Dissociating N400 Effects of Prediction from Association in Single-word Contexts

Abstract

When a word is preceded by a supportive context such as a semantically associated word or a strongly constraining sentence frame, the N400 component of the ERP is reduced in amplitude. An ongoing debate is the degree to which this reduction reflects a passive spread of activation across long-term semantic memory representations as opposed to specific predictions about upcoming input. We addressed this question by embedding semantically associated prime-target pairs within an experimental context that encouraged prediction to a greater or lesser degree. The proportion of related items was used to manipulate the predictive validity of the prime for the target while holding semantic association constant. A semantic category probe detection task was used to encourage semantic processing and to preclude the need for a motor response on the trials of interest. A larger N400 reduction to associated targets was observed in the high than the low relatedness proportion condition, consistent with the hypothesis that predictions about upcoming stimuli make a substantial contribution to the N400 effect. We also observed an earlier priming effect (205–240 ms) in the high proportion condition, which may reflect facilitation due to form-based prediction. In sum, the results suggest that predictability modulates N400 amplitude to a greater degree than the semantic content of the context.

The priming of priming: Evidence that the N400 reflects context-dependent post-retrieval word integration in working memory

“…This might suggest that early N400 priming effects may be generally due to lexical pre-activation mechanisms (such as ASA and prediction) whereas later N400 effects after 400 ms are at least partly influenced by post-lexical integration (see [26] for a related notion of staged N400 effects). The observation that both N400 priming effects (in LTM and WM) lasted beyond 500 ms in our study (and many other N400 studies, e.g., [7,27]) is likely related to the specific task. While our study only provides one example of post-lexical integration processes resulting in N400 effects, we believe this is a very common phenomenon that plays a crucial role in everyday sentence processing.…”

“…Secondly, neither the activation levels of a prime target pair in LTM nor the semantic links between them should change just because other words being processed happen to have similar links between them. With long SOAs, McKoon and Ratcliff's contextual list effect could possibly be accounted for by prediction mechanisms (in line with [7]), but their SOA was 250 ms and thus arguably too short for expectancybased priming. In other words, their findings are likely to reflect priming due to postlexical integration.…”

“…To test its assumptions, it should be possible to manipulate the SOA such that either only the N400 or both the N250 and the N400 would be affected. For example, results from [7] (including N250 reductions around 200 ms) should be fully replicable with SOAs of 450 ms (see Supplementary Materials for details on this argument).…”

“…A more recent ERP article by Lau et al [7] extended the 2008 article by clarifying that activation-based N400 effects during initial word retrieval in LTM may also include prediction-based priming. The authors manipulated the proportion of semantically related prime-target pairs (10% versus 50%) in two experimental blocks, where the higher proportion should encourage participants to use strategies to predict the target word.…”

“…(2) In contrast to ASA, the second priming mechanism is based on expectancy (prediction), is more controlled (i.e., subject to strategies), and usually gains influence at longer SOAs while it tends to be absent at short ones. Predicted target words are assumed to be active in working memory (WM) [7], such that this mechanism does not primarily operate within the semantic network in LTM, although the corresponding word nodes in the network may receive an additional (secondary) activation boost beyond ASA. (3) The third, and perhaps least understood, mechanism is semantic matching or semantic integration, which also takes place in WM, but only after both prime and target words have already been accessed (hence the term 'post-lexical integration').…”