Deﬁcient central mechanisms in tinnitus: Exploring the impact on speech comprehension and executive functions

in individuals with tinnitus. Given these previous ﬁndings, we theorize that deﬁcient central mechanisms may be responsible for the reported speech comprehension problems in tinnitus. 25 participants suﬀering from chronic subjective tinnitus and 25 control participants, between 23 and 58 years of age, were examined in a cross-sectional design. The groups were case-matched for age, sex, education, and hearing loss. A large audiometric battery was used ranging from threshold and supra-threshold tasks to spoken sentence level speech tasks. Additionally, four cognitive tests were performed, primarily covering the area of executive functions. Tinnitometry and tinnitus-related questionnaires were applied to complement sample description and allow for secondary analyses. We hypothesized that tinnitus participants score lower in complex speech comprehension tasks and executive function tasks compared to healthy controls, while no group diﬀerences in elementary audiometric tasks were expected. As expected, individuals with chronic subjective tinnitus scored lower in the SIN and gated speech task, while there were no diﬀerences in the basic speech recognition threshold task and the other elementary auditory perception tasks. The cognitive tests revealed clear deﬁcits in interference control in the Stroop task, but not in the Flanker task, in the tinnitus group. There were no diﬀerences in inhibition or working memory tasks. Our results clearly delineate diﬀerences between tinnitus individuals and control participants in two tests on speech intelligibility under adverse listening conditions. Further, the poorer performance in a task of interference control in individuals with tinnitus points towards an impaired central executive control in individuals with tinnitus. Taken together, our (partly) exploratory study provides novel evidence to the view that deﬁcient central executive system in individuals with tinnitus probably account for impaired speech comprehension.


