CC BY
9
MISMATCH NEGATIVITY GENERATED BY PROCESSING WORDS AND _PSEUDOWORDS: THE INFLUENCE OF LEXICAL CONTEXT_
Memetova Kristina Serverovna, Stankevich Ludmila Nikolaevna, Polyakova Nadezhda Vladimirovna.
Researchers, Saint Petersburg State University, St.Petersburg, Russia
The present study is designed to establish how lexical context influences the MMN latency and amplitude when the pseudowords are presented. The ERPs were recorded by using only pseudowords or a pseudoword and words with different lexical frequencies. We found the generation of different MMN patterns when the same pseudoword was presented in different contexts. The pseudoword presented in context with another pseudowords demonstrated the smaller amplitude and the bigger MMN latency. Whereas the same pseudoword presented in context with words led to the significantly enhanced amplitude and the decreased latency of MMN. It is supposed that the pseudoword presented in context with words is perceived as conceptually different stimulus leading to the significantly enhanced MMN. Moreover, the hypothesis of lexical frequency influence on MMN has been supported. We found that the presentation of a high-frequency word led to the significantly more pronounced MMN response relative to a low-frequency one [1; 2, p. 717; 3, p. 140; 9, p. 427]. We hypothesize that different amounts of activation depend on the words lexical representation strength.
Keywords: ERP, pseudowords, words frequency, MMN.
Introduction
In present work based on Russian language we have studied how linguistical differences between input signals (pseudowords and words) influence the true mismatch negativity brain potential latency and amplitude. The MMN means that it is not the refractoriness differences. The MMN responses relative to a physically identical stimulus were compared between two different scenarios for the MMN analysis: 1) when the stimulus was presented as a deviant in the odd-ball paradigm; 2) when the stimulus was presented as a standard in the odd-ball paradigm; [5, p. 140; 6, p. 177; 8, p. 1465]. Therefore, some acoustic confounds were reduced and the standard-deviant acoustic-phonetic contrast, the critical variable determining the MMN response, was identical in both conditions. The pseudoword had been chosen as the examined sound. We used the multistimulus odd-ball paradigm and presented two conditions: 1) experimental - including words with different frequencies of occurrence in the Russian language and 2) control - including only pseudowords. Thus, integrating a pseudoword in a sequence of acoustical similar high- and low-frequency words in the experimental condition, the lexical context has been supplied. The MMN responses were elicited either by high- or low-frequency deviant items. The experimental design gave us an opportunity to estimate the lexical influence because the MMN responses were elicited either by high- or low-frequency items. All stimuli were matched for their duration, fundamental frequency and peak amplitude. The word frequency was estimated according to the word frequency dictionary of the Russian language. In the control condition we used only acoustical similar pseudowords witch have no lexical representation in a mental lexicon. In this condition the ERP effect was elicited by the sounds from one pseudoword's category. To examine ERP effects further and minimize the acoustic confounds, we created pseudowords differing only by one phoneme in a real word according the Russian language rules.
Method
We used the passive multistimulus odd-ball paradigm witch was designed by R. Naatanen [5, p.140].
The important characteristic of this paradigm is the opportunity to present one standard and several deviant items. As a result the experimental time is shortened but the number of examined stimuli is risen. The aim of this study was to investigate the influence of lexical context on MMN elicited by pseudowords processing.
We calculated MMN values by subtracting the ERPs elicited by the same sound presented as the deviant and the standard stimulus. Firstly, the peak latencies of responses were obtained for each subject and condition. MMN peaks were determined as the highest amplitude of negative polarity at midline electrodes between 100 and 200 ms, when MMN peaks were the most typically reported. The analysis indicated the different mean long latencies windows for two main conditions: experimental and control. For the statistical assessment of physical acoustical confounds, we compared all standards items from the beginning to the point of divergence. For the statistical assessment of results, we performed a repeated measures analyses of variance (ANOVA) with Stimulus Type (three levels: examinee vs 2 contextual items), Condition (standard vs. deviant response), Electrode Position (six levels: F3, Fz, F4, C3, Cz, C4), the Bonferroni correction was applied.
Results
No significant main effects were found for the standards stimulus ^>0.05). This means that the influence of acoustic features was minimal.
In the control condition no significant main effects were found for the MMN responses. This may imply that multiple semantic representations do not exist for this pseudoword-form and the influence of acoustic features was minimal.
