Premium
The costs of being certain: Brain potential evidence for linguistic preactivation in sentence processing
Author(s) -
Freunberger Dominik,
Roehm Dietmar
Publication year - 2017
Publication title -
psychophysiology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.661
H-Index - 156
eISSN - 1469-8986
pISSN - 0048-5772
DOI - 10.1111/psyp.12848
Subject(s) - n400 , adverb , psychology , predictability , sentence , context (archaeology) , comprehension , linguistics , word (group theory) , sentence processing , cognitive psychology , event related potential , natural language processing , cognition , computer science , noun , mathematics , statistics , paleontology , philosophy , neuroscience , biology
Prediction in sentence comprehension is often investigated by measuring the amplitude of the N400 ERP component to words that are more or less predictable from their preceding context. The N400—linked to the activation of word‐associated semantic information—is reduced for words that are predictable, indicating that preactivation can lead to facilitated processing. We addressed the question whether there is measurable neural activity related to the preactivation of linguistic information before input confirms or disconfirms this prediction. We therefore measured ERPs not only to moderately to highly predictable target words, but also to preceding adverbs. Based on two separate cloze pretests, we quantified the impact of the adverb upon the predictability of the subsequent target word. Using linear mixed‐effects analyses, we could show that the N400 amplitude at the target word was inversely related to target cloze value, thus replicating the finding that prediction has a facilitative effect on semantic processing. Crucially, the N400 amplitude at the pretarget adverb was modulated by adverb impact: When adverbs increased the predictability of the following word, the N400 was more negative going. We argue that this effect is related to the preactivation of linguistic information. Our findings indicate that the specification of predictions can lead to additional processes before these predictions are confirmed or disconfirmed and that activation of word‐associated information through prediction is highly comparable to activation through actual input.