Structure and dynamics that specialize neurons for high-frequency coincidence detection in the barn owl nucleus laminaris

Abstract

AbstractA principal cue for sound source localization is the difference in arrival times of sounds at an animal’s two ears (interaural time difference, ITD). Neurons that process ITDs are specialized to compare the timing of inputs with submillisecond precision. In the barn owl, ITD processing begins in the nucleus laminaris (NL) region of the auditory brainstem. Remarkably, NL neurons are sensitive to ITDs in high-frequency sounds (kilohertz-range). This contrasts with ITD-based sound localization in analogous regions in mammals where ITD-sensitivity is typically restricted to lower-frequency sounds. Guided by previous experiments and modeling studies of tone-evoked responses of NL neurons, we propose NL neurons achieve high-frequency ITD sensitivity if they respond selectively to the small-amplitude, high-frequency fluctuations in their inputs, and remain relatively non-responsive to mean input level. We use a biophysically-based model to study the effects of soma-axon coupling on dynamics and function in NL neurons. First, we show that electrical separation of the soma from the axon region in the neuron enhances high-frequency ITD sensitivity. This soma-axon coupling configuration promotes linear subthreshold dynamics and rapid spike initiation, making the model more responsive to input fluctuations, rather than mean input level. Second, we provide new evidence for the essential role of phasic dynamics for high-frequency neural coincidence detection. Transforming our model to the phasic firing mode further tunes the model to respond selectively to the fluctuating inputs that carry ITD information. Similar structural and dynamical mechanisms specialize mammalian auditory brainstem neurons for ITD-sensitivity, thus our work identifies common principles of ITD-processing and neural coincidence detection across species and for sounds at widely-different frequencies.Author summaryDifferences in the arrival times of sounds at the two ears are essential for creating a sense of auditory space. For many animals, the utility of these interaural time-differences for sound source localization is thought to be restricted to relatively low-frequency sounds, due to limits of temporal precision in the auditory pathway. Barn owls, remarkably, use temporal processing to localize high-frequency (kilohertz-scale) sounds. This capability is critical for their activities as nocturnal predators. Building on insights from previous experimental and modeling studies, we propose that these neurons encode time differences in high-frequency sounds because they respond selectively to input fluctuations, and are relatively non-responsive to input mean. We use a biophysically-based computational model to show that electrical separation between a neuron’s input region (soma) and spike-generating region (axon) improves sensitivity to input fluctuations. This structural configuration produces linear integration of subthreshold inputs and rapid spike initiation, two dynamical features that improve time-difference sensitivity to high-frequency sound-evoked inputs. Neural coincidence detection in the neuron model is further enhanced if it operates in a phasic firing mode. Taken together, we provide new insights into the dynamical and structural mechanisms that support high-frequency sound localization by coincidence detector neurons.