Listeners have the remarkable ability to disentangle multiple competing sound sequences and organize this mixture into distinct sound sources. A previous study in human listeners has shown that the physical separation between sounds aids in “segregating” between sound sources, whereby sounds located further apart in space are more easily segregated. Furthermore, under anesthetized conditions, animal neurophysiology has been found to parallel conditions in which humans hear one stream or multiple streams. The goal for this dissertation is to evaluate the psychophysics of spatial stream segregation and, in the same species, record neural activity in auditory cortex in the absence of anesthesia. Cats have been used extensively in auditory research due to their well-developed auditory cortex and because they have evolved accurate sound localization ability to support their nocturnal predatory behavior. We developed a novel paradigm testing the spatial resolution of stream segregation in cats to measure psychophysical performance (Chapter 2) and to uncover the spatial cues that are utilized by cats to perform this task. We then implanted chronic electrodes into primary auditory cortex to record single- and multiple-unit neural activity in awake cats (Chapter 3 and Chapter 4). Our findings show that: (1) Cats can segregate streams of broadband sounds with spatial acuity approaching that of humans. In addition, performance was consistently better for high than for low frequencies which is consistent with previous cat physiological results but contrary to human psychophysics. (2) In the absence of anesthesia, neurons in cat cortex exhibit spectral and temporal properties which are not seen in anesthetized preparations but accord with previous observations in unanesthetized marmosets. (3) Lastly, neurons in auditory cortex of awake cats that are not engaged in an overt auditory task exhibit considerably weaker stream segregation than is observed in anesthetized preparations. Although there is little evidence for stream segregation while cats are not engaged in an auditory task, it may be that segregation is influenced by selective attention or that spatial stream segregation is processed in cortical areas beyond A1. Overall, these findings provide insight into auditory mechanisms underlying stream segregation and neural properties of the unanesthetized cortex.

In a complex auditory scene, listeners are capable of disentangling multiple competing sequences of sounds that originate from distinct sources. This process is referred to as “stream segregation”, where each “stream" represents the perception of a sound sequence from a particular source. Spatial separation of sound sources facilitates the recognition of multiple sequences of sounds (i.e., multiple “streams”) as belonging to distinct sources. Several neurophysiological studies in laboratory animals have shown that perceptual streams are represented by distinct mutually-synchronized neural populations in the auditory cortex. However, the mechanisms leading to those cortical responses are unknown. This dissertation explores the neural substrates and mechanisms of spatial stream segregation (“SSS”) at several stages in the ascending auditory pathway.

We recorded in vivo extracellular spike activity from neurons along different stations of the ascending auditory system of the anesthetized rat, from the midbrain, thalamus, and cortex. Several novel observations were made: (1) The rat primary auditory cortex (area A1) was exclusively tuned to the contralateral hemifield and and was level-tolerant across a 30-dB range of sound levels. (2) Level-tolerant contralateral hemifield spatial sensitivity arises independently along the tectal and lemniscal pathways, highlighting two parallel brainstem pathways for spatial hearing. (3) A linear discriminator analysis of cortical spike counts exhibited high spatial acuity for near-midline sounds and poor discrimination for off-midline locations, which is consistent with previous findings describing the rat’s sound localization behavior. (4) Under stimulus conditions at which human listeners report SSS, neural SSS is weak in the central nucleus of the IC (ICC), it appears in the nucleus of the brachium of the IC (BIN) and in about two thirds of neurons in the ventral MGB (MGBv), and is prominent in A1. Cortical SSS reflects the spatial sensitivity of neurons enhanced by forward suppression. (5) GABA receptor blockers showed no change on cortical forward suppression, suggesting that it does not result from GABAergic inhibition but might reflect synaptic depression at the thalamocortical synapse. Overall, these findings provide substantial evidence that auditory streams are increasingly segregated along the ascending auditory pathways that culminate in distinct mutually-synchronized neural populations in the auditory cortex.