to idioms was mainly distributed in the left frontal, left anterior temporal, and centrofrontal regions. The present study was carried out from March 2007 to November 2008. In this study, mismatching and matching ended Chinese idioms were used to elicit the N400 in healthy subjects, and the features of N400 were analyzed. It was also a preliminary work for future use of N400 in the psychiatric clinic.
Participants were 62 healthy adults (33 men and 29 women), who were enrolled from March 2007 to November 2008 from the Shanghai Mental Health Center (employees), Shanghai Jiao Tong University (teachers or students) and nearby from the community. All participants met the criterions: (1) no positive finding in physical examination; (2) normal in blood routine, biochemical routine items and X-ray examination; (3) without a history of psychiatric disorders themselves or in their first degree relatives; (4) with normal or corrected-to-normal vision, with good hearing; (5) right-handed; (6) using mandarin as their mother tongue. The mean age of the participants was (29.6±6.6) years (range 19-50 years), and the mean education time was (16.7±2.0) years (range 14-22 years). No significant difference was found in age and education between male and female participants.
The stimuli were 38 pairs of matching and mismatching ended Chinese idioms. All idioms were chosen from the Chinese Idioms Dictionary (seen below). The target character was the fourth (ending) character in every idiom, which was either matching in 38 idioms (congruent condition), or mismatching in 38 idioms (incongruent condition), with the original idioms. In incongruent condition, the ending characters are either homophonically, orthographically, or semantically related or totally unrelated to the matching Chinese characters. All 76 idioms were presented to participants in random orders under a rule that congruent or incongruent idioms could not be presented sequential more then three times.
The examples of the four types of incongruent condition of idioms were as follows (the italic characters were mismatching ones): (1) Homophonically related mismatch ending idioms and corresponding matching ones: ai bu shi shou (head)— ai bu shi shou (hand). (2) Orthographically related mismatch ending idioms and corresponding matching ones: ju shi zhu yue (moon) — ju shi zhu mu (eye). (3) Semantically related mismatch ending idioms and corresponding matching ones: bu kan hui tou (pate) — bu kan hui shou (head). (4) Unrelated mismatch ending idioms and corresponding matching ones: bai zhe bu long (dragon) — bai zhe bu nao (broken). The task was similar to the Chinese character classified matching task.9,19
Each participant was seated in a comfortable chair at a distance of one meter from the screen. At the beginning, oral instruments about the experiment's purpose and demand were told to the participant. The instructions were as follows: "Some Chinese idioms will be presented on the screen in front of you. Please judge the correctness of the idioms’ last character. If you think the character is correct, please press the button ‘1’, and if you think t he character is incorrect, please press the button ‘2’. After the experiment was ended, the participant would be informed of their number of incorrect answers.
Event related potentials (ERP) recording
The N400 was recorded by an event related potentials recording system (Guangzhou Runjie, China, WJ-1). Original electroencephalogram (EEG) was recorded by means of electrodes from Fp1, Fp2, Fz, F3, F4, Cz, C3, C4, Pz, P3, P4, O1, O2, F7, F8, T3, T4, T5, and T6 based on the International 10-20 System. All the EEG electrodes were referenced to linked earlobes (A1+A2). The ground electrode was positioned on FPz. Eye movements and blinks were monitored via an electrode placed beneath the left eye. Electrode impedances were kept below 5 kΩ. Scalp EEG data were recorded with a 0.53-60 Hz filter. All EEG data were stored in a hard disk for subsequent analysis.
ERP analysis and statistical analysis
Peak negativity was measured for each condition during the latency band of N400 (250-500 ms) on Cz, Fz and Pz and the different N400 waves were computed between incongruent condition and congruent condition. The trials contaminated by artifacts were excluded from the analysis by the ERP recording instrument automatically. The data were expressed mean ± standard deviation (SD) and the ERPs statistical analysis was performed using SPSS 10.0. The difference of N400 amplitudes and latencies was assessed by paired-sample t test. The significance level for statistical tests was set at P <0.05 (two-tailed).
After 250-500 ms from the stimuli, a negative going wave was found under both incongruent condition (with larger amplitude) and congruent condition (with smaller amplitude).
