Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Local Predictive Context Enhances and Speeds-Up Priming-Induced N400 Effect
Poster Session C, Saturday, September 13, 11:00 am - 12:30 pm, Field House
Anna Petukhova1, Florian Hintz1, Benjamin Straube1, Yifei He1; 1Philipps University Marburg
Imagine reading a shopping list: words such as apple, banana, and pear immediately create a strong sense of coherence, forming a predictive semantic context in which the next word orange fits naturally. In contrast, when the list begins with apple immediately followed by orange, the contextual support is weaker despite the identical semantic relatedness between the two words. This raises the question: How does local semantic context influence the processing of upcoming words, and how is it reflected in neural activity? Here, we conducted a priming experiment where participants (n = 31) performed a lexical decision task on targets that could be either real or pseudowords and were either related or unrelated to preceding primes. Each target was preceded by either one (short condition) or three primes of the same semantic category (long condition), while participants’ brain activity was recorded using EEG. Our analysis focused on the N400 component elicited by the real word targets that were either semantically related or unrelated to the primes. Notably, while the priming-induced N400 effect is often associated with semantic integration, we hypothesised that our condition-based (long vs. short) N400 difference may additionally engage predictive mechanisms. Specifically, we propose that richer semantic context in the long condition may facilitate prediction of upcoming targets, influencing the N400 response. Given that cognitive abilities, such as working memory, processing speed or vocabulary size, may influence predictive processes, we also examined whether individual differences in these skills modulate the effect. Our results provided several insights. First, in the short condition, we surprisingly found no priming-induced N400 effect for unrelated vs. related targets, whereas the long condition showed a robust effect. Although increased priming could naturally boost activation levels through spreading activation, the sharp contrast between conditions is also consistent with the involvement of predictive mechanisms when a stronger context is available. Second, the N400 effect in the long condition had an earlier onset (approximately 200 ms), suggesting that richer semantic context may facilitate faster access to word meaning. This pattern aligns with findings by Nieuwland et al. (2020), who reported that more predictable words elicited reduced N400 amplitudes starting as early as 200 ms after word onset. Finally, we observed a negative correlation between working memory tests scores and the N400 difference, with higher scores linked to a larger difference, suggesting that working memory capacity could benefit the employment of predictive mechanisms during semantic processing. Together, our findings provide supportive evidence that a coherent local context facilitates semantic processing. The enhanced target pre-activation can be explained by both spreading activation and predictive processing accounts. Moreover, the correlation between working memory and the N400 amplitude highlights how individual cognitive capacities can possibly hint at the nature of the mechanism behind semantic processing.
Topic Areas: Meaning: Lexical Semantics, Computational Approaches