Files in this item
|(no description provided)|
|Title:||Speech Reception Threshold in Noise|
|Author(s):||Lewis, Harold Donell|
|Department / Program:||Speech and Hearing Science|
|Discipline:||Speech and Hearing Science|
|Degree Granting Institution:||University of Illinois at Urbana-Champaign|
|Subject(s):||Health Sciences, Audiology|
|Abstract:||The audiological literature is replete with studies attesting to the disproportionate difficulty experienced by individuals with certain high-frequency hearing impairments when listening in noise. It has been suggested that these listeners exhibit a signal-to-noise ratio deficit when compared to normal-hearing listeners. The purpose of this investigation was to determine the effects of noise on the speech reception threshold (SRT) of these listeners. Specifically, this study sought to determine, if indeed, noise has more of a deleterious effect on these listeners than their normal-hearing counterparts, as evidenced by a signal-to-noise ratio deficit. The effects of levels, message set size, and methods for determining speech reception threshold were evaluated with signal-to-noise ratio as the dependent variable.
A total of 36 listeners participated in the study. Eighteen of the listeners had normal hearing while the others had sensorineural hearing impairments. Masking functions were determined for all listeners utilizing three methods for determining SRT. These methods included (1) method of single ascent, (2) method of limits, and (3) a multiple alternative, forced-choice procedure. All testing took place in two sessions at approximately one-week intervals.
The data were converted to signal-to-noise ratios and analyzed statistically using a nested three-factor partially hierarchical analysis of variance model (Winer, 1971). Orthogonal comparisons of means were conducted when so indicated. Data analysis produced no statistically significant differences among groups for either the method of single ascent or the method of limits. Slight differences were found between groups when using the forced-choice procedure; however, these differences were considerably smaller than previously reported. When differences were analyzed with respect to their source, it was found that those differences were mainly attributable to differences between the older and the young listeners.
When the effects of message set size were analyzed results revealed improvement in masked speech reception threshold for the young normal-hearing group, as set size was reduced from 16 to 4 alternatives. The data for the other three groups were anomalous, in that threshold either remained the same or got worse. These results suggest that the older listeners regardless of hearing status were not utilizing the potentially beneficial non-auditory cues afforded by a small message set. Based upon these data, it was concluded that message set size was not a relevant parameter of these subjects' listening strategies.
There were no level differences attributed to any of the three measures. This result suggests that presentation levels were not critical as long as the levels did not exceed distortion characteristics of the ear, and attempts are made to insure that the levels represent points on the psychometric function.
It was generally concluded that the existence of a signal-to-noise ratio deficit was not demonstrated for these listeners; however, the results do not preclude its existence under other circumstances. If indeed the deficit does exist, its locus does not appear to be of a peripheral nature. Further, this study suggests that the deficit might be a characteristic of these listeners' hearing in environments with long reverberant times.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1980.
|Date Available in IDEALS:||2014-12-14|
This item appears in the following Collection(s)
Dissertations and Theses - Speech and Hearing Science
Graduate Dissertations and Theses at Illinois
Graduate Theses and Dissertations at Illinois