UC Berkeley

The human auditory system easily solves the "cocktail party problem" - that is, even when multiple people are speaking at once, we can easily select and pay attention to a single voice while ignoring the others. Though this seems easy to do, the problem is known to be quite computationally complex. It requires identifying the important sound, selecting it for special processing, and using information from it to make behavioral decisions; meanwhile, the other voices must not be allowed to distract us.

How does the brain do this? In chapter 1, I review previous approaches to this question and motivate the choices I made in designing my experiments. In chapter 2, I present the data and conclusions I obtained in collaboration with my advisor, Dr Michael DeWeese. (We are submitting this chapter for publication separately.) In chapter 3, I present a detailed protocol for repeating our behavioral results.

The final chapter, Chapter 4, is broader in scope. I discuss how our models and results relate to existing models of prefrontal control over other brain regions. Finally, I consider what my results have taught me about the scientific process of investigating neural function and ruminate on where this field may be headed next.