A fast method for the determination of psychophysical tuning curves: further refining

Abstract

Psychophysical tuning curves (PTCs) are usually measured by determining the level of a narrowband noise required just to mask a fixed, low-level tone, for several masker centre frequencies. PTCs are often used to assess the frequency selectivity of the auditory system and they have also been used to detect "dead regions" in the cochlea, especially to define the frequency boundaries of the dead regions. However, the traditional method of PTC determination is too-time consuming for use in clinical practice. This paper is concerned with further evaluation and refining of a fast method for determining PTCs, based on the use of a sweeping band of noise. The fixed sinusoidal signal is turned on and off at regular time intervals and is masked by a band of noise, whose centre frequency sweeps over a range of two octaves during four minutes. A Békésy method is used to determine the masker level required for threshold; the subject presses a button to indicate that the signal is inaudible, and releases it when the signal is audible, and the masker level is adjusted accordingly by a computer. The fast method was evaluated using normally hearing subjects and showed good agreement with the results obtained with the use of the traditional method. The shapes of the PTCs, the slopes of the low- and high-frequency skirts, and the positions of the minima were very similar when the fast and the traditional methods were used. However from the point of view of clinical usage the determination of the PTC tip, that is the masker centre frequency at which the masker level is lowest is the most important issue. The position of the PTC minimum with reference to the tone frequency indicates the presence of a dead region. Therefore, in this study several methods of the PTC minimum estimation were evaluated and compared. It has turned out that a fitting method of a single PTC by means of a square function yielded the best results. The method gave the smallest standard deviation, the highest kurtosis and the narrowest range of the PTC minima.