UC Irvine

Mechanisms underlying selective auditory attention currently lack a comprehensive theoretical framework for considering many empirical phenomena. For instance, there exists little scientific vernacular for discussing differences in dichotic and diotic performance under conditions with equivalent information. The following work applies an empirical paradigm that adds noise to the stimulus as a means to study attention to individual components of the stimulus complex. The first part uses a sample discrimination task that presents tones to alternating ears while manipulating parameters such as informativeness, inter-tone interval, and binaural differences in loudness. Only 2 of 9 listeners showed an optimal listening strategy. Most, instead, show a left-ear advantage and are unduly influenced by loud tones that carry little information. The second part examines selective attention to the basic dimensions of sound - pitch, timbre (e.g., roughness), and loudness. Methods were developed to manipulate the stimulus and observe the effect of specific cue. A theoretical model associated with each cue yields a different pattern of decision weights in a discrimination task. Listeners are instructed to attend to target cues and estimated weights quantify the degree of selective attention to that cue. Some listeners matched theoretical predictions, whereas others appear to use a combination of cues, rather than a single cue. I show that there are clear changes to estimated weights that directly reflect perceptual changes to an isolated cue when a second and then a third cue wis introduced. Findings from both sets of studies introduce novel ways to study auditory selective attention and provide initial building blocks for a theoretical framework.