Send to

Choose Destination
Brain Lang. 2013 Oct;127(1):55-64. doi: 10.1016/j.bandl.2012.08.004. Epub 2012 Oct 2.

Predicting language: MEG evidence for lexical preactivation.

Author information

Sackler Institute for Developmental Psychobiology, Weill Cornell Medical College, NY, USA; New York University, Department of Psychology, NY, USA. Electronic address:


It is widely assumed that prediction plays a substantial role in language processing. However, despite numerous studies demonstrating that contextual information facilitates both syntactic and lexical-semantic processing, there exists no direct evidence pertaining to the neural correlates of the prediction process itself. Using magnetoencephalography (MEG), this study found that brain activity was modulated by whether or not a specific noun could be predicted, given a picture prime. Specifically, before the noun was presented, predictive contexts triggered enhanced activation in left mid-temporal cortex (implicated in lexical access), ventro-medial prefrontal cortex (previously associated with top-down processing), and visual cortex (hypothesized to index the preactivation of predicted form features), successively. This finding suggests that predictive language processing recruits a top-down network where predicted words are activated at different levels of representation, from more 'abstract' lexical-semantic representations in temporal cortex, all the way down to visual word form features. The same brain regions that exhibited enhanced activation for predictive contexts before the onset of the noun showed effects of congruence during the target word. To our knowledge, this study is one of the first to directly investigate the anticipatory stage of predictive language processing.


Language processing; Lexical priming; Lexical–semantic processing; Magnetoencephalography; Prediction; Top-down processing; Visual cortex

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Elsevier Science
Loading ...
Support Center