Presentation

Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams

Effect of semantic distance on informational masking in listeners with aphasia and controls

Poster C66 in Poster Session C, Friday, October 7, 10:15 am - 12:00 pm EDT, Millennium Hall

Sarah Villard¹, Isabelle Yap¹, Gerald Kidd, Jr.¹; ¹Boston University

Previous work has shown that persons with aphasia due to stroke (PWA) exhibit more difficulty than age- and hearing-matched controls (AHMC) in understanding target speech when background speech is present (Villard & Kidd, 2019). Additionally, other work has shown that, among other cognitive-linguistic deficits, PWA sometimes confuse semantically related words (e.g., “apple” and “peach”) during receptive language tasks (Butterworth et al, 1984). The primary aim of this study was to examine the effect of semantic distance between target and masker words on the ability of PWA and AHMC to understand target speech under speech-on-speech masking conditions. A secondary aim was to compare the results obtained by in-lab vs. remote testing, in order to learn more about the validity of remote testing paradigms. Thus far, 7 (of 8 anticipated) PWA and 8 AHMC have completed the in-lab version of the experiment. In this experiment, participants listened to simple 3-word target sentences starting with the carrier phrase “Betsy sees…” (e.g., “Betsy sees apples”), while ignoring two simultaneous maskers. Six item categories were included: fruits, clothing, birds, furniture, vegetables, and modes of transportation. The experiment included three conditions: “Same” (masker items were drawn from the same category as the target, e.g., “Karen gives peaches”), “Different” (masker items were drawn from a different category than the target, e.g., “Lucy loves T-shirts”), and “Noise” (maskers consisted of speech-shaped, speech envelope-modulated noise), as well as 4 target-to-masker ratios (TMRs): -10 , -5, 0, and 5 dB. After listening, participants clicked on a picture to report the target item. All participants demonstrated excellent comprehension of target sentences in quiet. A 2 x 3 mixed-model ANOVA examining the effect of group and masking condition on accuracy (at the most challenging TMR, -10 dB, only), revealed that accuracy was significantly higher for AMC than PWA, F(2,13) = 24.29, p < 0.001. There was also a significant main effect of masking condition, F(2,26) = 59.18, p < 0.001; post-hoc analyses confirmed significant differences (p < 0.05) in accuracy between each of the masking conditions (noise maskers > same-category maskers > different-category maskers). Finally, there was a significant interaction effect, F(2,26) = 5.29, p < 0.05. Results showed that PWA performance was worse than that of AHMC in the speech-on-speech masking conditions. Additionally, intelligible speech maskers resulted in lower accuracy than noise makers for both groups, with the majority of errors matching one of the presented masker words, indicating frequent target-masker confusions. Surprisingly, results suggested that an intelligible speech masker from the same semantic category as the target resulted in higher accuracy than a masker from a different semantic category. Additionally, the same experiment was repeated via a remote testing setup, and it was observed that results obtained through remote testing were broadly comparable to results obtained in a laboratory setting, providing support for the validity of remote testing paradigms. Work supported by a New Investigators Research Grant from the American Speech-Language Hearing Foundation (PI: Villard), NIH R01DC004545 (PI: Kidd), and a Boston University Undergraduate Research Opportunities Program (awarded to Yap).

Topic Areas: Disorders: Acquired, Speech Perception