Retrieval (N400) and integration (P600) in expectation-based comprehension

Abstract

Expectation-based theories of language processing, such as Surprisal theory, are supported by evidence of anticipation effects in both behavioural and neurophysiological measures. Online measures of language processing, however, are known to be influenced by factors such as lexical association that are distinct from-but often confounded with-expectancy. An open question therefore is whether a specific locus of expectancy related effects can be established in neural and behavioral processing correlates. We address this question in an event-related potential experiment and a self-paced reading experiment that independently cross expectancy and lexical association in a context manipulation design. We find that event-related potentials reveal that the N400 is sensitive to both expectancy and lexical association, while the P600 is modulated only by expectancy. Reading times, in turn, reveal effects of both association and expectancy in the first spillover region, followed by effects of expectancy alone in the second spillover region. These findings are consistent with the Retrieval-Integration account of language comprehension, according to which lexical retrieval (N400) is facilitated for words that are both expected and associated, whereas integration difficulty (P600) will be greater for unexpected words alone. Further, an exploratory analysis suggests that the P600 is not merely sensitive to expectancy violations, but rather, that there is a continuous relation. Taken together, these results suggest that the P600, like reading times, may reflect a meaning-centric notion of Surprisal in language comprehension.

Fig 1. Grand-average ERPs.

Grand-average ERPs on three midline electrodes in the four conditions crossing adverbial clause association and expectancy. Negative voltages are plotted upwards. Ribbons indicate standard error computed from the per-subject per-condition averages.

Fig 2. Scalp distributions.

Topographic distributions of the average potentials in the N400 (row 1) and P600 time windows (row 2), relative to the baseline condition (columns 1-3) or relative to the unexpected-associated condition (column 4). Topographies computed from all non-reference electrodes.

Fig 3. Residual error: Cloze.

Residual error between observed voltages and estimated voltages in Conditions A and C using raw Cloze (left) or log(Cloze) (right) as predictor. Larger deviations from zero indicate larger model error. Ribbons indicate standard error computed from the per-subject per-condition averages.

Fig 4. Residual error: Association.

Residual error between observed voltages and estimated voltages in Conditions C and D using noun-target (left) or verb-target association (right) as predictor. Larger deviations from zero indicate larger model error. Ribbons indicate standard error computed from the per-subject per-condition averages.

Fig 5. Estimated ERPs and residual error.

Estimated ERP waveforms (left) and residual error (right) computed from lmerERP models with log(Cloze) and noun-target association as predictor. Ribbons indicate standard error computed from the per-subject per-condition averages.

Fig 6. ERP coefficients and z-values.

Coefficients (left; added to their intercept), effect sizes (z-values) and corrected p-values (right) from the lmerERP model with log(Cloze) and noun-target association as predictors. Ribbons indicate the standard error on the coefficients from the statistical model.

Fig 7. Exploratory analysis.

Coefficients (left; added to their intercept) and estimated ERPs (right) for exploratory LMER models fitted only on Condition A. Error bars indicate the standard error on the coefficients from the statistical model (right) and standard error computed from the per-subject per-value averages (right).

Fig 8. Reading times.

Log Reading Times per condition on the pre-critical, critical, spillover, and post-spillover region. Error bars indicate standard error computed from the per-subject per-condition averages.

Fig 9. Estimated RTs and residual error.

Estimated log-Reading Times (left) and residual error (right) per condition on the pre-critical, critical, spillover, and post-spillover region. Error bars indicate standard errors on the condition means.

Fig 10. RT coefficients and z-values.

Coefficicents (left, added to their intercept), effect sizes (z-values) and p-values (right) for each predictor on the pre-critical, critical, spillover, and post-spillover region. Error bars indicate the standard error on the coefficients from the statistical model.

Fig 11. Exploratory RT analysis.

Coefficients (left; added to their intercept) and estimated log-RTs (right) for exploratory LMER models fitted only on Condition A. Error bars indicate the standard error on the coefficients from the statistical model (right) and standard error computed from the per-subject per-value averages.