CBSU bibliography search
To request a reprint of a CBSU publication, please click here to send us an email (reprints may not be available for all publications)
Localising the average or averaging the localisations? It does not matter for the early left-anterior Mismatch Negativity! (Poster)
HAUK, O., SHTYROV, Y. & PULVERMULLER, F.
NeuroImage 19(2) part 2, S57, June 2003
Year of publication:
Objectives: Syntactic violations within sentences have been found to elicit a negative event-related potential in left-anterior scalp regions as early as ~100ms [1,2,3].  applied minimum estimation to grand-mean ERP data and found a fronto-lateral peak of activation in the left hemisphere. This approach suffers from two shortcomings: 1) It does not provide statistical estimates for the reliability of features of the solution (e.g. peaks); 2) sources that are consistently activated in the same brain regions across subjects, but vary considerably in orientation, would produce a blurred ERP pattern possibly too weak to produce a significant current source in the corresponding source estimate. Here, we tried to remedy these shortcomings by computing group statistics on minimum-norm estimates obtained for each condition in individual subjects.Material and Methods: We presented the words "come" and "comes" either out of linguistic context or in grammatical vs. ungrammatical contexts in a passive oddball paradigm to elicit a left-anterior Mismatch Negativity [cf. 3]. Source current distributions were computed with the minimum L2-norm method for each individual subject and each condition. A spherical head and source model was used. To control for the influence of noise, regularisation parameters were determined for each data set separately using the L-curve method. The grand-mean estimates were projected on a standard brain surface.Results: The root-mean-square of the difference ERP peaked around 120ms after the point of stimulus divergence (fig. 1A). The corresponding grand-mean difference ERP topography shows a broad frontal negativity with a left-lateralised maximum (fig. 1B). The corresponding grand-mean source estimate exhibits its largest peak at leftinferior frontal sites (fig. 2). The source strength for this location was determined for all four conditions in each subject and submitted to ANOVA analysis with the factors CONTEXT (in/out of) and STIMULUS (come/comes). This analysis revealed a significant interaction between both factors. Because the source estimate revealed a bilateral inferior-frontal source, a similar analysis was performed for the right-hemispheric source as well. However, this analysis did not reveal any significant effects. Conclusion: The early electric left-anterior Mismatch Negativity related to grammatical processing is mainly generated by sources in left fronto-lateral cortex. This is consistent with earlier results . There was no evidence of syntactally-related activation in the superior temporal lobe, as was found in previous MEG studies