Variables related to speech discrimination in competition for normal hearing subjects

Much of a person's time is spent in the listening form of communication. Informational campaigns by people in business and education have indicated the importance of this skill in school and on the job. In our noise polluted environment, listening rarely takes place in the quiet one-to-one situation found in the audiologist's test booth. Perhaps the present determination of normal hearing with tests of speech discrimination in quiet and pure tone acuity is failing to identify subjects with problems understanding speech in noise.

In the present study the diversity of responses was examined when competing backgrounds were added to the speech discrimination test words heard by normal hearing subjects. Other tests were also given to identify possible measures related to speech discrimination performance in noise conditions. These other measures included: speech discrimination in quiet; pure tone thresholds at 250, 500, 1000 and 4000 Hz; , threshold of octave masking (TOM); three conditions of masking level differences (MLDs); pitch discrimination and tonal memory subtests of the Seashore Test of Musical Talents; and the Embedded Figures Test (EFT).

Following the test session the subjects were given a questionnaire. These subjective responses were divided into variables concerned with the listening experience of the subject (years of musical study, time in Dept. of Speech Pathology-Audiology, etc.), noise experience and predicted performance listening and studying in a variety of everyday situations. The correlations were examined between these variables and the speech discrimination tests.

Fifty students in Speech Pathology-Audiology were asked to repeat 100 monosyllabic NU-6 words given in a background of quiet, white noise and three talker discourse at an S/N ratio of -10 dB. The speech discrimination tests showed good test-retest reliability and increased inter-subject variability in the competing background conditions.

These individual differences on speech discrimination in competition were significantly correlated with performance on the measure of visual figure-ground performance (EFT). Those subjects who were more figure dependent in the visual task tended also to be better able to repeat correctly the target words embedded in a white noise or competing speech background. A factor analysis showed these three figure ground tests loaded on a single factor.

The MLD condition which gave the largest release from masking (SπNπ-S_oNπ) was positively correlated with speech discrimination in white noise. Other auditory tests such as threshold of octave masking, pure tone thresholds, pitch discrimination and tonal memory were generally not significantly related to speech discrimination in competition.

Responses to the questionnaire indicated listener sophistication did not cause improved scores. A self-rating of performance in a variety of listening situations showed that subjective responses may be fruitful in identifying normal hearing subjects who may have difficulty understanding speech in a competing speech background.

Note: Abstract extracted from PDF file via OCR.