N400 latency changed with stimuli types (sPdSdM, sSdPdM, sMdPdS, dPdSdM) and conditions (congruent, incongruent). The longest N400 latency was elicited by dPdSdM. No significant difference was found among the electrode sites of Fz, Cz and Pz. In four kinds of stimuli types, N400 showed longer latency in incongruent conditions than in congruent conditions (P <0.05, Table 1).
N400 amplitudes were also influenced by stimuli types (sPdSdM, sSdPdM, sMdPdS, dPdSdM). The largest N400 amplitude was elicited by dPdSdM in incongruent condition. In four kinds of stimuli types, incongruent conditions elicited larger N400 amplitude than congruent conditions (P <0.05, Table 2).
N400 could be elicited from all brain areas of scalp, but the amplitudes and latencies had slight differences in different brain area. The largest N400 was elicited at Cz.
Seven to sixteen days (mean 12 days) after the first experiment, four men and three women were selected randomly to perform the N400 experiment again. No significant difference was found between the first experiment and the second, indicating that N400 was a stable ERPs component in health subjects.
No gender differences were found between the 33 males and 29 females in their N400 latency and N400 amplitudes.
In the present study, we consider the fact that Chinese characters have a unity of shape, phoneme and meaning,20,21 and the cognitive processing of Chinese characters is a complex process.22 The incongruent stimuli used in this study consist of four kinds of characters; sPdSdM, sSdPdM, sMdPdS, and dPdSdM. In the analysis, ERPs data was averaged by congruent and incongruent conditions separately.
We demonstrated in the present study that when the last character (target character) of the Chinese idioms was inappropriate with the first three characters, i.e. if the last character was different (incongruent) to the character the participant expected, a N400 would be elicited in healthy adults. As the target character's congruity is determined by the participants after the processing of the stimuli, N400 can reflect the influence of expectation to the character before the stimuli presentation, i.e. the influence of congruent or incongruent, on inner language processing (shape, phoneme, meaning).12 Chinese idioms are comprised of four characters, and can not be understood from the individual meanings of the characters. As the characters in idioms are peculiar to themselves grammatically, participants have more expectations for the matching last character (congruent condition). This makes Chinese idioms good stimuli to study the cognitive processing of Chinese language.9,12 When the participants read silently the first three characters of an idiom, the expectation of the last character became stronger based on the context of the idiom. So, before the last character shows up on the screen, the last character can be activated in the participants’ brain. In the present study, the N400 elicited by matching ending-characters are smaller than that elicited by mismatching ending-character, demonstrating that participants expected less from the mismatching ending-character. N400s elicited by different types of stimuli showed different latencies and amplitudes, and the longest N400 latency and largest N400 amplitude were elicited by ending-words with dPdSdM, which indicates a good semantic priming effect also exists in cognitive processing of Chinese idioms.
According to the guidelines of event-related potential recording and published standards suggested by the International Organization of Psychophysiology in 2000, the authors set down the standard N400 testing and analyzing procedure of Chinese idioms. The participants in this study were all healthy adults, with good cognitive function and education. The results showed that N400 could be elicited clearly by the stimuli of congruent and incongruent ending-words of Chinese idioms, and the N400 in this study was stable. The placement of electrodes in this study was based on the international 10/20 system. N400 was recorded from participants' frontal, temporal, parietal, and occipital sites over scalp, and the electrode location should be chosen carefully in analysis if N400 is to be used in psychiatric measurement.2 We suggest that Cz, Fz and Pz be the analyzing sites in psychiatric patients, considering the amplitudes of N400 in these sites were greater and clearer than in other sites.
These results are different from the N400 elicited by the English stimuli,2,6,12 but much like the N400 elicited by picture stimuli. Such a phenomenon could be explained by the difference of the procedure stimuli presented in the studies, but more likely, the phenomenon was the result of the cognitive differences between Chinese characters and phonetic words themselves. Researchers generally believe that N400 reflects a violation of semantic expectations, and N400 amplitude is, in fact, an inverse function of the semantic congruence between the target word and the priming effect afforded by its prior context. Our findings suggest that the violation of semantic expectations in language cognition is different between Chinese and English.
