Cognitive control states influence real-time sentence processing as reflected in the P600 ERP

ABSTRACT

As sentences unfold, people often encounter conflicting cues that vie for influence during processing and comprehension. We measured event-related potentials (ERPs) to test the hypothesis that cognitive control increases activation of the most plausible analysis of the input to resolve such conflicts. Participants read sentences like “The bathroom floor was mopping yesterday”, where conflict arises between a syntactically-licensed interpretation (“floor” as Agent of “mop”) and a semantically plausible one (“floor” as Theme of “mop”). Participants also completed pseudorandomly interleaved Congruent and Incongruent Stroop trials, which manipulated cognitive-control engagement immediately before sentence-reading trials. Target sentences elicited a P600 effect, replicating previous results suggesting that readers typically make morphosyntactic edits to yield the plausible interpretation (e.g. “mopping” → “mopped”). Crucially, the P600 was larger and more broadly distributed following Incongruent versus Congruent Stroop items, consistent with increased structural repair activity when an individual’s state of cognitive control is relatively upregulated.

KEYWORDS:

• Language comprehension
• cognitive control
• P600 ERP
• conflict adaptation
• rational inference

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 An alternative approach would be to quantify the P600 across a pre-specified set of electrodes. We decided against this approach because we were testing for a difference in the spatial extent of two P600 effects (post-Congruent and post-Incongruent) and did not have a priori reason to know which electrodes would be part of either P600 pattern. Moreover, it is our experience that the P600 can vary in scalp distribution across contexts (experiments, conditions, samples)—the effect is reliably central-parietal but can appear at a number of electrodes within this wide region—making it difficult to establish a single electrode group that optimally captures the P600 in all contexts.

2 The threshold of t>1.7 identified electrodes at which the one-directional t-test (62 d.f.) was significant at p<0.05.

Additional information

Funding

This work was supported by NSF: [Grant Number 2020490 ] and NSF: [Grant Number 2020932].