Encoding sound in the cochlea: from receptor potential to afferent discharge

Abstract

Ribbon-class synapses in the ear achieve analog to digital transformation of a continuously graded membrane potential to all-or-none spikes. In mammals, several auditory nerve fibres (ANFs) carry information from each inner hair cell (IHC) to the brain in parallel. Heterogeneity of transmission among synapses contributes to the diversity of ANF sound-response properties. In addition to the place code for sound frequency and the rate code for sound level, there is also a temporal code. In series with cochlear amplification and frequency tuning, neural representation of temporal cues over a broad range of sound levels enables auditory comprehension in noisy multi-speaker settings. The IHC membrane time constant introduces a low-pass filter that attenuates fluctuations of the receptor potential above 1-2 kHz. The ANF spike generator adds a high-pass filter via its depolarization-rate threshold that rejects slow changes in the postsynaptic potential and its phasic response property that ensures one spike per depolarization. Synaptic transmission involves several stochastic subcellular processes between IHC depolarization and ANF spike generation, introducing delay and jitter that limits the speed and precision of spike timing. ANFs spike at a preferred phase of periodic sounds in a process called phase-locking that is limited to frequencies below a few kilohertz by both the IHC receptor potential and the jitter in synaptic transmission. During phase-locking to periodic sounds of increasing intensity, faster and facilitated activation of synaptic transmission and spike generation may be offset by presynaptic depletion of synaptic vesicles, resulting in relatively small changes in response phase. Here we review encoding of spike-timing at cochlear ribbon synapses.

Keywords: first-spike latency; phase-locking; receptor potential; ribbon synapse; synaptic delay; temporal code.

Figure 1.. Sound is Encoded as Spikes at the Cochlear Hair Cell Afferent Synapse.

A) In the organ of Corti, inner hair cells (IHCs) are stimulated by the movement of cochlear fluids, amplified and tuned by sound-evoked activity of the outer hair cells (OHCs). The graded transmembrane potential of the IHC (analog) is encoded as spikes (digital), generated at the nearby heminode on the auditory nerve fiber (ANF). The ANF peripheral axon is myelinated. Spikes are propagated at nodes of Ranvier in the osseous spiral lamina (OSL), across the soma in the spiral ganglion (SG), and along the central axon toward the brain. In the cochlear nucleus (CN), the ANF central axon branches to terminate on several principal cells. Extracellular sharp-electrode recordings of spike trains in ANFs are generally made from axons as they enter the brainstem. PNS: peripheral nervous system; CNS: central nervous system. B) Expansion of boxed area in panel A, showing the basolateral area of one IHC with its afferent ribbon synapses. Each ANF receives excitatory glutamatergic input from a single ribbon synapse (R: ribbon). The ANF dendrites are unmyelinated within the hair cell epithelium where they also receive efferent synaptic input (not shown). C) Expansion of boxed area in panel B, showing one afferent synapse containing the presynaptic ribbon and the voltage gated Ca²⁺ channels in the membrane (Ca, thick orange line) that control release of synaptic vesicles (SV). Postsynaptic AMPA-type glutamate receptors in the membrane (AMPARs, thick green line) are required for the synaptic transmission that mediates hearing. M: mitochondria. D) Upper: a silent period is interrupted by a 50 ms tone burst at 1 kHz or 5 kHz. Lower: Idealized post-stimulus time histograms; representations of average responses of a single ANF to repetitions of a 1 kHz tone (red) or a 5 kHz tone (green), constructed by recording responses to N repetitions of the same 50 ms tone. The spike sequences are aligned at stimulus onset, the observation period is divided into time bins of size T, and spikes are counted in each bin. The mean spike rate per second for each bin is calculated by dividing the number of spikes per bin by N × T. Before and after the tone burst, the spontaneous spike rate is observed. At sound onset, the spike rate increases to a peak rate that rapidly adapts to a steady-state rate primarily due to depletion of synaptic vesicles. The response to the 1 kHz tone is synchronized to the periodicity of the tone (red); the response to the 5 kHz tone is not (green). E) Upper: Approximately 3.5 cycles of a 1 kHz tone. Lower: ANF spike trains for two presentations of the 1 kHz tone demonstrate responses at a preferred phase (i.e., phase-locking). The ANF does not fire on every stimulus cycle due to failures of synaptic transmission (*) or failures of spike generation (+). The vertical dashed lines refer to the same arbitrary phase within each cycle. Imperfect synchronization results from the temporal jitter in synaptic transmission. F) Upper: Approximately 18 cycles of a 5 kHz tone. Lower: ANF spike trains for two presentations of the 5 kHz tone demonstrate apparently random timing of individual spikes relative to the phase of the stimulus. Synaptic transmission fails to evoke a spike (+) if the excitatory postsynaptic potential (EPSP) is too slow or too small (spike train #1) or if the EPSP follows a spike at very short latency while the ANF is refractory (spike train #2). Failures of synaptic transmission occur on every cycle lacking an EPSP or spike, but asterisks are omitted for clarity. The vertical dashed lines refer to the same arbitrary phase for two adjacent periods around each spike. Lack of synchronization results from low-pass filtering of the IHC membrane time constant and from temporal jitter in synaptic transmission.

