Event-related Potentials Evoked By Logic Anomalies

Since ERP technique is considered to provide reliable analysis of processing human language in human brain, previous studies of language and human brain, utilizing ERP, have focused on mainly three components: N400 reported to be elicited by semantically related stimuli, ELAN evoked by first stage of syntactic processes, P600 correlated with the second stage of syntactic processes. However, very few studies have suggested language processing of logically anomalous sentences. Since the study of logic is boosted by the use of formal languages with precisely outlined rules of syntax and semantics, the present study utilizes ERP technique, assessing N400 and late positive component (LPC) to evaluate and provide information on the language processing of logic anomalies. Objective: The present study adopting purely logically anomalous sentences without semantic contamination, and purely semantically incorrect sentences, conducts the comparison study between 30 pairs of logic anomalies and logically correct sentences, and comparison study of the other 30 pairs of semantically incongruity and congruity in order to find out the features of logic processing in the cognitive processing as well as the distinctness among the components of ERPs brain activities evoked by logic processing and other levels of language processing, which is semantic processing in current study. Thus, logic anomalies in present study refer to those sentences whose two propositions are logically contradictory. For instance, the main sentence is with truth value P, whereas the subordinate sentence is with truth value–P, which also means the truth values of atomic statements are true (1), and false (0) respectively, but the truth value of the complex statement is false (0). Method: 20 native Chinese subjects ranging in age from 19 to 26 years (M=21.4375) are studied on a paradigm in which logically incorrect sentences (30 sentences), logically correct sentences (30 sentences), semantically incorrect sentences (30 sentences), semantically correct sentences (30 sentences) and 120 filler sentences are displayed phrase by phrase on the computer screen. 64 electrodes of ERP data are recorded and analyzed by descriptive statistics and ANOVA in SPSS. Amplitudes and latencies of early ERPs and late ERPs are measured by peak-to-peak and peak latency measure, and mean amplitude measures respectively. Greenhouse-Geisser correction is applied to all analyses with more than one degree of freedom in the numerator. Scalp distribution of ERPs elicited by those four types of stimuli is shown in the topography. Results: (1) In the contrast study of logically incorrect and correct sentences, the amplitudes of EPRs evoked by logic anomalies are much larger than those evoked by logically correct sentences. ERPs component LPC (late positive component) elicited by logic anomalies, beginning from 230ms to 430ms without a clear peak, are widely distributed over central and parietal areas. No distinct differences are discovered between the mean amplitudes for LPC in logically anomalous sentences over right and left hemispheres. (2) Consistent with previous findings, deviant words lexically-semantic incorrect sentence structures evokes a widely distributed negativity that started from 250ms and peaked around 400 ms (N400). The amplitudes of N400 elicited by semantically incorrect sentences are larger than those elicited by semantically correct sentences. The negative component is most yielded over frontal area. Conclusions: We found significant abnormalities of LPC in logically incorrect sentences, with potentials showing markedly sensitivity to logic processing. Striking findings are shown that incorrect logic sentences structures are not processed as purely semantic errors, as evidence by the LPC response. Significance: Results indicate that logic anomalies require extra computational operations, which are not required for the processing of semantically incorrect structures. When processing logically incorrect sentences, the system appears to need resources in most areas of the scalp except lateral areas. It appears from this data that incorrectly logic structures cause the neural system to seek for a logically correct solution in other areas of the brain, therefore, increasing the activity in the central and occipital area.