Effect of target versus distracter-tone frequency region and variance on sample discrimination of frequency differences

Abstract

This experiment examined how performance in a complex frequency-discrimination task was influenced by the frequency region and relative frequency variability of target versus distracter tones. In a 2IFC sample-discrimination task, normal-hearing listeners were asked to select the interval containing target tones drawn from the higher of two overlapping Gaussian frequency distributions. The target distributions were placed in one of three frequency regions: ‘‘low’’ (500 Hz), ‘‘middle’’ (1414 Hz), or ‘‘high’’ (4000 Hz). For a particular target region (e.g., low), two distracter tones were then drawn at random and added, one from each of the Gaussian distributions at the two remaining frequency regions (e.g., middle and high). All distributions were equivariant on a logarithmic frequency scale. Based on the CoRE model [R. A. Lutfi and K. A. Doherty, J. Acoust. Soc. Am. 96, 3443–3450 (1994)], it was predicted that tones with greater variance would have a greater influence on performance independent of frequency region. The results indicate that the influence of variance depends on the frequency region of targets versus distractors, with notable individual differences and training effects. [Work supported by NIDCD.]