Mechanical amplification by hair cells in the semicircular canals

Abstract

Sensory hair cells are the essential mechanotransducers of the inner ear, responsible not only for the transduction of sound and motion stimuli but also, remarkably, for nanomechanical amplification of sensory stimuli. Here we show that semicircular canal hair cells generate a mechanical nonlinearity in vivo that increases sensitivity to angular motion by amplification at low stimulus strengths. Sensitivity at high stimulus strengths is linear and shows no evidence of amplification. Results suggest that the mechanical work done by hair cells contributes approximately 97 zJ/cell of amplification per stimulus cycle, improving sensitivity to angular velocity stimuli below approximately 5 degrees /s (0.3-Hz sinusoidal motion). We further show that mechanical amplification can be inhibited by the brain via activation of efferent synaptic contacts on hair cells. The experimental model was the oyster toadfish, Opsanus tau. Physiological manifestation of mechanical amplification and efferent control in a teleost vestibular organ suggests the active motor process in sensory hair cells is ancestral. The biophysical basis of the motor(s) remains hypothetical, but a key discriminating question may involve how changes in somatic electrical impedance evoked by efferent synaptic action alter function of the motor(s).

Fig. 1.

Experimental set-up. (A) Schematic of surgically exposed region of the membranous labyrinth showing location of indentation stimulus, single-unit afferent recording, and fluorescent microbead placement. (B) Bipolar stimulating electrodes placed in the brainstem efferent vestibular nucleus. (C) (Lower) Fluorescent microbeads adhered to the cupula. (Upper) Intensity of a single bead viewed through the transparent membranous labyrinth in the living animal.

Fig. 2.

Afferent discharge and cupula displacement. (A–C) Afferent discharge (A) and cupula motion (B) recorded simultaneously in response to sinusoidal mechanical indentation of the membranous duct (C). The electronic shutter of the CCD camera was open ∼50 ms to collect each image. Image acquisition times (“+” in B and E) define the times when bead positions were measured; straight lines connect the data points to clarify the waveform. (D–F) Afferent discharge and cupula motion for step indentation stimuli showing the excitatory–inhibitory asymmetry adaptation of afferent discharge that was not present in cupula motion (which was nearly symmetric at these stimulus levels) and differences in time constants of afferent vs. cupular adaptation. Note that 1-μm sinusoidal mechanical indentation evokes afferent responses and cupula displacements equivalent to ∼4 °/s angular velocity of the head (31); hence the 20-μm stimuli are equivalent to an angular head velocity of 80 °/s (peak-to-peak). The small periodic oscillations in D and E are the result of an uncorrected respiration movement artifact.

Fig. 3.

Efferent action on afferent discharge. (A–C) Activation of the efferent vestibular system by delivering 200 shocks per second to the brainstem (30) typically increased the average discharge rate (spk/s) of sensitive afferents and reduced the peak-to-peak amplitude of modulation. For sinusoidal stimuli greater than ∼2-μm indentation (equivalent to ∼8°/s angular head rotation), efferent activation had no obvious effect on displacement of the cupula (e.g., B). (D and E) Continuous activation of the efferent system evoked tonic changes in afferent discharge rates far outlasting transient discharge modulations induced by mechanical stimuli (*). Subsequent results shown in Figs. 4 and 5 report changes in cupula motion during tonic efferent activation.

Fig. 4.

Nonlinear micromechanical displacements for stimuli near the threshold of sensation. (A and B) Displacement of the cupula as a function of stimulus level at 0.3 Hz. Cupula motion linearly followed the stimulus at high levels of stimulation. At low stimulus levels cupula motion exhibited increased harmonic distortion and higher gain associated with nonlinearity in response. (C and D) At high mechanical stimulus strengths, electrical activation of the efferent system had no effect on motion of the cupula, but at low stimulus strengths harmonic distortion and nonlinear gain observed in the control condition were eliminated. Insets show direct comparisons of responses in the control (A) and the efferent activated conditions (C) for high (a, linear) and low (b, nonlinear) stimulus strengths.

Fig. 5.

Mechanical compressive nonlinearity and elimination by efferent activation. (A) First plus second harmonic displacement of the cupula as a function of equivalent angular head velocity for sinusoidal stimuli at 0.3 Hz in three animals (indicated by red and blue symbols, by gray symbols, and (in Inset) by green symbols, respectively). A compressive nonlinearity amplifying the response of the cupula was present for cupula displacements below ∼500 nm (red curves, filled squares). The nonlinearity was eliminated by electrical activation of the efferent vestibular system (linear growth with stimulus; blue line, filled circles). (B) Nonlinearity is manifested primarily in the second harmonic illustrated here for the first animal from A. The second harmonic of cupula motion shows an active process that feeds mechanical power into the response (active, solid red line). The large second harmonic was eliminated during efferent activation (blue filled circles). Third harmonics are shown also (open symbols) and were consistent with passive harmonic distortions that grow in linear proportion to the stimulus (passive; solid blue line). See Results and Discussion for further details.

Fig. 6.

Afferent compressive nonlinearity. The first plus second harmonic discharge rate (spk/s) modulation of semicircular canal afferent neurons is shown as a function of stimulus strength for sinusoidal motion at 2 Hz (three animals, 14 neurons). Afferent data were normalized by their responses at 7°/s to allow multiple units to be compared using a single vertical axis. The solid blue line with a slope of 1 (log-log scale) illustrates the response of a linear system in which doubling the stimulus doubles the response. For stimuli below ∼7°/s, most afferent responses fell above the blue line and were more sensitive than would be expected from linear theory. Results for low-strength stimuli exhibited a compressive nonlinearity as demonstrated by a slope significantly <1.