Introduction
Chronic subjective tinnitus is the constant perception of a sound or noise that is not relatable to an external acoustic source and does not have a signaling or informational character for the person affected (Hallam et al., 1984;Eggermont and Roberts, 2004).De Ridder et al. (2021) proposed a further differentiation of the tinnitus phenomenon, where 'tinnitus' refers to the conscious perception of a tone or noise as just described, and 'tinnitus disorder' describes tinnitus that is accompanied by emotional and/or cognitive dysfunction, autonomic arousal, possibly resulting in functional disability and changes in behavior.The prevalence of tinnitus increases with age, as does hearing loss; however, tinnitus can appear at any age (Humphriss et al., 2016).Depending on the definition of tinnitus and the questionnaires used, between 5 and 43% of the population worldwide are affected (McCormack et al., 2016).Between 1 and 4% of all people suffer severely from tinnitus disorder (Langguth et al., 2013;Biswas et al., 2022).Tinnitus manifests very heterogeneously and is related to many risk factors, e.g., hearing loss, aging, and temporomandibular joint disorder (Beukes et al., 2021;Cederroth et al., 2019).The general consensus in the scientific community is that hearing loss is a major risk factor for tinnitus (Baguley et al., 2013;Cederroth et al., 2013;Nelson and Chen, 2004).However, it's important to note that not everyone with hearing loss experiences tinnitus, and not everyone with tinnitus has measurable hearing loss.Therefore, there are some studies that highlight the complexities in the relationship between tinnitus and hearing loss (Liberman and Liberman, 2015;Schaette and McAlpine, 2011;Guest et al., 2017).
Individuals suffering from chronic tinnitus often complain about impaired and limited spoken language comprehension (Andersson et al., 2001;Sanchez and Stephens, 2000;Tyler and Baker, 1983), in particular in everyday communicative situations, which can impact their normal functioning considerably.Nevertheless, self-reported hearing impairment does not necessarily predict hearing performance or speech-in-noise (SIN) performance (Nabelek et al., 2004;Choi et al., 2016).Beyond that, multiple factors may exert influence on subjective speech comprehension, including demographic, health-related, and social variables (Pronk et al., 2018).A further perspective posits that subjective hearing in noisy environments primarily pertains to evaluative aspects of auditory processing, rather than sensory aspects (Wöstmann et al., 2021).Taken together, it is thus critical to avoid fixating on individual aspects of auditory perception exclusively and consider an all-encompassing view on the interplay of tinnitus, complex speech comprehension, auditory perception, and cognition.Most important, we think that spoken language must not be considered a modular entity, but functions in dense intertwining with other cognitive skills.Hence, we aim at construing a feasible model of the interplay between tinnitus, speech comprehension and general executive functions (EF),2 possibly culminating in a deficient central executive system (Tai and Husain, 2019).To the best of our knowledge, this is the first study which combines methods of comprehensive audiometry, speech comprehension, and cognitive tests, to investigate to what extent speech comprehension may be directly or indirectly affected by tinnitus, which allows for more in-depth insights into the interplay of the three domains.
Our study largely focuses on speech comprehension, therefore it is important to note that hearing and comprehension occur at different levels, from sensory perception to understanding complex spoken language utterances.In addition to the peripheral auditory system and the auditory pathway (i.e., sensorineural hearing loss), components of the central auditory system seem to be affected in chronic subjective tinnitus as well (Eggermont and Roberts, 2004;Jagoda et al., 2018;Profant et al., 2020).The appropriate perception of supra-threshold auditory features such as frequency specificity and temporal compression, as well as speech recognition in noisy environments, is also dependent on the functional integrity of the central auditory system (Giroud et al., 2018).Therefore, it is meaningful to use several supra-threshold tasks, e.g., perception of temporally modulating acoustic cues, in our study here.Some studies suggest that supra-threshold functions are affected in people with chronic tinnitus (Fournier and Hébert, 2013;Ibraheem and Hassaan, 2017;Sanches et al., 2010) while others do not (Boyen et al., 2015;Bureš et al., 2019;Moon et al., 2015).An established device to investigate the temporal perception ability of the auditory system, i.e., to measure the ability to perceive rapid temporal changes, is the gap detection task (GDT) (He et al., 1999).Many studies with individuals without chronic tinnitus have shown significant effects of age (He et al., 1999;Moore et al., 1992;Schneider et al., 1994;Snell and Frisina, 2000) and hearing loss (Pichora-Fuller et al., 2006) on poorer gap detection.In individuals with tinnitus, there is no clear picture yet; some studies have found differences between tinnitus and control groups (Fournier and Hébert, 2013;Gilani et al., 2013;Ibraheem and Hassaan, 2017;Sanches et al., 2010), while others have not (An et al., 2014;Boyen et al., 2015;Bureš et al., 2019;Kenneth and Werff, 2019;Zeng et al., 2020).A similar picture emerges for other auditory measures that are closely related to temporal perception, but also include sensitivity to intensity and frequency, namely acoustic modulation tasks.In natural acoustic signals, amplitude and frequency modulations (AM and FM) are key features and cues that critically help detect, discriminate, identify, and localize sound sources (Joris et al., 2004).Paul et al. (2017) applied an amplitude modulation task with background noise.Results showed that the individuals with chronic tinnitus had poorer amplitude modulation sensitivity for a first tone in their procedure but not for a second one.On the contrary, Bureš et al. (2019) have used a frequency modulation procedure, which showed no significant differences between tinnitus and control group.Another aspect of elementary auditory perception is the ability to discriminate intensities, that is, to perceive differences in loudness between sounds (Carlyon and Moore, 1984).There is little tinnitus-related literature on this topic.Zeng et al. (2020) have demonstrated a small difference in an intensity discrimination task (i.e., when subjects have to differentiate the intensity of two tones) at a high frequency, 30 dB SPL tone, where individuals with tinnitus performed superiorly compared to a control group.Bureš et al. (2019) also have shown better performance among individuals with tinnitus compared to a control group in an intensity discrimination task.Nevertheless, not all evidence points in the direction of enhanced performance.For example, Epp et al. (2012) found diminished performance in a tinnitus group at a specific sound intensity.Further methods to assess the integrity of the central auditory system are frequency and duration pattern tests (Marshall and Jones, 2017).They were designed to measure frequency sensitivity and temporal coding.These tests are used in part in audiological practice, for example, to determine central processing disorders in children (Balen et al., 2019).However, there is little literature on these tests in tinnitus, with some studies not finding any differences (e.g., Gilani et al., 2013;Raj-Koziak et al., 2022).
