Nunn, Kristen
(2018)
The noisy channel model and sentence processing in individuals with simulated broadened auditory filters.
Master's Thesis, University of Pittsburgh.
(Unpublished)
Abstract
Noise is abundant in every day communication. This high prevalence of noise means we need a language processing mechanism that can recover intended meanings when given noisy input. Research suggests that we do this by maintaining uncertainty about linguistic input and interpreting sentences in a way that is unfaithful to the literal syntax (Gibson, Bergen, & Piantadosi, 2013; Levy, 2011; Levy, Bicknell, Slattery, & Rayner, 2009). People with communication disorders like aphasia or hearing loss have an even higher prevalence of noise. Research has shown that both groups show higher degrees of uncertainty than controls (Gibson, Sandberg, Fedorenko, Bergen, & Kiran, 2015; Nunn, 2016, & Warren, Dickey, & Liburd, 2017). The present study aims to examine how different aspects of cochlear hearing loss influence certainty about linguistic information. While having their eyes tracked, 40 individuals were administered the Gibson Task with sound files simulating broadened auditory filters (BAF). The Gibson Task is a forced-choice picture task that requires participants to select which image best represents a sentence they heard. One illustration represents the literal syntax and one represents an alternate interpretation that may be obtained through edits to the literal syntax. Sentences of different structure (double object/prepositional object, active/passive) require different types and amounts of edits to switch between interpretations. Sentences of different plausibility are more or less likely to be interpreted literally. Using data collected by Nunn (2016), comparisons were made between groups with simulated BAF, simulated reduced audibility of high frequency information (LPF), and no hearing loss (NoHL). The BAF and LPF groups were less accurate and showed higher degrees of uncertainty than the NoHL group. The BAF group was more faithful to the literal syntax than the LPF group for the double object/prepositional object condition. When comparing people with aphasia (PWA) to LPF and BAF, BAF outperformed PWA in all structures but the LPF group only outperformed PWA on actives/passives. Finally, groups with high accuracy scores sometimes showed covert signs of uncertainty in eye-tracking data. The variability between groups implies a complex relationship between noise, syntactic structure, and fidelity to a perceived linguistic signal.
Share
| Citation/Export: |
|
| Social Networking: |
|
Details
| Item Type: |
University of Pittsburgh ETD
|
| Status: |
Unpublished |
| Creators/Authors: |
|
| ETD Committee: |
|
| Date: |
25 May 2018 |
| Date Type: |
Publication |
| Defense Date: |
2 March 2018 |
| Approval Date: |
25 May 2018 |
| Submission Date: |
29 March 2018 |
| Access Restriction: |
No restriction; Release the ETD for access worldwide immediately. |
| Number of Pages: |
137 |
| Institution: |
University of Pittsburgh |
| Schools and Programs: |
School of Health and Rehabilitation Sciences > Communication Science and Disorders |
| Degree: |
MS - Master of Science |
| Thesis Type: |
Master's Thesis |
| Refereed: |
Yes |
| Uncontrolled Keywords: |
uncertainty, noisy channel model, rational sentence processing, low pass filter, vocoded, simulated hearing loss, eye-tracking |
| Date Deposited: |
25 May 2018 13:22 |
| Last Modified: |
25 May 2018 13:22 |
| URI: |
http://d-scholarship.pitt.edu/id/eprint/33985 |
Metrics
Monthly Views for the past 3 years
Plum Analytics
Actions (login required)
 |
View Item |
![[feed]](/images/feed-icon-32x32.png){kind=icon}