Abstracts Submitted: 510
Number of Users: 710
View Abstracts Submitted
Back to home Page
Title:Comparison of two psychophysical procedures in estimation of Gap Detection Threshold. Introduction: Psycho-acoustical experiments involve measurement of individual’s ability to detect a change in the physical aspects of the stimulus (Grose & Mamo, 2012; Schneider & Hamstra, 1999). Measurement of threshold is an important aspect in psycho-acoustical experiments. In psycho-acoustical experiments, as the level of the change, or the stimulus magnitude, increases, the probability of detecting the stimulus also increases. The function relating the signal strength and probability detecting the stimulus is called as psychometric function(Leek, Hanna, & Marshall, 1992) and the threshold is the magnitude of signal at the centre of the dynamic range of the psychometric function. There are different types of psychometric procedures for threshold estimation among which commonly used is Two-down One-up adaptive procedure.One recent addition to the adaptive psycho-acoustic procedure is maximum likelihood procedure (MLP). There are dearth of studies comparing these two procedures in estimating auditory threshold. Aim of the study: To compare the performance of two-down one-up adaptive procedure and MLP in estimation of auditory threshold Method: In the current study, performance of both the procedures were compared in estimating gap detection threshold (GDT). Eight subjects (3 females, 5 males, ages 20- 55yrs, mean age = 42.1 yrs) participated in this experiment. All the subjects had normal hearing thresholds. Both MLP and two-down one-up adaptive was implemented along with 3-interval 3-alternate forced choice task. Numbers of stimulus presentation trials for MLP were 20, 25 and 30. Responses obtained from each stimulus presentation was fitted to candidate psychometric function and the threshold was estimated as the stimulus magnitude corresponding to 70.07% point on the psychometric function.Thresholds were estimated separately for 20, 25 and 30 trails condition.Two-down one-up procedure was implemented with 12 reversals.Stimulus magnitude was reduced by 10% following two consecutive positive responses and decreased by 10% following a single negative response. Threshold was estimated by geometrically averaging the mid-point of last 8 reversals. Threshold was estimated twice with the time gap of 15 minutes for each condition to assess the test re-test reliability. Results: Separate paired-t tests were performed to investigate the main effect of number of trials on GDT. Analysis revealed that GDT for 20 and 25 trials were significantly different (t7= 2.85, p= 0.02). However, no significant difference was found between 25 and 30 trials (t7= -0.87, p= 0.40). This result suggests that, minimum 25 trails are required to estimate threshold using MLP procedure. Intra-class correlation co-efficient was computed to test the test-retest reliability. Test-retest reliability of MLP procedure for 20 trials (r=-0.06, p=0.55), 25 trials (r=-0.63, p=0.48) and 30 trials (r=--0.53, p=0.33) was poor. However, the test-retest reliability of two-down one-up procedure was good (r=0.60, p=0.03). When comparison was made between two-down one-up and MLP procedure (while ignoring the non-reliable responses) there was no significant difference was found (t7= 0.07, p= 0.93). Conclusion: Though MLP can be used to quickly estimate the threshold, the test-retest reliability of the procedure was poor. However, two-down one-up procedure can be used to reliably estimate the threshold.
© Copyright 2017 All Rights Reserved