To date, studies found only small or even no direct (causal) effects of chronic subjective tinnitus on elementary auditory perception.This looks substantially different for speech comprehension.On the one hand, basal speech comprehension (i.e., phonetic decoding) does not seem to be affected in people with chronic tinnitus (Hennig et al., 2011).Speech comprehension tasks in noisy environments on the other hand, present a significant challenge especially for those individuals affected by tinnitus (e.g., Ivansic et al., 2017;Gilles et al., 2016;Huang et al., 2006;Vielsmeier et al., 2016).EF, like attention control, may thus be crucial for individuals with tinnitus to overcome the challenges of speech comprehension in noisy environments.In order to investigate the ability to comprehend speech under adverse listening conditions, researchers have widely employed SIN tests (Taylor, 2003).Studies investigating speech comprehension in individuals with tinnitus using SIN tests have reported impaired performance of the tinnitus groups when compared to control groups.Remarkably, these impairments have been observed even in the absence of measured threshold hearing loss, as evidenced by inconspicuous audiograms (Gilles et al., 2016;Huang et al., 2006;Moon et al., 2015;Niewiarowicz et al., 2022;Ryu et al., 2012;Mertens et al., 2013).A comprehensive review encompassing studies conducted from a twenty-year period further supports the notion of poorer SIN performance in individuals with tinnitus compared to hearing-matched controls.This finding remains mostly consistent across a heterogeneous tinnitus population (Ivansic et al., 2017) and it suggests that reduced speech comprehension in tinnitus patients is not solely attributed to impaired hearing but represents a distinct phenomenon.Nevertheless, it is worth noting that a subset of studies has not replicated deficits in speech comprehension among individuals with chronic subjective tinnitus (Bureš et al., 2019;Zeng et al., 2020).Zeng et al. (2020) have employed different methodologies and Bureš et al. (2019) have included a sample of older individuals, which may account for the variability in their findings.Recent modeling (Schilling et al., 2022) and opinion (Schilling and Krauss, 2022) papers surmise that SIN performance in tinnitus could even be improved by mechanisms of stochastic resonance in the subcortical auditory pathway.Yet, this assumption currently lacks both confirmation from the literature and original data from a dedicated paradigm.A recent neuroanatomical study explored the relationship between SIN performance and cerebral morphometry in tinnitus (Tai et al., 2023).Their findings revealed a noteworthy negative correlation observed between SIN performance and gray matter volume in distinct regions of the brain in the tinnitus group, amongst others the left cerebellum (Crus I/II) and the left superior temporal gyrus.Intriguingly, this correlation was absent in the control group, thereby implying potential neuroanatomical anomalies in the tinnitus population.Possible reasons for the contradictory findings may be the heterogeneity of individuals with chronic tinnitus themselves (Meyer et al., 2014;Vanneste et al., 2019) that may consolidate in distinct neural network organization of tinnitus as a function of distress or pathology, and/or methodological aspects.It may also stand to reason that the relationship between chronic tinnitus and auditory as well as speech comprehension in noise is not as straightforward or unidimensional as previously surmised.Therefore, it is of paramount importance that we extend the past research by including a wide range of threshold, supra-threshold, speech comprehension and cognitive tasks, to comprehensively investigate these aspects.Taken together, the majority of studies suggest that a SIN deficit is present in people with tinnitus, which, in turn, may be indicative of a deficit in the central auditory system, and/or top-down executive control.
Another critical addition to the above-mentioned aspects, with respect to hearing loss and tinnitus, is cognition, namely EF.Akeroyd (2008), in their review article, concluded working memory,3 more specifically verbal working memory, to be the best cognitive predictor associated with SIN performance for older, hearing-impaired listeners.In contrast, a meta-analysis with young listeners with normal threshold hearing showed that individual differences in working memory only account for a relatively small proportion in SIN performance (Füllgrabe et al., 2015).However, several reviews and meta-analyses suggest a potential correlation between tinnitus and cognitive domains that warrants further investigation and consideration (Brueggemann et al., 2021;Mohamad et al., 2016;Tegg-Quinn et al., 2016;Trevis et al., 2018;Clarke et al., 2020).Araneda et al. (2018) for example showed significant differences in behavioral outcomes in interference control4 in Stroop task with the tinnitus group performing poorer.Several other studies found a similar picture using a Stroop task (Araneda et al., 2015;Gonendik et al., 2021;Stevens et al., 2007).Another aspect of inhibitory control is response inhibition, which refers to the suppression of actions that are usually not favorable for the achievement of a goal (Verbruggen and Logan, 2008).By means of the stop-signal task response inhibition can be measured.This task has been used in the study by Leong et al. (2020), where no differences were found between the tinnitus and control group.A further study only found a significant difference when controlling for working memory scores (Trevis et al., 2016).Nagaraj et al. (2020) showed a significantly lower performance in a tinnitus group compared to the control group in a working memory task.Trevis et al. (2016) also observed significant limitations in the reaction time in the n-back task5 in the tinnitus group.However, the literature is not unanimous and there is evidence where no differences or even superior performances of the tinnitus group were found in the n-back task (Waechter et al., 2019(Waechter et al., , 2021)).A recent study suggested that in older Hispanic adults tinnitus might even be related to enhanced cognitive performance (Hamza and Zeng, 2021).The authors examined retrospective data from the National Health and Nutrition Examination Survey (NHANES, 2011(NHANES, -2012)), and found a specific tinnitus subgroup to exhibit superior cognitive performance, contrary to their initial hypothesis of diminished functioning related to tinnitus.Considering the study's limitations and unanticipated findings, these results should be regarded as singular and not as a feasible counterargument to the prevailing knowledge stemming from comprehensive original experimental work on EFs and tinnitus to date.
Undoubtedly, there are indications of clear cognitive deficits in individuals with tinnitus, especially in the domain of EF (e.g., Araneda et al., 2018;Clarke et al., 2020;Gonendik et al., 2021;Mohamad et al., 2016;Nagaraj et al., 2020;Tegg-Quinn et al., 2016;Trevis et al., 2018Trevis et al., , 2016;;Neff et al., 2021).Currently, the exact mechanisms underlying the EF impairments observed in individuals with tinnitus remains elusive.In conjunction with noted shortcomings in SIN performance, these challenges in EFs may indicate a deleterious, albeit yet insufficiently acknowledged, influence of tinnitus on essential aspects of (normal) daily functioning.Alternatively, it may represent a straightforward additive effect of tinnitus and deficits in EFs, which can considerably impair daily functioning like speech comprehension in adverse listening conditions.
To close this critical research gap and extend knowledge about the interaction of tinnitus, SIN performance, and EFs, we apply a probing design with a comprehensive set of several audiometric tests, speech tasks, cognitive tests, and standard questionnaires related to tinnitus and health status.These tests include pure-tone audiograms, pure-tone audiograms in noise, GDT, AM, FM, difference limen for intensity (DLI), speech recognition threshold in silence (SRT), SIN, gated speech (GS), Stroop task, Flanker task, stop-signal task and dual-n-back task.All of these tests were chosen to assess the multiple factors underlying impaired speech comprehension in chronic subjective tinnitus as comprehensive as possible.By incorporating innovative, novel tests such as the GS, we enhance previous methodologies by introducing a speech comprehension evaluation of greater complexity that putatively surpasses conventional techniques such as SRT and SIN tests in terms of cognitive demands, particularly on working memory (Moradi et al., 2014b).To the best of our knowledge, this is the first study which combines aformentioned investigation methods, especially sophisticated audiometry, comprehensive assessments of SIN and EF performance, into a single focused investigation.

