Authors:

SHTYROV, Y. & PULVERMULLER, F.

Reference:

Year of publication:

2001

CBU number:

5177

Abstract:

Mismatch negativity (MMN), an automatic acoustic change-detection response, has been shown to indicate existence of memory traces for linguistic information (phonemes, words) in the brain. Here, we have tested whether it might be used for investigating grammatical affixes as well. We recorded MMN elicited by identical phonemic contrast incorporated into two sets of spoken language stimuli: (1) frequent (standard) stimulus being non-inflected verb stem [come] and infrequent (16% deviant) stimulus being an inflected form [comes]; up to the divergence point - where [s] started - the stimuli were identical, (2) the same contrast, but in reversed condition, in which the deviant was the stem form, served as control condition. The stimuli were presented in passive oddball paradigm to native English speakers whose brain activity was simultaneously recorded using 64-channel EEG set-up. Both conditions produced MMN responses peaking at 180-280 ms. However, the MMN to the inflectional affix occurred ~50 ms later than for the non-inflected form in the control condition (p<0.001). Furthermore, we found different scalp distribution of the responses elicited by inflected and non-inflected deviant forms: the functional affix presented as a deviant elicited slightly more posterior and left-lateralized distribution (Affixation x Topography interaction, p<0.01). These effects cannot be attributed to mere acoustic differences, since omissions in deviant stimuli are known to result in delayed MMN, while here we observe a reverse. We suggest that the MMN response topography and latency may reveal cortical distribution and activation dynamics of the memory trace (cell assembly) representing the inflectional affix [-s].