In summary, our findings prove that N400 is an ERP component closely related to the language processing of Chinese characters. It should be noted that the cognitive processing of Chinese characters is not merely the processing from the character's phoneme to its meaning.9,20,21 Obviously, in the cognitive course of Chinese characters, repeated processing and reprocessing of phoneme, shape, and meaning exist. The cognitive processing of phoneme, shape and meaning of Chinese characters can be overlapped temporally, and the cognitive course of phoneme, shape and meaning can hardly be divided. Several issues about Chinese cognition deserve further research.
1. Kutas M, Hillyard SA. Reading senseless sentences: brain potentials reflects semantic incongruity. Science 1980; 207: 203-205.
2. Zhang MD, Chen XS. Evoked brain potentials, 2nd ed. Shanghai: Shanghai Scientific & Technology Publishing; 2000: 311-333.
3. Kiang M, Kutas M. Association of schizotypy with semantic processing differences: an event-related brain potential study. Schizophr Res 2005; 77: 329-342.
4. Olichney JM, Riggins BR, Hillert DG, Nowacki R, Tecoma E, Kutas M, et al. Reduced sensitivity of the N400 and late positive component to semantic congruity and word repetition in left temporal lobe epilepsy. Clin Electroencephalogr 2002; 33: 111-118.
5. Castle PC, Van-Toller S, Milligan GJ. The effect of odour priming on cortical EEG and visual ERP responses. Int J Psychophysiol 2000; 36: 123-131.
6. Wei JH, Luo YJ. Tutorial of cognitive events related potentials. Beijing: Economic Times Press; 2002: 220-228.
7. Franklin MS, Dien J, Neely JH, Huber E, Waterson LD. Semantic priming modulates the N400, N300, and N400RP. Clin Neurophysiol 2007; 118: 1053-1068.
8. Wei JH, Yan KL. Foundation of cognitive neuroscience. Beijing: People's Education Press; 2008: 31-61.
9. Yang WJ. Neural electrophysiology of higher function of brain. Beijing: China Science and Technology Press; 1998: 102-143.
10. Debruille JB, Kumar N, Saheb D, Chintoh A, Gharghi D, Lionnet C, et al. Delusions and processing of discrepant information: an event-related brain potential study. Schizophr Res 2007; 89: 261-277.
11. Chen XS, Lu YZ, Wang JJ, Wang HX, Zhang MD, Lou FY, et al. Relationship between event-related potential P300 and first episode shizophrenia. Chin Med J 2007; 120: 339-341.
12. Chen X, Yang WJ. Brief review of N400 research of Chinese and western language (1). J Clin Electroneurophysiol (Chin) 2004; 13: 53-55.
13. Wang L, Hagoort P, Yang Y. Semantic illusion depends on information structure: ERP evidence. Brain Res 2009; 1282: 50-56.
14. Jiang X, Zhou X. Processing different levels of syntactic hierarchy: an ERP study on Chinese. Neuropsychologia 2009; 47: 1282-1293.
15. Li X, Yang Y, Hagoort P. Pitch accent and lexical tone processing in Chinese discourse comprehension: an ERP study. Brain Res 2008; 1222: 192-200.
16. Zhang Y, Zhang J. Brain responses to agreement violations of Chinese grammatical aspect. Neuroreport 2008; 19: 1039-1043.
17. Meng X, Jian J, Shu H, Tian X, Zhou X. ERP correlates of the development of orthographical and phonological processing during Chinese sentence reading. Brain Res 2008; 1219: 91-102.
18. Zhou S, Zhou W, Chen X. Spatiotemporal analysis of ERP during Chinese idiom comprehension. Brain Topogr 2004; 17: 27-37.
19. Dong RG, Sun XR, Gao SR. An ERP Study on classification of the shape, the phoneme and the meaning of Chinese character. J Clin Electroencephalogr (Chin) 1996; 5: 195-198.
20. Yu BL, Cao HQ. A new exploration on the effect of stroke-number in the identification of Chinese characters. Acta Pharmacol Sin (Chin) 1992; 24: 120-126.
21. Yu BL, Cao HQ, Feng L, Li WL. Effect of morphological and phonetic whole perception of Chinese characters on the perception of components. Acta Psychol Sin (Chin) 1990; 22: 232-239.
22. Li WB, Zhang T, Song LP, Yang J, Feng H. Chinese word processing in nonfluent aphasic patients. Chin Med J 2009; 122: 1901-1906.