Profile analysis in the inferior colliculus: physiology and modeling studies

Abstract

Profile-analysis is a strategy for assessing sensitivity to differences in spectral shape [Green, Oxford University Press, 1988]. Despite extensive behavioral data, the neuronal mechanisms underlying profile-analysis remain elusive. To bridge this gap, extracellular recordings were made in awake rabbit inferior colliculus (IC) and simulated using a computation model. The standard stimulus was a log-spaced, n-component, zero-phase, equal-amplitude complex tone, presented diotically. The central component was incremented in the target stimulus. The discharge rate as a function of characteristic frequency (CF) was inferred by shifting the stimulus spectrum past the CF of each neuron. When the increment frequency was near CF, IC neurons had rates that decreased as the increment increased, contradicting a simple energy-based code. In some cases, the differences between rates for target and standard were largest for stimuli with component spacing that yielded the lowest thresholds in listeners. Responses of IC neurons that were excited by amplitude-modulated stimuli were consistent with model predictions for profile-analysis stimuli [Maxwell et al., JASA 147, 3523 (2020), Guest et al., bioRxiv (2023)]; however, the models did not explain responses of other types of IC neurons. Potential factors contributing to this discrepancy, such as off-CF inhibition and chirp selectivity, will be further explored. Support: NIHR01-DC010813