Figure 2.. Transduction, Membrane Time Constant and Frequency-dependence of the Hair Cell Receptor Potential.

A) Hair cell cartoon showing the path of mechano-electric-transduction current (I_Tr) through the hair bundle to charge the basolateral transmembrane potential (V_m). Circuit diagram of the basolateral membrane with membrane capacitance (C_m) and resistance (R_m) in parallel. B) Normalized transduction current amplitude (pA) as a function of normalized hair bundle displacement. In the resting position (d = 0), I_Tr is non-zero; the “silent current” depolarizes the hair cell at rest to generate the spontaneous ANF spike rate. Displacement in the excitatory direction increases the depolarizing current; displacement in the inhibitory direction decreases the depolarizing current. The dotted blue line is the symmetric current-displacement function observed for OHCs when the hair bundle is bathed in endolymph (low Ca²⁺, high K⁺). The solid blue line is the asymmetric current-displacement function observed when the OHC bundle is bathed in perilymph (high Ca²⁺, low K⁺). In endolymph, IHCs from the apical cochlea may have current-displacement functions similar to the dotted line, while those of IHCs from the basal cochlea are similar to the solid line (Johnson et al., 2011; Johnson, 2015). C) A square pulse of depolarizing current is injected to the hair cell and the membrane potential is recorded. The hair cell membrane time constant (τ_m) limits the rate of change in V_m as current flows first to charge the membrane capacitance (C_m) and then to cross the membrane resistance (R_m). τ_m = R_m*C_m, the time to charge the membrane to 63% of the total voltage change. The total change in transmembrane potential, or depolarization (ΔV_m), is the product of the injected transmembrane current and resistance, ΔV_m = I_inj*R_m. D) In vivo intracellular recordings of IHC receptor potentials in response to tones of different frequencies, as indicated (Russell and Sellick, 1983). The responses contain alternating components (AC) that synchronize to the tone stimulus and steady or ‘direct’ components (DC). At 300 Hz the response is purely AC. As frequency increases, the AC component is filtered out due to the membrane time constant. The DC component is depolarizing due to asymmetry of the IHC current-displacement function. The intensity of the sound was 80 dB SPL for all frequencies. The resting potential of the IHC was −37 mV.

Figure 3.. Latency of Synaptic Transmission and Spike Generation.

A) Cartoon depiction of an experiment in the excised organ of Corti, utilizing paired pre- and post-synaptic whole-cell recordings to measure transmission at the ribbon synapse. A step depolarization (red trace) evokes a presynaptic Ca²⁺ current in the IHC (green trace). When the depolarizing step is repeated three times, the resulting presynaptic Ca²⁺ current is very reproducible (shown only once, for clarity). Bottom traces show the first excitatory postsynaptic current (EPSC) in the ANF in response to three identical depolarizing pulses. Traces are representative of data from a young rat, at room temperature, showing a long synaptic delay of approximately 2 ms, on average (Goutman and Glowatzki, 2011). Due to the stochasticity of presynaptic release, the EPSC latency will vary (not shown) and the EPSC size may range from tens of pA to hundreds of pA in response to identical presynaptic Ca²⁺ current. B) Stochasticity of presynaptic release reflected in the heterogeneity of EPSC peak amplitude (y-axis) and 10–90% rise-time (x-axis). Data is from one postsynaptic recording of a train of spontaneous EPSCs (i.e., no stimulus applied to the presynaptic hair cell). The gray-, orange-, and purple-filled circles mark the parameter space of the three EPSCs, with corresponding color, shown in panel A and simulated below. C) Cartoon depiction of an experiment in the same preparation, utilizing a post-synaptic recording to measure spike generation. Synaptic transmission has been blocked with the glutamate receptor antagonist CNQX. In current-clamp mode, simulated synaptic current is injected (I_inj.) via the recoding pipette (red, which has an access resistance, R_a) directly into the ANF, which was modeled as two RC compartments in series, separated by an axial resistance (R_axial). Upper traces: The simulated synaptic current injections from the parameter space indicated in color in panel B produce the gray, orange, and purple excitatory postsynaptic potentials (EPSPs). Lower traces: Larger EPSPs evoke spikes with briefer latencies. The traces in panels A and C have the same time scale. D) Same experiment as in panel C. Spike-onset latency (ms) as a function of stimulus (I_inj.) peak amplitude (pA). When stimuli were > 300 pA, latencies were < 0.5 ms and temporal jitter (s.d. of latency) was < 20 μs. The gray-, orange-, and purple-filled circles mark the amplitudes of the 3 stimuli; the vertical lines of corresponding color mark the spike-onset latencies for the 3 stimuli. Upper graph is on a linear scale. Lower graph is on a log-log scale. As simulated synaptic currents (x-axis) approach 300 pA, spike-onset latencies (y-axis) become much faster (y = ~ 1/x), large-dashed line. Above 300 pA, spike-onset latencies change relatively little (y = ~ 1/$\surd$x), small-dashed line (Rutherford et al., 2012).