In experimental condition no significant main effects were found for the MMN responses elicited by the pseudoword and low-frequency word (p = 0.709) while the presentation of a pseudoword led to the more pronounced MMN response relative to a low-frequency one. The word is occurred very rare in Russian language. Thus, the lexical dissociation between a low-frequency word and a pseudoword was minimal and that contrast was not enough for statistical significant effect in our conditions. However, the response elicited by
high-frequency word demonstrated the significantly greater MMN amplitude (5,16 mV) than the response elicited by a pseudoword (1,95 mV) (p = 0.036).
The results of the study have demonstrated significant main effects (p = 0.006) for the MMN responses elicited by the same pseudoword that was presented in different conditions: in the control condition with pseudowords only the smaller amplitude and the bigger MMN latency was shown. Whereas the same pseudoword presented in a context in the experimental condition with words led to the significantly enhanced amplitude and the decreased latency of MMN. It is supposed that the pseudoword presented in context with words is perceived as conceptually different stimulus leading to the significantly enhanced MMN. Therefore, this may imply that if a strong dissociation between stimuli (like word vs pseudoword), so-called "novel" reaction is shown this led to the significantly enhanced MMN. Furthermore, the results of this study support the idea that if only pseudowords are presented, the reaction time is risen because pseudowords have no lexical representation in the mental lexicon and they are processed like unknown sounds. However, if the pseudoword is presented in a lexical context with real words, the MMN response demonstrates the processing of the mechanism for novel stimuli.
Finally, we found that the high-frequency stimulus led to a significantly more pronounced MMN response than a low- frequency one in the long latencies window. This finding is similar to earlier reports that the enhancement of lexical word frequency lead to the enhancement of MMN amplitude [1; 2, p. 717; 3, p.140; 9, p. 427].
Conclusion
In sum, this results support the hypothesis about the influence of lexical context on MMN. It was shown the generation of different MMN patterns when the same pseudoword was presented in different contexts. The pseudoword presented in a context with another pseudowords demonstrated the smaller amplitude and the bigger MMN latency. Whereas the same pseudoword presented in context with words led to the significantly enhanced amplitude and the decreased latency of MMN in 100-200 ms. It is supposed that the pseudoword presented in context with words is perceived as conceptually different stimulus leading to the significantly enhanced MMN as for processing on the mechanism for novel stimulus .
Moreover, the hypothesis of lexical frequency influence on MMN has been supported. We found that
the presentation of a high-frequency word led to the significantly more pronounced MMN response relative to a low-frequency one. We hypothesize that different amounts of activation depend on the words lexical representation strength.
Acknowledgements
The work was supported by the Russian Foundation for Humanities (project # 15-06-10806)
References
1. Alexandrov A. A., Boricheva D. O., Pulvermuller F., Shtyrov Y. Strength of word-specific neural memory traces assessed electrophysiological^. PLoS One. 2011. V. 6. Doi: 10.1371/journal.pone.0022999.
2. Aleksandrov A.A., Memetova K.S., Stankevich L.N., Uplisova K.O. Lexical frequency influence on MMN: an ERP-study in Russian language. Rossiiskii fiziologicheskii zhurnal imeni I.M. Sechenova / Rossiiskaia akademiia nauk. 2016. V. 102. № 6. P. 717.
3. Aleksandrov A., Memetova K., Stankevich L. Influence of lexical context on MMN in recognition linguistic stimuli: an ERP-study. The proceedings of the 3rd international conference of the European Society for Cognitive and Affective Neuroscience. 2016. P.140
4. Naatanen R., Alho K. Mismatch negativity-the measure for central sound representation accuracy. Audiol. Neurootol. 1997. V. 2. P. 341.
5. Naatanen R., Pakarinen S., Rinne T., Takegata R. The mismatch negativity (MMN): towards the optimal paradigm. Clin. Neurophysiol. 2004. V. 115. P. 140.
6. Pakarinen S., Takegata R., Rinne T., et al. Measurement of extensive auditory discrimination profiles using the mismatch negativity (MMN) of the auditory event-related potential (ERP). Clinical Neurophysiology. 2007. V. 118. P. 177
7. Shtyrov, Y., Pulvermuller, F. Neurophysio-logical evidence of memory traces for words in the human brain. Neuroreport. 2002. V. 13 P. 521.
8. Shtyrov Y., Kujala T., Pulvermuller F. Interactions between Language and Attention Systems: Early Automatic Lexical Processing? J. Cogn. Neuro-sci. 2010. V. 22. P. 1465.
9. Меметова К.С., Александров А.А., Станкевич Л.Н. Влияние речевой частотности и семантического контекста на Т400: ЭЭГ-исследование с использованием разночастотных омонимов русского языка. Седьмая международная конференция по когнитивной науке. Тезисы докладов. Ответственные редакторы: Ю. И. Александров, К. В. Анохин. 2016. С. 427-428.
