STEROWANA OCENA PROGÓW PSYCHOFIZYCZNYCH

Abstract

ObjectiveThe estimation of perceptive thresholds is a basic element of psychoacoustics. One of the drawbacks of commonly used adaptive algorithms is the lack of reliability when the behavioral response is not robust. To address this issue an adaptive algorithm, TEMA (Threshold Estimation by Managed Algorithm), has been developed.DesignTEMA seeks the 50% point on the psychometric curve based on an up-down staircase procedure. Internal controls and stochastic processes aim at enhancing the reliability. The development of TEMA is described, together with its validations with reference to common adaptive procedures. Both Monte Carlo simulations and real subject testing were performed to assess the psychoacoustic threshold in intonation perception tests and the number of stimulus presentations needed.Study sampleA total of 29 adult listeners participated in the within-subjects comparison; 19 listeners had normal-hearing, the other 10 were hearing impaired (5 aided, 5 unaided).ResultsThe results show that TEMA outperformed commonly used algorithms in non-robust responders, with a minimal cost in terms of test duration.ConclusionsTEMA’s adaptive algorithm was shown to be significantly more resistant to gambling or cheating behavior and threshold drift than traditional, reversal-based algorithms. TEMA increases the accuracy of threshold estimation and the test reliability in non-robust responders. This makes TEMA applicable for automated threshold measurements in clinical settings.