Related Experiment Videos
Evaluation of maximum-likelihood threshold estimation with tone-in-noise masking
1Centre for Human Communication and Deafness, The University of Manchestre, UK. richard.baker@man.ac.uk
British Journal of Audiology
|April 21, 2001
Summary
Maximum-likelihood estimation (MLE) offers faster auditory threshold testing but increases variability. Optimizing the up-down method provides comparable results to MLE with reduced testing time.
Area of Science:
- Auditory psychophysics
- Signal detection theory
- Acoustics
Background:
- Adaptive psychophysical procedures are crucial for efficient auditory threshold determination.
- Maximum-likelihood estimation (MLE) is a computationally intensive method gaining traction for minimizing test duration.
Purpose of the Study:
- To compare the speed and accuracy of MLE with a standard transformed up-down algorithm.
- To evaluate the impact of different tracking parameters on threshold estimation in normal-hearing listeners.
Main Methods:
- Estimation of auditory detection thresholds for a 2 kHz tone in broadband and notched noise.
- Comparison of a transformed up-down algorithm (2-8 turnarounds) with MLE (70-90% correct tracking).
- Utilized a two-interval forced-choice task with three normal-hearing participants.
Main Results:
- MLE was consistently faster but exhibited higher threshold variability.
- MLE tracking at 70% or 80% correct led to overestimation of auditory sensitivity.
- Reducing up-down algorithm turnarounds significantly decreased trial count, yielding results comparable to MLE at 90% correct.
Conclusions:
- While MLE accelerates threshold measurement, it may compromise accuracy and overestimate sensitivity.
- Optimizing the transformed up-down procedure by reducing turnarounds offers a viable alternative for efficient and accurate auditory threshold estimation.