Neural indices of discourse coherenc

The goal of the present set of experiments was to examine the influence of semantic and syntactic cues on focusing attention during discourse comprehension. Experiments 1 and 2 employed event-related brain potentials (ERPs) and reading times respectively to examine temporal shifts in discourse. Experiments 3 and 4 used similar methodology and scenarios to examine spatial shifts in discourse. Experiment 6 employed a novel methodology - simultaneous self-paced reading times and ERP methodologies - to examine the influence of syntactic cues, specifically cleft sentence constructions, on focusing attention. This technique, which was validated in Experiment 5, allowed participants to read at a natural and comfortable pace for comprehension. Repeated NP anaphors were employed to examine the accessibility of information presented prior to an attentional shift. Previous research has demonstrated that these anaphors are employed to reference entities that are no longer in working memory. Thus, if semantic and syntactic cues shift attention, repeated NP anaphors should be used to refer to antecedents mentioned prior to a shift but should not be used if the antecedent is still in a readers' focus of attention (i.e., following a continuity). Results of Experiments 1 and 2 demonstrated that integrating temporal shifts into the preceding context incurs a processing cost at the shift itself, as evidenced by the largest N400 evoked to words that signal large temporal shifts and the smallest N400 evoked to words that signal temporal continuities. In addition, these shifts serve to immediately decrease the accessibility of information prior to the shift, demonstrated by a larger amplitude N400 evoked to repeated NPs anaphors following a temporal continuity. Experiments 3 and 4 demonstrated that readers integrate spatial information into a discourse representation only when there is a direct causal link between described events, providing some evidence that causal information may be critical for tracking information along different dimensions. Future work aims to further examine the influence of causality on tracking other dimensions. In addition, a symbolic distance effect was immediately observed at the point of integrating distances into the preceding discourse context; readers had more difficulty evaluating events that occurred in close proximity relative to more distant events. Finally, Experiment 6 demonstrated that syntactic information can also be used to focus a reader's attention on specific entities in the text and that both lexical and contextual factors most likely independently rather than interactively influence the ease of integrating an anaphor into the preceding context.