Reliability and Validity of Infant Speech-Sound Discrimination-in-Noise Thresholds

Infants were tested on a speech-sound discrimination-in-noise task using the visual reinforcement infant speech discrimination (VRISD) procedure with an adaptive (up-down) threshold protocol. An adult control group was tested using the same stimuli and apparatus. The speech sounds were synthetic /ba/ and /ga/. The masker was band-passed noise presented continuously at 48 dB SPL. Test-retest reliability was good for both groups, although test-retest differences were smaller for adults. For infants the mean of the absolute values of the differences between tests was only 5.2 dB, and there was less than a 10-dB difference between the two tests of 14 (87.5%) of the 16 infants completing the study. The infant-adult difference in discrimination threshold in noise was 6.9 dB, which agrees well with detection-in-noise thresholds from earlier studies and with discrimination-in-noise thresholds obtained on a subset of subjects in our earlier work. Advantages of the adaptive threshold procedure and its possible applications both in research studies and in the clinic are discussed.

KEY WORDS: auditory development, infant speech discrimination, discrimination-in-noise, adaptive thresholds

Submitted on January 8, 1990
Accepted on July 28, 1990

CiteULike    Connotea    Del.icio.us    Digg    Facebook    Reddit    Technorati    Twitter    What's this?