Michael Stahl

ZEST as a tool for rapid assessment of frequency discrimination

DATE: April 12, 2004

Abstract:

The purpose of this study is to evaluate the ZEST psychophysical procedure (King-Smith et al., 1994) as a tool for rapidly and reliably measuring auditory discrimination thresholds. Thresholds for frequency discrimination were obtained by ZEST in 30 trials with a 2AFC paradigm. Subsequent analysis allowed calculation of thresholds for any number of trials up to 30. The stimuli were 600-ms tones at seven frequencies from 250 to 7000 Hz. Data for six normal listeners obtained with three different initial slopes (beta) of the assumed psychometric function indicate that reliable thresholds can be obtained in 10 to 15 trials. Likewise, simulations indicate that threshold estimates have only a slight percent error, which declines to less than 15% after only nine trials. The simulations also show that this bias can be reduced with only a small increase in the variability if the assumed psychometric function used by ZEST is made steeper than the listener?s actual psychometric function, (i.e., if beta is increased from 1.23 to 6). Altogether these results suggest that ZEST combined with a 2AFC paradigm is a promising candidate for rapid and reliable assessment of discrimination thresholds.

Committee: Prof. Søren Buus (Advisor)
Prof. Mary Florentine
Prof. Dana Brooks