| Sensitivity to the temporal relationships between sounds is crucial for auditory perception. However, little is known about the neural circuitry underlying such relative-timing judgments, or how different training procedures affect learning on perceptual tasks such as these. This dissertation has two aims: (1) to use a behavioral learning paradigm to shed light on the neural circuitry governing auditory relative-timing judgments, and (2) to use performance on auditory relative-timing judgments to compare different training procedures. To address both aims, listener performance was assessed on two types of relative-timing tasks: asynchrony judgments (distinguishing synchronous from asynchronous sound onsets or offsets) and order judgments (distinguishing the order of sound onsets or offsets). To address the first aim, in each of four perceptual-learning experiments, all listeners took a pretest consisting of relative-timing conditions that varied in the task (asynchrony vs. order), temporal position (onset vs. offset), and stimulus frequencies, and 6-10 days later, an identical posttest. Between these tests, trained listeners practiced one of these relative-timing conditions with tones at 0.25 and 4.0 kHz, while controls did not receive this multihour training. When the trained condition was asynchrony onset, order onset, or order offset, the trained listeners learned more than the controls, but only on the trained condition, indicating that the circuitry affected by the training on these conditions is specialized to separately process each relative-timing task, at each temporal position, for a given frequency pair. When the trained condition was asynchrony offset, the trained listeners instead learned more than the controls on the trained condition as well as on the untrained task and temporal position, suggesting that the circuitry affected by this training was less specialized. To address the second aim, in two experiments, listeners were trained and tested on relative-timing conditions with an identification (or single-interval labeling) or a discrimination (or two-interval comparison) procedure. In contrast to previous proposals that identification training is superior, listeners learned equivalently in both experiments. These results may inform the development of better perceptual training regimens for relative-timing skills, as well as contribute to models and physiological explorations of relative-timing mechanisms. |