Hypotheses
We compared a chronic subjective tinnitus sample to a control group that was carefully matched for age, sex, education, and threshold hearing acuity.As introduced above, much of the existing literature points towards poorer SIN performance in chronic subjective tinnitus individuals.In order to provide a comprehensive assessment of auditory perception, we also look at more elementary functions of the auditory system (e.g., gap detection ability), which are not expected to be different between the groups.We expect that the tinnitus group performs inferiorly to the control group in complex speech comprehension tasks (as measured with SIN and GS), but not in more basic speech perception tasks (as measured with SRT).With respect to EFs, impairments in working memory and interference control have been observed.Therefore, we chose respective assessment instruments to thoroughly test these aspects of EF (interference susceptibility with Stroop and Flanker tasks, response inhibition with stop-signal task, and working memory with n-back task).We here hypothesize that the tinnitus group performs poorer than the control group in the cognitive tests.

Participants
A total of 30 Swiss German-speaking participants (12 female) with tinnitus and 26 participants (12 female) without tinnitus between 23 and 58 years of age were recruited.Tinnitus participants reported having chronic tinnitus for at least 24 months.Due to excessive hearing loss (defined by mean thresholds greater than 20 dB HL, measured by the pure-tone average of 0.5, 1, 2 and 4 kHz), five participants in the tinnitus group as well as one participant in the control group had to be excluded, resulting in two groups of 25 participants each.Among participants with tinnitus, 16 individuals reported the absence of a discernible precipitating factor for their tinnitus, while 6 individuals attributed its onset to exposure to loud music during a concert, 2 individuals associated it with a period of heightened stress, and 1 individual linked it to the occurrence of a cold.Participants reported the absence of neurological conditions (such as a history of stroke, brain tumor, meningitis, epilepsy, Parkinson's disease, multiple sclerosis, dementia, schizophrenia, or Huntington's disease) as well as psychological disorders at the time of assessment.Participants also negated any professional musician status (with none having received formal musical training beyond that typically provided in school music programs), also none of them devoted more than two hours per week to playing a musical instrument.
Tinnitus participants were recruited through the participant pool from the Evolutionary Neuroscience of Language Group from the Department of Comparative Language Science of the University of Zurich.The control group participants were recruited mainly from the participant pool from the Department of Psychology of the University of Zurich and the Marketplace of the University.The study was approved by the Ethics Committee of the Faculty of Philosophy of the University of Zurich (Permit No. 21.4.18) and written informed consent was obtained from all participants.For participation, participants received monetary compensation.

Procedure
Each participant attended an appointment lasting between 2 and 3 hours.Tinnitus participants completed the tinnitus questionnaires prior to the laboratory visit.On the day of the examination, first the declaration of consent was signed, and a health questionnaire was completed by all participants.For the tinnitus group, the first part of the tinnitometry was performed afterwards, which comprised a tinnitus frequency and loudness matching.Then, four cognitive tests were completed by both groups.For the tinnitus group, the second part of the tinnitometry followed, which consisted of another frequency and loudness matching paradigm.Subsequently, the audiometry was completed by both groups.Test order was not randomized and identical for all participants.Details of the tests performed are given in the following sections.Due to its importance and complexity, audiometry is explained first.

