Many of the most important sounds humans hear, including speech and music, are harmonic, and are in part defined by their fundamental frequency (f0). The perception of f0, often referred to as ?pitch?, has historically been thought to reflect mechanisms that estimate f0 from a sound?s harmonics. But the variety of situations where pitch is utilized raises the possibility of underlying mechanistic diversity. For example, while the absolute f0 of a sound can be important, as when recognizing the gender of a speaker, it is often more important to determine how the f0 has changed over time, as when differentiating a question from a statement. This could be done by estimating the f0 of two sounds and comparing those two estimates, or alternatively, by recognizing some other change between two sounds without first estimating the f0s, such as tracking how the harmonics shift together. In addition to conveying pitch information, harmonic sounds are often important for sound segregation. Harmonic structure can be used to segregate a single sound from a mixture, for example, picking an individual speaker out of a group. Harmonic structure may also aid in detecting sources in other kinds of background noises (such as air vent noise ? noise that does not have harmonic structure). The purpose of this project is to uncover mechanisms used to hear harmonic sounds, both to extract pitch information and to segregate concurrent sounds, by leveraging the differences among individuals and between cultures. We will develop an online crowd-sourcing paradigm to test hundreds of participants on a battery of pitch perception tasks; our hypothesis is that tasks that rely on the same underlying mechanism should correlate. We will also develop a novel task using singing to probe pitch perception, bypassing the explicit judgments traditionally used in psychophysical research. Sung responses provide rich analog measures that can be analyzed for both relative and absolute pitch information. Because singing is a natural human behavior, experiments using singing are well suited to cross-cultural settings in which participants are less able or willing to perform more traditional psychophysical tasks. We will compare sung responses of Western musicians and non-musicians, and the Tsimane? (an indigenous population of hunter agriculturalists living in the Bolivian Amazon), in order to test for multiple discrete representations of pitch. Additionally, we will test the same populations on tasks measuring harmonicity based sound segregation. Studying populations with diverse auditory experiences could help better characterize mechanisms for perceiving harmonic sounds; this could lead to an increased understanding of pitch and speech processing deficits in individuals with hearing loss, assisting in the refinement of existing clinical interventions. This fellowship, along with my research and training plan, will provide me with the foundation for a successful career as an independently funded scientist.
Speech and music are arguably the two most important types of sounds humans hear, and the two domains in which the effects of hearing impairments are most acutely evident. The proposed research will uncover mechanisms used by normal listeners to hear speech and music, both to extract pitch information and to segregate concurrent sounds, by leveraging the differences among individuals and between cultures. The results will provide insight into the specific auditory mechanisms that could be impaired in different hearing disorders, and this understanding could lead to improvement and refinement of existing clinical interventions.