Figure 4.. Phase-locking and Dependence of Response Phase on Sound Level.

A) Paired recording of phase locking from the bullfrog amphibian papilla. Whole cell voltage-clamp recording from the hair cell, which was held at −90 mV and then at −55 mV, a potential close to the in vivo resting potential of a hair cell in silence. A 400 Hz sinewave voltage command was then applied to the hair cell (blue traces). A patch pipette in the cell-attached mode records EPSPs and spikes (black traces). The gray rectangular box area on left is shown in greater time resolution, on right. Note that many EPSPs do not trigger a spike in the amphibian papilla. Spikes occur during a preferred phase of the sine wave (Li et al., 2014). B) Paired recording of EPSCs in the presence of tetrodotoxin to block spikes at the rat IHC-ANF synapse in response to 100 Hz sinusoidal stimulation at two levels (upper). Responses to the larger stimulus (black) are more frequent, due to increase in the presynaptic Ca²⁺ current (not shown), as shown for two individual responses (middle). Response phase is relatively unchanged with the increase in stimulus level, as can be seen more clearly in the average postsynaptic responses, on bottom (Goutman, 2012). C) Period histograms of in vivo spike trains from a high spontaneous-rate ANF of the squirrel monkey, showing the phase of occurrence of spike-times in the period of the 1 kHz tone stimulus. As SPL increased by 10,000-fold from 10 dB to 90 dB the mean spike rate remained relatively unchanged and the peak in the period histogram remained relatively stationary near 180 degrees, or about half-way through the 1 ms period of the tone (Rose et al., 1967).

Figure 5.. Facilitation and Depression May Offset to Maintain Synaptic Latency at Different Stimulus Levels.

A) Paired recording of a ribbon synapse in rat organ of Corti. Top traces show three trains of 2 ms depolarizing steps from the holding potential of −70 mV to −20, −30, or −40 mV. Middle traces show the presynaptic Ca²⁺ current in the IHC. Bottom traces show the EPSC trains in the ANF. At the end of the pulse train, the dashed blue box around the 10 ms pulse to −20 mV is a “Test” to probe for the remaining synaptic vesicle pool. The larger pulse train (black, −20 mV) results in greater depletion, thus, the response to the test pulse is smaller and longer in latency. The smaller pulse train (green) results in less depletion, thus, the response to the test pulse is larger and faster. B) Open symbols: Synaptic latency in response to the “Test” pulse in panel A. Latency increases as a function of the level of the preceding stimulus train. Filled symbols: Synaptic latency decreases as a function of pulse level, in response to the single pulses in panel C. These two effects may offset in the context of an ongoing stimulus (Goutman, 2012). C) In response to 10 ms depolarizing pulses, larger depolarizations evoked larger Ca²⁺ currents and larger EPSCs with briefer latency (black traces). Smaller depolarizations evoked smaller Ca²⁺ currents and smaller EPSC with longer latency (green traces). D) Left: In response to a large stimulus train, release from depleted ribbons produces a delay in response to the test pulse. Right: In response to a smaller stimulus train, release from non-depleted ribbons produces an advance in response timing to the test pulse. E) Left: The larger single pulse will open more presynaptic Ca²⁺ channels, resulting in facilitation which tends toward a phase advance. Right: The smaller single pulse will open fewer channels, tending to delay release.