Tinnitus pitch matching by residual inhibition and contrast gain control.

Several authors have addressed the relationship between hearing impairment and
persistent tinnitus. In the majority of cases, patients' tinnitus percept
overlaps the hearing loss frequency region (Tunkel et al., 2014) (Henry 1999)
(Norena, Micheyl, Chéry-Croze, & Collet, 2002). Animal studies have shown that
after a cochlear damage some changes in the response properties of auditory
neurons take place (Noreña & Eggermont, 2003). One of these changes is that
neurons in cortical and subcortical auditory structures suffer an increase of
their spontaneous firing rates after the input from the ear is diminished.
Another change is the temporally synchronous activity of a population of
neurons when compared to control animals. The neural synchrony model of
tinnitus suggests that these changes might be responsible for generating the
tinnitus percept.

There is an increasing interest in sound-based therapies for tinnitus treatment
(e.g., Henry et al. 2008; Hobson et al. 2010; McNeill et al. 2012; Shekhawat et
al. 2013). Some of these treatments are based on applying a background noise to
the hearing loss region, but their effectiveness is still questioned (Davis et
al. 2008 positive) (Vanneste et al. 2013 negative). However, the knowledge
about neurophysiology of tinnitus is constantly growing and some of the
treatments are focused on reversing the changes of maladaptive plasticity,
reporting a benefit for the patients. That is the case of the notched music
listening approach, aiming at reducing the spontaneous activity in neurons by
enhancing lateral inhibition from the frequencies above and below the tinnitus
frequency. (Okamoto et al, 2010; Herraiz et al. 2010; Teismann et al, 2011;
Pantev et al, 2012). Like other approaches, the goal is to stimulate either the
predominant tinnitus frequency or the frequencies around it.

The problem which arises in this context relates to the difficulty of targeting
the frequency region of the tinnitus percept, which is crucial for ensuring the
correct application of these therapies and is not directly observable. The
process of defining the fundamental frequency of the tinnitus percept is known
in the tinnitus field as pitch matching (Henry & Meikle, 2000). The most common
approach consists of a psychoacoustical task where the patient listens to
several pure tones and reports which one is the most similar to the tinnitus
percept. The literature is plenty of different approaches to carry out the
pitch matching, many of them consisting on several steps of choices where the
distance in frequency between the presented tones is narrowed step by step,
just as in the case of the two-interval forced choice (2IFC) or the
forced-choice double staircase (FCDS). (ref) Despite of the amount of available
methods, tinnitus pitch matching is strongly criticized because of its
subjective nature. Moreover, some authors claim these are unreliable cause
repeated pitch matches often vary over 2 to 3 octaves (Penner 1983; Tyler &
Conrad-Armes 1983; Burns 1984; Henry et al. 2004). Variations might be produced
by the reporting difficulties patients have when performing the test or by the
tinnitus changes due to the stimulus (Tyler 2000).

Residual inhibition

Some authors have tried to shed light on other representations of the tinnitus
pitch through different psychoacoustical measurements, such as patients*
audiograms (Schaette & Kempter, 2009) or masking levels. The masking intensity
of tones or noise throughout the entire frequency range have shown to follow
specific patterns that might be related to the tinnitus pitch for some cases
(Feldmann, 1971). Some stimuli used in these masking tests have a particular
outcome: a temporary suppression of the tinnitus percept can happen even after
the cessation of the stimulation. The effect is known as residual inhibition
(RI) and its depth and duration can be measured along frequencies, resulting in
the residual inhibition functions.
It has been shown that RI functions overlap the tinnitus spectrum and the
region of auditory threshold shift (Roberts, Moffat, Baumann, Ward, & Bosnyak,
2008), consistent with the hypothesis of the increased neural synchrony. This
difference between responses in the tinnitus frequency region and outside of it
during residual inhibition has been also observed in the 40-Hz auditory
steady-state response (ASSR) and the event-related potential component N1
(Roberts, Bosnyak, Bruce, Gander, & Paul, 2015). These results suggest that
residual inhibition might be a more reliable marker for tinnitus pitch
matching. One major problem approaching RI is the time-consuming aspect to
measure it. The traditional method consists in measuring the time needed for
the tinnitus to reappear after the cessation of a 30s or 60s noise presentation
at 10 dB above the Minimum Masking Level (MML), meaning the lowest intensity
level required to mask the tinnitus percept. Performing this test with
different carrier frequencies to study the frequency dependence and waiting the
correspondent recovery times between trials might be impracticable.
A new method has recently been designed, measuring the Minimum Residual
Inhibition Level (MRIL) (Fournier & Norena, 2018) for reducing considerably the
testing time by using pulsed acoustic stimulation of fixed duration, thus
facilitating the study of this feature in depth. Although in this study narrow
and wide band noise were used as stimuli, recent evidence has suggested that
amplitude modulated tones might reduce the tinnitus loudness even faster (Tyler
& Stocking, 2014), specially in the frequencies around the tinnitus pitch (Neff
et al., 2017). Therefore, the new method will be used in this project and it
will include these stimuli for greater efficiency.