Audiometry
The audiometry, as well as the cognitive testing, were conducted in a sound-attenuated audiometric chamber.The audiometry was performed with a device developed by the Third Faculty of Medicine of the Charles University in Prague, which combines both hardware and software (Bureš et al., 2019).The hardware is a self-made audiometric device with a high-quality audio interface (RME Fireface, RME, Germany), which includes a programmable attenuator.The device allows digital-to-analog conversion as well as amplification and attenuation of auditory signals.Furthermore, Sennheiser High-Frequency Audiometric Headsets HDA 300 (Sennheiser, Germany) and an Arturia BeatStep (Arturia, France) controller with backlit buttons were used.The software to control the appliance was developed in a Matlab environment (Mathworks Inc., Natick, MA, USA).Certain presets were prepared specifically for our study that could be loaded to ensure a standardized procedure and to avoid errors.The results were exported in the form of a text file and for some tests, an additional manual evaluation had to be performed.The equipment was calibrated according to the ISO 389-5, ISO 389-8, ISO 8253-3 and IEC 60645-3 standards using the artificial ear 4153 from Brüel and Kjaer.
The procedures and parameters of each individual audiometric test are listed in Table 1.Please refer to Table T.1 in the supplemental material for a more detailed description of the individual tests.

Cognition
To measure EF, four computerized tests were used: a Stroop task, a Flanker task, a stop-signal task, and a n-back task.The 'PEBL' software, version 2.1 was used for the Stroop, the Flanker and the n-back task (Mueller and Piper, 2014) and the 'STOP-IT' software by Verbruggen et al. (2008), 'Exe' version was used for the stop-signal task.
The procedures and parameters of each individual cognitive test are listed in Table 2. Please refer to Table T.2 in the supplemental material for a more detailed description of the individual tests.

Tinnitometry
The tinnitometry included tinnitus loudness and frequency matching.They are briefly presented in the following section.

Loudness and frequency matching
The matching of tinnitus loudness and frequency involved participants adjusting a tone or sound to closely correspond to their individual perception of tinnitus.A method of adjustment was employed, allowing participants to independently set the frequency and loudness using a guided user interface (Henry et al., 2004(Henry et al., , 2006;;Neff et al., 2019;Tyler and Conrad-Armes, 1983).The specific tools utilized for this purpose included the 'Match Your Tinnitus' app, running on a Huawei MediaPad M5 tablet (Huawei, China), for pure-tone-like tinnitus (Android-version 8.0.0).Additionally, a custom software MAX 7 program (Cycling 074, USA) and a modular hardware controller (Palette Expert Kit; Palette, Canada) were employed for noise-like tinnitus.Sennheiser HD 25 headphones (Sennheiser, Germany) were used in both cases.
Following the adjustment of frequency and volume, a manual octave confusion test was performed to assess if the perceived tinnitus frequency would better correspond to a tone one octave higher or lower (Graham and Newby, 1962;Vernon et al., 1980).To enhance the reliability of the procedure, it was repeated a second time after the completion of the cognitive tests.

Questionnaires
To evaluate various health factors and psychopathological aspects related to tinnitus, we administered a set of three established questionnaires.The first questionnaire focused on general health and encompassed inquiries regarding medical conditions, medication usage, lifestyle factors, and language details.For the tinnitus-specific assessment, two questionnaires were utilized.The Tinnitus Sample Case History Questionnaire (TSCHQ; 35 items, comprising dichotomous questions, rating scales, and open-ended questions) to gather comprehensive information about tinnitus, including its characteristics and associated factors (Langguth et al., 2007), and the Tinnitus Handicap Inventory (THI; 25 items) to measure the impact of tinnitus on daily life (Newman et al., 1996), were used.

Data analysis
Statistical analyses were conducted using Version 4.2.1 of the R statistical software (R Project, Vienna, Austria) and the following packages: 'caret', 'corrplot', 'ggplot2', 'lme4', 'psych', 'qgraph', 'rstatix', and 'sjPlots'.Descriptive analyses included the presentation of means, medians, standard deviations, as well as minimum and maximum values.Statistical significance was defined as  < 0.05, and trends were reported at  < 0.1.One-sided tests were performed for directed hypotheses, while two-sided tests were used for all other undirected hypotheses.To account for multiple comparisons within a hypothesis, Bonferroni correction was applied to adjust the p-values and mitigate the accumulation of alpha errors (audiometry and speech tasks were corrected for 14 comparisons, and cognitive tasks were corrected for 4 comparisons).Given the probing and partly exploratory nature of the study, no a priori power and/or sample size analyses were performed.
Outliers were identified visually through boxplots and subsequently tested using z-scores.An outlier was defined as a data point with a z-score greater than three, indicating a deviation of more than three standard deviations from the mean.ANCOVAs were employed to compare the two groups, with age, sex, education, and PTA serving as covariates in all models.Assumptions of the ANCOVAs were evaluated visually and analytically.Homoscedasticity was assessed through residual plots, while the Durbin-Watson test was employed to test the uncorrelatedness of residuals.Normal distribution of residuals was assessed using qq-plots and a Shapiro-Wilk's test.Multicollinearity was examined using a correlation matrix.In cases where the assumptions of the ANCOVAs were not met, a non-parametric alternative was employed, specifically a rank-based ANCOVA-like approach (Olejnik and Algina, 2016).The effect size  2 was reported for each model.
Exploratory statistical analyses were conducted to further explore interactions among tinnitus-related variables and the significant outcomes of the group contrast within the tinnitus group.First, Spearman partial correlations were calculated using the qgraph package, with corrections made for all other correlations.The variables included in these analyses were the dependent variables that exhibited differences between the two groups, the covariates (age, sex, education, and PTA), tinnitus-specific self-reported measures (THI, self-reported loudness, and tinnitus duration), and psychoacoustic tinnitus measures (loudness and frequency of tinnitus matching).The resulting partial correlations were visualized as edges connecting nodes (i.e., variables) in a network diagram (Fig. 3).
Second, for readers interested in examining the interrelationships among all the cognitive, auditory, and speech variables of this study, we have computed a bivariate correlation matrix using the Spearman method using the corrplot package.Please refer to Figure A.1 in the supplemental material for a detailed visualization of these associations.

