Response Properties of Neighboring Neurons in the Auditory Midbrain

The inferior colliculus, the primary nucleus in the mammalian auditory midbrain, occupies a central position in the ascending auditory pathway. Nearly all ascending neural pathways converge and synapse in the central nucleus of the inferior colliculus (ICC). Further, the anatomical arrangement of axons and neurons in the ICC suggests the existence of functional regions which may play a role in organizing different types of physiological information. To investigate this organization, we characterized the response properties of neighboring neurons in the ICC. To record reliably from neighboring neurons, we adopted a relatively new electrophysiological technique, tetrode recordings. Tetrodes have four closely spaced recording sites (<20ptm) which record multi-unit activity from a small number of neighboring neurons. The recorded signals contain action potentials originating from more than one neuron. Based on action potential wave shape differences across the four channels, we can reconstruct the contributions of individual neurons. Applying tetrode recordings to the ICC of anesthetized cats, we successfully reconstructed individual spike trains for 190 neurons at 52 recording sites. To quantify the advantage of tetrodes, we compared our multi-channel recording results with waveform sorting from single-channel electrode recordings. At best, only 32% of the single-units from tetrode sorting were correctly identified using single-channel recordings. We used tetrode to characterize pure tone responses of neighboring neurons in the ICC in terms of frequency selectivity, level dependence, temporal discharge patterns, and sensitivity to interaural time differences. We find similarities in best frequency and puretone threshold among neighboring neurons; however, we find large disparities in bandwidth, level dependence, temporal discharge patterns, and sensitivity to interaural time differences. These results suggest that neighboring neurons in ICC can greatly differ in membrane properties and/or their patterns of synaptic input from different brainstem nuclei and tonotopic regions. Using tetrode recordings, we investigated how well multi-unit responses represent the response properties of the contributing single-unit responses. We find that multi-unit