Contrast gain control

Although several neurophysiological models of tinnitus have been proposed,
there is still no consensus on the specific mechanisms generating this
disorder. One of the most recent proposals is the Central Gain Model of
tinnitus. After noise exposure, the neural spontaneous activity is reduced in
the auditory nerve, but the activity in some higher auditory structures, such
as the auditory cortex (AC), medial geniculate body (MGB) and inferior
colliculus (IC), is increased. The Central Gain Model proposes that tinnitus is
caused by an increase in gain or neural amplification in the central auditory
system, aimed to compensate the loss of sensory input from the cochlea. In this
model, it is important to note the role of *neural noise* in tinnitus
generation. Even in silence, the central auditory system is constantly
receiving input so there is a minimum level of activity defining this silence
or baseline. Central gain enhancement does not only produce an increased
sensitivity to sound-evoked activity, but it also amplifies spontaneous
activity, resulting in the auditory perception of neural noise. Several
behavioral studies with animals suggest a correlation between this central gain
enhancement and tinnitus, and moreover, there is some evidence restricting gain
enhancement to only the region of hearing loss and tinnitus pitch (Brozoski et
al., 2002), which places emphasis on the need for further studies to correlate
neural enhancement to the tinnitus pitch (Auerbach, Rodrigues, & Salvi, 2014).
One of the most important statistical properties of the auditory system for
representing sounds is the spectrotemporal contrast in the auditory
environment. Contrast gain control is the property by which neurons adjust
their sensitivity to changes in sound level in response to contrast of sound
stimuli. As a result, the cortical representation is very robust to the noise
increase and invariant to stimulus contrast, which in addition provides a more
efficient use of the neural dynamic range. There are few existing studies
addressing this feature of the auditory system in ferrets (Rabinowitz,
Willmore, Schnupp, & King, 2011) (Willmore, Cooke, & King, 2014), but there is
still no behavioral research in humans. In tinnitus patients, an abnormal
functioning of the central gain around tinnitus frequencies could be measured
by using behavioral test consisting on contrast of stimuli.

The main objective is to improve the existing techniques for measuring the
tinnitus pitch in patients, by exploring both the frequency dependence of the
residual inhibition and the contrast gain control effect in the auditory
system. Behavioral measures and EEG responses will be related in order to
improve the traditional self-report-based methods for pitch matching and to
provide new evidence for the contribution of maladaptive reorganization in
tinnitus generation and maintenance.

The current study has two parts. Part I will be a series of behavioral
experiments with the ultimate goal of finding correlation between residual
inhibition and the tinnitus pitch, in order to shed light on whether it is a
reliable measure for matching the pitch. Part II will consist on a behavioral
and ERP study on contrast gain control in tinnitus patients, aimed to
investigate the relation between this feature and the tinnitus pitch. Both
parts and the corresponding experiments are described below:

Part I: Residual inhibition

Exp. 1: Tinnitus pitch matching by audio samples and automatic method
Subjects will fill out questionnaires and undergo audiometry testing. In this
first session of this experiment, subjects will perform an active listening to
a list of tinnitus synthesized audio samples, previously labelled and
categorized. Participants will relate their tinnitus percept to one or more of
these audio samples by means of a forced-choice procedure, providing relevant
information about their tinnitus such as timbre and pitch.
A non-supervised automated tinnitus pitch-matching procedure (also known as
bracketing method) will be carried out in a second session. There will be a
training session before this procedure starts, in order to familiarize the
participants with the task. In this session, participants will perform a
forced-choice test using pure tones as stimuli (Vernon, 1983).
Additionally, participants will undergo the regular pitch matching procedure
that is currently used in the KNO clinic for assisting tinnitus patients.
The overall duration of the experiment is about 1 hour and a half. Blinding,
randomization, placebo-control, or cross-over are not applicable in this study
design. Between the two measurements, the participants will be asked to take a
break in. The study will be carried out in the tinnitus lab at the
Otorhinolaryngology Department at UMCG. The participants will be placed in
front of a computer monitor in a sound proof chamber.
Once the participants finish the experiment, a second appointment will be
scheduled within one week for repeating the same experiment, in order to test
the reproducibility of these methods and to check the variability of their
tinnitus.

Exp. 2: Minimum Residual Inhibition Level and tinnitus pitch
In the second experiment, the Minimum Residual Inhibition Level (MRIL) will be
measured for several frequencies by means of stimulus sequence consisting of
amplitude modulated tones at increasing intensity levels. Subjects will be
asked, before, after and between the assessment, to rate the loudness of their
tinnitus. For each carrier frequency, the trial finishes when a complete
suppression of the tinnitus is achieved or the maximum intensity level of the
stimulus is reached. A paradigm similar to Fournier & Norena (2018) will be
used, obtaining the MRIL for several frequencies as it is usually done in
audiometries. This experiment will explore the relationship between patients*
RI measures and tinnitus pitch, expecting to find lower values of RI in the
frequencies around the tinnitus pitch.
The overall duration of this experiment is about 20 or 30 minutes. Blinding,
randomization, placebo-control, or cross-over are not applicable in this study
design.
We will conduct a pilot study in which 10 participants will perform the same
test by using narrow band noise as stimulus. The objective is to compare the RI
curves obtained by both amplitude modulated tones and narrow band noise
stimuli. Same duration will be needed for this pilot study.
The study will be carried out in the tinnitus lab at the Otorhinolaryngology
Department at UMCG. The participants will be placed in front of a computer
monitor in a sound proof chamber.

Part II: Contrast gain control

Exp. 3: Loudness discrimination in background noise
In the third experiment, differences in gain enhancement and contrast gain
control between normal hearing, tinnitus and hearing loss patients will be
investigated.
Participants will perform a psychoacoustic experiment consisting of stimulus
discrimination in masking and no-masking conditions with different intensity
levels by using a forced-choice task. On the neuronal level, discrimination
capacity will be measured by means of ERP responses. The overall duration of
this experiment is about 1 hour.

Neither risks nor benefits are known with participation.