Participants' characteristics
The mean age of the experimental groups was matched as closely as possible (TI: 38.88 years (SD = 11.90) and CG: 39.76 years (SD = 12.19)), to exclude possible age-related differences ( = 305.50, = 0.900).The PTA of the two groups was also matched as close as possible (TI: 7.56 dB HL (SD = 4.02) and CG: 7.47 dB HL (SD = 4.72)), to exclude potential differences due to hearing loss ((47) = 0.07,  = 0.942), see Fig. 1.In addition, each frequency of the pure-tone audiometry was compared between the two groups and none of them revealed a difference.None of the participants showed asymmetry in the individual frequencies between the left and right ear.The lateralization of the tinnitus was evaluated during the loudness and frequency matching procedure and, on average, was in the middle between left and right.21 of the participants had a pure-tone-like tinnitus and four a rather noise-like tinnitus with a peak frequency in the noise.There were no differences between the two types of tinnitus in any of the tests or questionnaires used.The individuals with tinnitus all were diagnosed with subjective chronic tinnitus and the characteristics were as follows: the average tinnitus duration was 132.96 months (SD = 93.42), the mean THI was 30.80 (SD = 17.91), the average loudness in the matching was 38.21 dB SPL (SD = 15.81), the average pitch in the matching was 7.44 kHz (SD = 3.26), more details can be found in Table 3.   AP = adaptive procedure, SSD = stop-signal delay, SSRT = stop-signal reaction time.

Audiometry
Results of ANCOVA analyses can be found in Table 4, and are described in more detail in the following.

Gap detection task
In the tinnitus group, the mean length of the gaps was found to be 4.10 ms (SD = 0.77) and in the control group 4.15 ms (SD = 0.85).Due to the non-normal distribution of the residuals, a non-parametric alternative was employed.The statistical analysis did not yield a difference between the two groups ( (1,44) = 0.31,  = 0.583,   = 0.999,  2 < 0.01).

Difference limen for intensity task
The mean threshold for perceiving the smallest difference between two tones was determined to be 0.97 dB SPL (SD = 0.76) in the tinnitus group and 1.00 dB SPL (SD = 0.88) in the control group.There were outliers identified in both the tinnitus and control groups, and their removal did not alter the results.Due to the non-normal distribution of residuals, a non-parametric alternative was used.The findings indicated no difference between the two groups in terms of their thresholds ( (1,44) = 0.44,  = 0.510,   = 0.999,  2 < 0.01).

Frequency pattern test
The mean percentage of correct answers in the tinnitus group was found to be 95.87%(SD = 7.02), while in the control group it was 96.13% (SD = 3.56).Two outliers were identified within the tinnitus group; however, their removal did not substantively alter the obtained results.Given the non-normal distribution of the residuals, a non-parametric alternative was employed for analysis.The results revealed no difference between the two groups ( (1,44) = 1.87,  = 0.179,   = 0.999,  2 = 0.03).

Duration pattern test
The mean percentage of correct answers in the tinnitus group was determined to be 97.73%(SD = 3.43), whereas in the control group it was 96.93% (SD = 3.59).No outliers were identified within either  (Newman et al., 1996), SD = standard deviation, Min = minimum, Max = maximum.group.The residuals were not normally distributed, therefore we used a non-parametric alternative.The results showed no difference between the two groups ( (1,44) = 1.45,  = 0.235,   = 0.999,  2 = 0.04).

