Auditory System at the Cocktail Party

The "cocktail party" is the classical example of a complex listening environment. Listeners attempt to understand the speech stream from a selected talker amidst a cacophony of speech sounds from other talkers, clinking of glasses, maybe a jazz combo over in the corner, etc. The normal human auditory system accomplishes this feat with impressive accuracy. Similarly, birds attend to particular individuals among other birds singing stereotyped songs, mammalian predators and prey detect or evade each other in the presence of ambient wind and water noise, and frogs identify potential mates in the noisy environment of a breeding chorus. Hearing in a complex auditory scene integrates all the functions of the peripheral and central auditory systems. Major auditory cues for object formation and segregation include fundamental frequency (i.e., corresponding to pitch), spectral or temporal envelope, and location. Processing of these cues relies on spectral and temporal pitch mechanisms, precise spectral representation, analysis of temporal coherence among widespread neural populations, and extraction and integration of monaural and binaural localization cues. There is good evidence for bottom-up auditory object formation and segregation and top-down selection among objects. Many of these individual functions traditionally have been studied in isolation. In recent years, a growing body of research is attempting to explain how these multiple elements are integrated to solve the cocktail party problem.