UC Irvine

The ability to engage in auditory selective attention relies on listeners discriminating between acoustic features that are simultaneously available within a sound and can differ between auditory objects. Quantifying the qualitative state of these acoustic features has been advanced via COSS analysis. This methodology employs identical stimuli across trials, while directing the listener to attend to one of several acoustic features (e.g. loudness, timbre, or pitch). Weight profiles for loudness and timbre judgments were replicated from previous experiments, while under EEG recordings. While listeners were engaged in making loudness judgments, a characteristic ERP response of the stimulus fundamental frequency (F0) and second harmonic (2F0) was observed. Timbre ERPs displayed a similar response to F0, but a decrease in activity related to 2F0. The dissertation then moves to studying automatic attention. Transition bandwidths are calculated by estimating thresholds at various bandwidths as the number of components in a complex sound is increased at a fixed frequency difference. Thresholds increase with the number of components as the signal to noise ratio is decreased. Eventually a breakpoint is achieved where thresholds decrease as listeners perform profile analysis. This experiment manipulates previous experiments by fixing the phase of each component, instead of randomly sampling each phase independently. Results show a dramatic decrease in threshold estimates associated with profile listening from randomized phase to fixed phase conditions. The physiological basis of this change is investigated on the cortical level. EEG recordings show that F0 power increases with stimulus bandwidth.