Neurons in the medial superior olive (MSO) perform one of the most time sensitive computations in the brain. Situated in the mammalian brainstem just one synapse removed from the auditory nerve, MSO neurons encode the location of low-frequency sounds in the horizontal plane by detecting the simultaneous arrival times of auditory inputs driven from the two ears, a process termed ‘binaural coincidence detection’. Given that sounds travel at ∼34 cm ms−1, the physiologically relevant range of interaural time differences (ITDs) is sub-millisecond for humans and smaller animals. Consequently, the reliable encoding of ITDs requires extraordinary temporal precision that must be maintained at sound frequencies approaching ∼2 kHz, the limit of phase locking exhibited by excitatory inputs to the MSO in most mammals.
At first glance, a sensory system so heavily dependent on temporal precision would seem a poor candidate for neuromodulation, especially neuromodulation of presynaptic release. Yet a study in this issue of The Journal of Physiology provides compelling evidence that presynaptic GABAB receptors modulate neurotransmitter release at synapses onto MSO neurons and that such modulation improves the resolution of binaural coincidence detection (Fischl et al. 2012).
GABAB receptors are G-protein-coupled receptors that, when expressed presynaptically, act through a variety of potential mechanisms to inhibit release probability. How might this contribute to ITD detection? An important clue comes from an insightful study by Brenowitz et al. (1998), who asked how GABAB activation affects the reliability of high-frequency synaptic transmission. Recording from neurons in the chick nucleus magnocellularis, avian analogs of the input neurons to the mammalian MSO, the authors stimulated auditory nerve afferents to evoke trains of EPSCs. They observed that while GABAB receptor activation reduced EPSC amplitudes generally, at frequencies above 100 Hz short-term depression was dramatically reduced, rendering later EPSC amplitudes greater than un-modulated controls. Thus, the strength and reliability of high-frequency synaptic transmission was improved, presumably by preserving vesicles for later release. A subsequent study by Kuba et al. (2002) in the avian equivalent of the MSO showed that artificial reductions in release probability could sharpen ITD detection by narrowing the time window for suprathreshold summation of bilateral EPSPs.
Do such mechanisms reside in mammalian MSO?Hassfurth et al. (2010) found functional and anatomical evidence that excitatory and inhibitory inputs express GABAB receptors in MSO neurons. This expression appeared to be developmentally regulated for excitatory inputs, as GABAB-mediated depression of EPSC amplitudes dropped to smaller, but still significant, levels in the 3 weeks following the onset of hearing at around postnatal day 12. In contrast, GABAB-mediated depression of IPSC amplitudes remained consistent across development.
In this issue, Fischl and colleagues provide an elegant examination of the functional implications of GABAB receptor modulation in the MSO (Fischl et al. 2012). Recording from MSO neurons in slices, they confirmed the earlier finding that activation of GABAB receptors depresses individual EPSCs and IPSCs. In contrast to the Hassfurth study, they found that IPSC decay kinetics were slowed by GABAB activation at mature ages and that EPSC depression remained strong across development. Next, in a result reminiscent of the Brenowitz findings, the authors demonstrated that GABAB activation reduced the amount of depression within 50–200 Hz trains of PSCs, yielding trains with more stable amplitudes. Fischl and colleagues then simulated binaural ITDs in the slice using trains of bilateral synaptic stimuli to evoke EPSPs. The firing of MSO neurons was highly sensitive to these simulated ITDs. Upon addition of the GABAB agonist baclofen to the bath, ITD functions narrowed significantly. Conversely, when a GABAB antagonist was added, ITD functions widened. This latter result provides important confirmation that GABA can be released under physiological conditions (albeit in a brain slice), and even more importantly, GABA can achieve concentrations sufficient to modulate ITD detection. As conventional stimulation techniques do not permit separate stimulation of excitatory and inhibitory inputs to the MSO, ionotropic inhibitory receptors were blocked in these experiments. To circumvent this, the authors developed a model incorporating excitatory and inhibitory inputs driven by in vivo-like patterns. Simulations revealed that both the amplitude modulation of EPSCs and IPSCs and the kinetic slowing of IPSCs by GABAB modulation significantly sharpened ITD tuning in MSO neurons. Moreover, this sharpening persisted over a range of simulated sound intensities.
Together, these results provide compelling evidence that GABAB receptors on presynaptic terminals in the MSO afford a mechanism for rapidly and reversibly sharpening the receptive fields of MSO neurons. When this mechanism comes into play is an exciting question for future studies. At present, the source of GABAergic input to the post-hearing MSO remains a mystery (the primary inhibitory inputs to MSO are glycinergic), but two likely possibilities present themselves. GABA release from an auditory source might be linked to sound intensity levels and provide a way to dynamically dampen highly active synapses. Alternatively, GABAergic inputs from a non-auditory centre might link receptive field modulation with attention or other behavioural states, possibly enabling the system to select between detecting weak auditory stimuli and precisely localizing strong ones. Getting to the bottom of this will probably require a combination of careful anatomy and electrophysiology. More broadly, the present results highlight the growing evidence that computations in the auditory brainstem are subject to modulation (e.g. Kotak et al. 2001; Magnusson et al. 2008). Given the highly dynamic nature of auditory stimuli, this is perhaps not as surprising as it once seemed. It does hint, however, that a rich repertoire of modulatory mechanisms awaits discovery.