Mapping Event-related Brain Potential Microstates to Sentence Endings

We analyzed topography and strength of 20 channel event-related potential maps to sentence endings differing in correctness, verbal vs. nonverbal surface form, priming, and repetition count. Seventeen healthy subjects silently read correct and incorrect versions of simple sentences with predictable color endings, and of more complex sentences with predictable composite word endings. Color endings appeared in verbal and nonverbal form. Measures of map topography (centroids of the positive and negative areas of the average referenced maps) and strength (Global Field Power) were analyzed. Adaptive segmentation distinguished a pre-N400 and a N400 microstate in the N400 time range. Topography differed between these two microstates, between verbal and nonverbal endings, and between correct color, incorrect color, and incorrect noncolor words. All verbal endings evoked left-laterlized negativity and right lateralized positivity in the pre-N400 microstates. Correct verbal endings evoked consistent posterior postivity and anterior negativity with left-lateralized gradient strength suggesting language-specific processing. New, incorrect noncolor words evoked reversed anterior-posterior N400 and pre-N400 map topographies with more anterior positivity and more posterior negativity than correct colors in each subject. Gradient strength and current source density maps also differed from those to correct colors. Strongest gradients were left-posterior in the pre-N400 but anterior in the N400 microstate, consistent with anterior activity contributing to the posterior N400 negativity. Incorrect and correct colors, which were semantically primed and repeated, showed smaller topographic differences and N400 effects with a different topography. These different maps can not arise by modulation of a single pattern of neural activity and show that the N400 time range consists of multiple distinct microstates.