Search Abstracts | Symposia | Slide Sessions | Poster Sessions
A drift-diffusion decomposition of conditions that underlie shallow “good enough” processing of spoken sentences
Poster Session C, Saturday, September 13, 11:00 am - 12:30 pm, Field House
Ryan M. O Leary1, Natalie Omori-Hoffe1, Griffin Dugan1, Arthur Wingfield1; 1Volen National Center for Complex Systems, Brandeis University
Although listeners may have the competence to engage in a word-by-word analysis to build a syntactic representation of a heard sentence, in everyday conversation, listeners may, almost by necessity, process the sentence only to a shallow or “good enough” level to derive its meaning. The possibility has been raised that processing schemata may be flexible, such that under some circumstances, comprehension decisions are more likely to be based on an incomplete analysis where heuristic-based sampling and plausibility are used rather than a full syntactic analysis. We report two experiments in which adult participants were tested on their ability to determine the meaning of auditorily presented sentences. In some cases, the thematic roles had been reversed to produce an implausible version of the sentence using the same set of words as the original sentence. After a sentence had been heard, participants were presented with a visually displayed question that probed their understanding of the meaning of the sentence. A participant providing a plausible, yet syntactically incorrect judgment about the meaning of an implausible sentence was taken as an index of shallow processing. In Experiment 1, 32 participants were given two different orienting instructions while completing the listening task (instructions that emphasized either the speed or accuracy of the comprehension decision). To vary perceptual challenge, spectral clarity was also manipulated such that speech was either heard clearly or degraded in spectral richness. Hierarchical Bayesian drift-diffusion modeling was used to probe latent decision-making processes that influenced the participants’ comprehension decision. Results indicate that orienting instructions that emphasize speed and perceptual challenge both increase the likelihood that the assumed meaning of implausible sentences will be based on plausibility. Of special interest, drift-diffusion modeling revealed a dissociation where orienting instructions selectively influenced the amount of evidence required for the participant to make a comprehension decision (a parameter often associated with frontal regions), while sentence plausibility selectively influenced the rate of evidence accumulation (a parameter often associated with parietal and temporal areas). In Experiment 2, 64 participants were tested on the same materials without the use of vocoding to replicate the main drift-diffusion model results. The comprehension, reaction time, and drift-diffusion results of Experiment 2 replicated the results of Experiment 1. The findings from these experiments suggest that comprehension processes are highly flexible and are modulated by both the demands of a task as well as the clarity of the speech signal. These results are interpreted in terms of an underlying neural decision-making mechanism in which the listeners’ development of the meaning of an utterance is due to the accumulation of evidence from the syntactic and semantic content.
Topic Areas: Speech Perception, Syntax and Combinatorial Semantics