Speech recognition threshold in silence task
The mean SRT was found to be 35.83dB SPL (SD = 3.83) in the tinnitus group and 35.54 dB SPL (SD = 3.58) in the control group.Within each group, there was one identified outlier; excluding these outliers did not yield a change in the overall outcome.No difference between the two groups was observed ( (1,44) = 0.81,  = 0.374,   = 0.999,  2 < 0.01).Nonetheless, an influence of the PTA on the SRT was detected ( (1,44) = 30.06, < 0.001,   < 0.001,  2 = 0.25).

Speech-in-noise task
The mean SIN SNR was determined to be -4.67 dB SPL (SD = 1.17) in the tinnitus group and -5.63 dB SPL (SD = 0.89) in the control group.A group difference was observed, indicating poorer performance in the tinnitus group compared to the control group (Fig. 2).This effect size was deemed large based on the statistical analysis ( (1,44) = 13.49, < 0.001,   = 0.009,  2 = 0.18).

Gated speech task
In the GS the mean duty cycle was found to be 45.93% (SD = 4.38) in the tinnitus group and 42.86% (SD = 3.39) in the control group.Due to the non-normal distribution of residuals, a non-parametric alternative was employed.The results showed a significant group difference, with the tinnitus group exhibiting poorer performance (Fig. 2).This ef-fect size was deemed large ( (1,44) = 11.16, = 0.002,   = 0.024,  2 = 0.16).

Stroop task
The average Stroop effect observed was 126.72 ms (SD = 86.27)among participants with tinnitus, while it was 64.90 ms (SD = 46.09)among participants in the control group.Due to a potential trade-off between accuracy and reaction time, one participant from the control group was excluded.The two groups exhibited differences (Fig. 2), indicating that the tinnitus group performed poorer, with a large effect size ( (1,44) = 7.17,  = 0.010,   = 0.042,  2 = 0.17).

Flanker task
The mean Flanker effect was 58.23 ms (SD = 17.42) for participants with tinnitus and 66.59 ms (SD = 18.87) for participants in the control group.In order to account for the potential trade-off between accuracy and reaction time, one participant from each group was excluded.The results did not indicate a difference between the two groups ( (1,44) = 1.39,  = 0.250,   = 0.980,  2 = 0.05).However, age was found to have an influence on the performance ( (1,44) = 15.62, < 0.001,   = 0.001,  2 = 0.25).

Stop-signal task
The mean SSRT was found to be 307.65 ms (SD = 26.07)among participants with tinnitus and 308.89 ms (SD = 40.92)among participants in the control group.In order to account for the potential trade-off between accuracy and reaction time, three participants from the tinnitus group and two participants from the control group were excluded due to their deviation from the expected 50% response prob-Table 4 Results of the ANCOVA analyses.The results are displayed for all participants for the effect of tinnitus on elementary auditory perception and speech comprehension as well as cognition, controlled for age, sex, education, and PTA.ability on stop trials, as indicated by the 'ANALYZE-IT' software output (Verbruggen et al., 2019).The results did not reveal a significant difference in SSRT between the two groups ( (1,44) = 1.37,  = 0.248,   = 0.994,  2 = 0.02).

Secondary analyses
Partial correlations were computed to examine the relationships between different variables.The network graph in Fig. 3 illustrates several positive associations between different variables.Age demonstrated a correlation with PTA ( = 0.54,  = 0.037) and tinnitus loudness (from matching) ( = 0.52,  = 0.048).Education exhibited a correlation with tinnitus duration ( = 0.57,  = 0.027).THI was correlated with self-reported tinnitus loudness ( = 0.62,  = 0.015).Moreover, tinnitus loudness (from matching) displayed a correlation with tinnitus frequency (from matching) ( = 0.80,  < 0.001).SIN performance was associated with GS performance ( = 0.71,  = 0.003).These correlations are further depicted in the correlation matrix shown in Fig. 4. Explorative bivariate Spearman correlations of the entire variable set contained in this study are plotted in Supplemental Figure A.1.Notably, no correlations were found between the set of speech comprehension and cognitive variables, while we found a positive correlation between working memory and good performance in PTA  .

Discussion
This study systematically and comprehensively examined the differences in speech and cognitive performance, as well as elementary auditory perception, between a chronic subjective tinnitus sample and a matched control group.We used a wide range of speech tasks, including novel tasks such as the GS, and several cognitive measures within the same sample.Our results indicate a poorer SIN and GS performance  (Newman et al., 1996), DUR = tinnitus duration, SRL = self-rated tinnitus loudness, LDS = matched tinnitus loudness, FRQ = matched tinnitus pitch, ST = Stroop effect.

