Reversible induction of phantom auditory sensations through simulated unilateral hearing loss

Abstract

Tinnitus, a phantom auditory sensation, is associated with hearing loss in most cases, but it is unclear if hearing loss causes tinnitus. Phantom auditory sensations can be induced in normal hearing listeners when they experience severe auditory deprivation such as confinement in an anechoic chamber, which can be regarded as somewhat analogous to a profound bilateral hearing loss. As this condition is relatively uncommon among tinnitus patients, induction of phantom sounds by a lesser degree of auditory deprivation could advance our understanding of the mechanisms of tinnitus. In this study, we therefore investigated the reporting of phantom sounds after continuous use of an earplug. 18 healthy volunteers with normal hearing wore a silicone earplug continuously in one ear for 7 days. The attenuation provided by the earplugs simulated a mild high-frequency hearing loss, mean attenuation increased from <10 dB at 0.25 kHz to >30 dB at 3 and 4 kHz. 14 out of 18 participants reported phantom sounds during earplug use. 11 participants presented with stable phantom sounds on day 7 and underwent tinnitus spectrum characterization with the earplug still in place. The spectra showed that the phantom sounds were perceived predominantly as high-pitched, corresponding to the frequency range most affected by the earplug. In all cases, the auditory phantom disappeared when the earplug was removed, indicating a causal relation between auditory deprivation and phantom sounds. This relation matches the predictions of our computational model of tinnitus development, which proposes a possible mechanism by which a stabilization of neuronal activity through homeostatic plasticity in the central auditory system could lead to the development of a neuronal correlate of tinnitus when auditory nerve activity is reduced due to the earplug.

Figure 1. Audiograms, earplug attenuation characteristics, and characterization results for the phantom auditory sensations.

a) Mean pure tone hearing thresholds (in dB hearing level) for left (grey line) and right ears (black line) of all participants (n = 18). All error bars denote ± s.e.m. b) Mean attenuation provided by the earplugs (n = 18). c) Mean tinnitus spectrum (TS) rating of the participants who perceived a stable phantom sound on day 7 (n = 11).

Figure 2. A computational model illustrates how attenuation through an earplug could lead to the development of a neural correlate of phantom sounds.

a) Architecture of the model covering auditory nerve (bottom) and cochlear nucleus (middle) with projection neurons (PNs), narrow- (NBIs) and wide band inhibitor neurons (WBIs), 4 frequency channels are shown. Circles denote neurons, black lines excitatory and grey lines inhibitory connections. The strength of inhibition from WBIs and NBIs onto the PNs is determined by the gain factors g_w and g_n. b) Attenuation through an earplug is modelled by shifting AN rate-vs.-intensity functions to higher intensities, two different degrees of attenuation are shown (black line – normal, dark grey line –20 dB attenuation, light grey line –40 dB attenuation). c) The mean AN activity is reduced in proportion to the degree of attenuation. d) Attenuation reduces the mean activity of the principal neurons in the CN stage of the model (grey line). By increasing excitation and decreasing inhibition, homeostatic plasticity is able to restore the mean activity to its healthy target level (black line). e) As a side-effect of activity stabilization through homeostatic plasticity, spontaneous firing rates in the model PNs are increased in dependence upon the degree of attenuation. f) Average hearing thresholds of our participants with the earplug in place. g) After homeostatic plasticity has compensated for the earplug-induced decrease in mean activity, the PNs in the CN stage of the model display a pattern of increased spontaneous activity in the high frequency range.