THE LATE POSITIVE ERP IN LANGUAGE PROCESSING
Maxwell Ruckstuhl, Megan Bardolph, Seana Coulson.
University of California, San Diego, La Jolla, CA.
Electroencephalography (EEG) is a noninvasive measure of brain activity that has proven especially valuable in the study of neural information processing. Event-related potentials (ERPs) are cross-trial averages of EEG signals following a stimulus. ERP parameters, including timing, amplitude, and distribution, have been shown to index discrete components of language processing. By analyzing ERPs across stimulus conditions, we can break down the complex problem of understanding language and glean insight about the brain's activity over time and space. In this study, we examined 2 theories regarding the post-N400 positivity (PNP), an ERP component that has been shown to index both syntactic and semantic violation. We hypothesized that a parietal PNP is caused by the forced reinterpretation of a sentence, including syntactic violation. Independently, a frontal PNP indexes incorrect word prediction including semantic violation. To evaluate these hypotheses, we took EEG measurements at 32 standard scalp locations on 19 subjects who were presented with 290 sentences under 4 conditions to elicit the separate PNPs. We observed a PNP consistent with our prediction of the frontal PNP, but it was distributed across both the frontal and parietal lobes, precluding confirmation of our theory. Improved stimuli that more selectively elicit the 2 PNPs may allow us to tease out the differences. Nevertheless, this finding supports the use of ERPs as a valuable tool in cognitive science to determine how the brain processes information and how it reacts to different unexpected stimuli.