Coding deficits in hidden hearing loss induced by noise: the nature and impacts

Abstract

Hidden hearing refers to the functional deficits in hearing without deterioration in hearing sensitivity. This concept is proposed based upon recent finding of massive noise-induced damage on ribbon synapse between inner hair cells (IHCs) and spiral ganglion neurons (SGNs) in the cochlea without significant permanent threshold shifts (PTS). Presumably, such damage may cause coding deficits in auditory nerve fibers (ANFs). However, such deficits had not been detailed except that a selective loss of ANFs with low spontaneous rate (SR) was reported. In the present study, we investigated the dynamic changes of ribbon synapses and the coding function of ANF single units in one month after a brief noise exposure that caused a massive damage of ribbon synapses but no PTS. The synapse count and functional response measures indicates a large portion of the disrupted synapses were re-connected. This is consistent with the fact that the change of SR distribution due to the initial loss of low SR units is recovered quickly. However, ANF coding deficits were developed later with the re-establishment of the synapses. The deficits were found in both intensity and temporal processing, revealing the nature of synaptopathy in hidden hearing loss.

Figure 1. Noise induced changes in ribbon density (ribbon#/IHC) and click evoked CAP amplitude.

(A) representative images showing ribbon changes by the noise exposure. (B) ribbon#/IHC cochleogram. (C) the changes of overall ribbon density after noise (% was calculated against the value of control). (D) Click-evoked CAP input-output functions. A one-way ANOVAs were performed on the overall ribbon density changes and the CAP amplitude measured at 90 dB SPL. Both showed significant effects of noise. The significance was indicated by the asterisks as the result of post-hoc pairwise tests against the control. ***p < 0.001.

Figure 2. Distribution of threshold-BF (left) and threshold-SR (right) for all the units recorded from the four groups.

The impact of the threshold elevation was seen at 1DPN. At this time, there was a disproportional loss of units with extremely low SR.

Figure 3. Noise-induced SR changes in units with BF >4kHz.

(A) The changes in SR median. (B) The changes in SR distribution. One-way ANOVA of rank shows that a significant increase in SR exists only for those high BF units at 1DPN. *p < 0.05. n ≅ 100 in each group.

Figure 4. Noise-induced changes of driven spike rates.

Significant effects were seen only in low-SR units and only appeared at 1WPN and 1MPN. (A) peak rate, (B) total rate. (C) Representative PSTHs from units obtained from the ctrl, 1DPN and 1MPN groups **p < 0.01, ***p < 0.001. The total units were 200 in each group and low-SR units were ~100.

Figure 5. Noise induced changes on temporal responses of ANF units.

(A) prolongation in peak latency, (B) reduction in peak-sustained spike ratio, (C) cumulative unit distribution of peak latency, and (D) cumulative unit distribution of peak/sustained ratio. **p < 0.01, ***p < 0.001 for the results of post-hoc pairwise comparison against the control.

Figure 6. Impact of noise exposure on the recovery functions of click2 responses.

Upper: high-SR units (A,B), middle: low-SR units (C,D), left: click2 spike rate-ICI functions; and right: click2/click1 ratio-ICI functions. (E) typical examples of PSTH for the spike rates evoked by the 2^nd clicks from 3 units obtained from the ctrl, 1DPN and 1MPN groups respectively. Two-way ANOVAs on the factors of groups (noise-treatment) and ICI showed significant effects of noise in both low- and high-SR units and in both the rate and the ratio. Asterisks indicated the significance in the post-hoc tests against the control within the factor of ICI (***p < 0.001, **p < 0.01). The significance was seen only at 1MPN in the high-SR units, but at both 1WPN and 1MPN in the low-SR units. The number symbols indicated the significance in the post-hoc pairwise tests within factor of IC and between 1MPN and 1WPN (^##p < 0.01, ^#p < 0.05).

Figure 7. TEM images of a whole IHC (A) and serial sections of two ribbon synapses, one on the pillar side (B) and the other on the modiolar side (C).

The larger ribbon volume at modiolar side is resulted from the larger cross-section area and/or the larger number of sections on which the ribbon is seen. The white lines show the measure of PSD angle.

Figure 8. TEM images of ribbon synapses obtained at the 3 time points after the noise exposure.

Swollen post-synaptic terminals were seed clearly in images at 1DPN. The presynaptic ribbons appeared to be larger, distorted and having a hollow center.

Figure 9. Comparisons of ribbon vesicles (A) and PSD angle (B) showing effect of sides (modiolar vs pillar) and noise groups.

“#s” for the within side comparison against control, “*s” for the within group (time point) comparison between sides. ^###/***p < 0.001, ^##/**p < 0.01, ^#/*p < 0.05.