Jonas Fisch

• Excitatory and inhibitory synapses are major components of interneuronal information processing. A robust interplay between these two types is crucial in the lateral superior olive (LSO). LSO neurons integrate excitatory and inhibitory inputs from the ipsilateral and contralateral ear, respectively. They are remarkably sensitive to interaural level differences (ILDs) and interaural time differences (ITDs), making them well-suited for azimuthal sound source localization. Together with their input neurons from the cochlear nucleus and the medial nucleus of the trapezoid body, LSO neurons form an optimal circuit to investigate the mechanisms underlying behaviorally relevant synaptic integration of excitation and inhibition. The mechanisms by which the relative strength and timing of excitatory and inhibitory inputs are integrated in the LSO are not yet fully understood. In this study, I examined synaptic integration in LSO neurons in four steps. First, I evaluated excitatory transmission by combining whole-cell voltage-clamp recordings with electrical fiber stimulations in young adult mice. LSO neurons receive up to 40 excitatory inputs, with each input contributing a synaptic weight (SWexc) of 1 nS/input. I used these values, along with four inhibitory inputs of SWinh of 8 nS/input, as canonical synaptic inputs in the LSO circuit. Second, I analyzed the biophysical membrane properties of LSO neurons that underlie synaptic integration. Fast signal integration is enabled by a short membrane time constant and low input resistance along the putative tonotopic axis. Voltage sag behavior and membrane resonances indicate an extension of passive features by active conductances. Suprathreshold responses were observed as biphasic action potentials (APs), likely reflecting compartment-specific activation of voltage-gated sodium channels. Moreover, the AP threshold varied with the slope of stimulation. Third, I employed the dynamic-clamp in conjunction with an input pathway model to study synaptic integration. Upon sinusoidal stimulation, the neurons behaved as coincidence detectors. They transformed a primary-like input pattern with physiological SWexc into onset responses. The onset can be employed for rate-level coding in combination with inhibition. Conversely, the sustained phase does not support rate-based level difference coding, regardless of SWexc and SWinh configurations. The coincidence mechanism creates band-pass filtering during sinusoidal stimulation and sets the highest sensitivity to transient activations. These activations enable robust rate-difference coding and coding of temporal disparities between excitation and inhibition. Fourth, in vivo data from anaesthetized mice showed that LSO neurons are highly sensitive to variations in stimulus transience and can encode these up to high modulation rates. Moreover, integration of ILDs is not impeded by such high modulation rates. Together, my results provide a mechanistic rationale for the synaptic integration underlying the high perceptual acuity of sound source localization with transient sound signals.