Expect the Unexpected: Event-related Brain Response to Morphosyntactic Violations

TLDR

Recent debates on language‑specific neural systems question whether syntactic and semantic processing are independent, with event‑related potentials such as the P600 for syntactic violations and the N400 for semantic expectancy showing distinct patterns, yet the P600’s scalp distribution resembles the non‑linguistic P3b component. The study examined how the P600 relates to the P3b by manipulating sentence grammaticality, odd‑ball probability (via varying violation frequency), and event saliency (using two morphosyntactic violation types of differing strikingness). Results showed that P600 amplitude, like P3b, was modulated by odd‑ball probability and event saliency, with similar scalp distributions for probability and grammaticality effects, and an anterior negativity between 300–500 ms post‑word onset likely reflecting working‑memory processes.

Abstract

Abstract Arguments about the existence of language-specific neural systems and particularly arguments about the independence of syntactic and semantic processing have recently focused on differences between the event-related potentials (ERPs) elicited by violations of syntactic structure (e.g. the P600) and those elicited by violations of semantic expectancy (e.g. the N400). However, the scalp distribution of the P600 component elicited by syntactic violations appears to resemble that elicited by rare categorical events (ldquo;odd-balls”) in non-linguistic contexts, frequently termed the P3b. The relationship between the P600 and the P3b was explored by manipulating the grammaticality of sentences read for comprehension, as well as two factors known to influence P3b amplitude: odd-ball probability and event saliency. Oddball probability was manipulated by varying the frequency of morphosyntactic violations within blocks of sentences, and event saliency was manipulated by using two types of morphosyntactic violations, one of which was more striking than the other. The results indicate that the amplitude of the P600, like the P3b, was sensitive to both the probability and saliency manipulations, and that the scalp distributions for the effect of probability and grammaticality are essentially similar. An unexpected, but not wholly surprising, finding was the elicitation of an anterior negativity between 300 and 500 msec post-word onset, which may index working memory operations involved in sentence processing.