MMN: A tool for studying morphosyntactic processing?

MMN: A tool for studying morphosyntactic processing?

1. Study presents two MMN experiments to investigate whether morphosyntactic features are involved in the representations underlying the morphosyntactic processing

   • Methodology: Reported two MMN experiments in passive auditory oddball paradigm with pairs of french words, pronoun and verb, differing in agreement grammaticality. These two experiments differed in the number of morphosyntactic features producing agreement violations. (person and number or person)
     • Results: observed no effect of grammaticality on the MMN response for the two experiments.
       • Conclusions: study highlights the difficulties encountered in studying morphosyntactic level with the passive auditory oddball paradigm.
   • use of MMN makes it possible to manipulate morphosyntactic features but more specifically to control the physical differences between the grammatically correct sentences and sentences with morphosyntactic violations.

2. MMN originally shown in passive auditory oddball paradigms to be generated by any acoustical change.

   • More recently it has been applied to the study of higher order linguistic levels including the morphosyntactic level in spoken language comprehension.
   • In acoustic domain MMN is an ERP generated by automatic cerebral responses to any change in auditory stimulation.
     • MMN is a negative wave observed at frontocentral electrodes peaking at 110-250 ms from the onset of physical change in the absence or presence of focused attention to the stimuli
       • discovered more recently to also reflect long-term memory traces of language experience at various higher linguistic levels including phonological, lexical, semantic and syntactic levels.
         • (PHONOLOGICAL)- Naatanen et al (1997)- found larger effect for deviant finnish vowel /o/ than for deviant estonian vowel that did not exist in finnish. crucially they observed this result despite the fact that the acoustic difference between the estonian vowel and the standard finnish vowel /e/ was larger than that of the deviant finnish vowel /o/- shows MMN is a promising tool for studying higher level processes in speech perception in which the participant does not focus attention on the stimuli.
           • subsequently found MMN effect after a syllable larger when syllable presented in a deviant word than in a deviant pseudoword. (Pulver-muller et al 2001)
             • Peak latencies of the MMN sources in the left superior temporal areas were correlated with word recognition points (Pulvermuller et al 2006) (LEXICAL)
         • sensitive to SEMANTICS- words referring to face movements activated inferior frontocentral areas more strongly whereas words related to leg movements elicited larger activation in the superior central areas (pulvermuller et al 2006)
         • several MMN studies- MORPHSYNTAX by comparing MMN amplitude after correct sentences to that after sentences with morphosyntactic violations. Four studies reported enhanced MMN amplitudes associated with a violation in pronoun-verb agreement in several languages (finnish, english, french, german) at 100-200 ms after the violation point. (Shtyrov et al, 2003; Pulvermuller and shtyrov, 2003; Brunellierte et al 2007; Hasting et al 2007; Pulvermuler et al, 2008)
           • Menning et al (2005) also found an MMN effect modulated by the case violation 150-200 ms after the onset of the deviant word when studying case violations between a determiner and its noun in german
             • All these findings suggest the MMN is sensitive to morphosyntactic processing and hence allows us to tap into the nature of the morphosyntactic representations.
       • independent of attentional demands