Fig. 4. Partial correlation matrix.
The matrix shows only significant correlations.These were between age and PTA as well as tinnitus loudness (from matching), tinnitus duration and education, THI and self-rated tinnitus loudness, tinnitus loudness and tinnitus frequency (from matching), and SIN, and between GS and SIN.AGE = age, SEX = sex, EDU = education, THI = tinnitus handicap inventory (Newman et al., 1996), DUR = tinnitus duration, SRL = selfrated tinnitus loudness, LDS = matched tinnitus loudness, FRQ = matched tinnitus pitch, ST = Stroop effect.
in the tinnitus group, whereas the speech perception in the SRT and the elementary auditory perception, including supra-threshold tasks did not show any differences between groups.Further, we found differences in the Stroop task showing diminished performance for the tinnitus group, while no differences were found in the other cognitive tasks.Moreover, our results revealed no correlation between speech and EF task performance, nor any relationship between these test scores and various demographic, hearing, and tinnitus-related factors.

Elementary auditory perception
First, we investigated elementary auditory perception and found that none of the threshold and supra-threshold tests performed (i.e., GDT, AM, FM, DLI, PTA  , FPT and DPT) differed between the groups.Our data replicates several findings that also did not find differences between tinnitus individuals and control groups (Bureš et al., 2019;Gilani et al., 2013;Moon et al., 2015;Raj-Koziak et al., 2022;Zeng et al., 2020).Our results are thus in line with our hypotheses as well as with previous literature, while these null findings primarily serve to rule out any influence of peripheral elementary auditory processes on speech comprehension in individuals with tinnitus.

Speech comprehension
Second, our investigation on (impaired) speech comprehension in individuals with chronic subjective tinnitus revealed compelling findings that support the yet existing body of literature.In our study, we are able to clearly demonstrate lower performance of the tinnitus group in the SIN task, with a notable difference in SIN SNR of approximately 1.00 dB SPL between the tinnitus and control group, while the total range of the two groups was 4.37 dB SPL.This finding is in accordance with previous studies (e.g., Gilles et al., 2016;Moon et al., 2015;Niewiarowicz et al., 2022;Vielsmeier et al., 2016), which consistently reported differences in SIN performance between tinnitus and control groups.Therefore, our study not only aligns with previous research but also contributes to the robustness and generalizability of these findings.
Moreover, our study ventured into a novel area by examining the performance in the GS task.To the best of our knowledge, merely one prior study by Bureš et al. (2019) has investigated the GS task performance in individuals with tinnitus, yet in a sample with only older individuals and without finding differences between the tinnitus and control group.However, our present results diverged from this prior work, as we observed a clear difference with a poorer performance in the GS task among young individuals with tinnitus.The observed discrepancy between our findings and those reported by Bureš et al. (2019) may be ascribed to variances in participant age ranges or tinnitus handicap.In particular, Profant et al. (2019) and also Fogerty et al. (2022) elucidated that age exerts a discernible influence on performance in the GS task, whereby older individuals exhibit poorer performance.Another potential explanation for the lack of differences in the study by Bureš et al. (2019) could be attributed to the tinnitus group having lower tinnitus severity, as measured with the THI, compared to ours.Studies have already indicated that tinnitus participant with a higher THI exhibit lower SIN performance (e.g., Jagoda et al., 2018).Our study thus adds valuable insights to the understudied aspect of tinnitus-related speech comprehension deficits by providing novel and genuine GS data for a young tinnitus population.
Furthermore, we would also like to emphasize that the impact of tinnitus on speech comprehension extends beyond individuals with measurable hearing loss (e.g., Tai and Husain, 2019;Weisz et al., 2006).As in our study, results from several other research groups demonstrated that even individuals with tinnitus and with no measurable hearing loss experienced a negative effect on masked speech recognition (e.g., Gilles et al., 2016;Niewiarowicz et al., 2022).Additionally, Tai and Husain (2019) proposed that tinnitus can alter cognitive control processing, trying to shed light on the potential mechanisms underlying the observed speech comprehension deficits.These findings align with our results, further supporting the notion that tinnitus exerts an adverse detrimental influence on both speech perception and higher-level cognitive functions, negatively impacting individuals affected.Our results may reflect this reasoning, as, especially when it comes to supra-threshold speech perception (speech in adverse listening conditions: SIN and GS), tinnitus participants show deficits, but not in threshold speech perception (SRT).
Following Vielsmeier et al. (2016), this observation could be termed the 'central inhibitory deficit' hypothesis in individuals with chronic tinnitus.Such a top-down mechanism could also explain why we find deficits in the GS task, as it requires top-down control to fill in the missing information.These observations may also imply plastic changes in central auditory and non-auditory regions in individuals with tinnitus.Since supra-threshold perception is also dependent on the auditory cortex and surrounding auditory association areas (Giroud et al., 2018), the poorer performance in the SIN and GS task could indicate that on these higher levels an inhibition system is needed to filter out and suppress irrelevant sounds.Hence, it stands to reason that this system is deficient in tinnitus individuals (Meyer et al., 2016).Supporting this reasoning, smaller cortical auditory areas are generally reported in tinnitus while some studies report the contrary or null findings (Adjamian et al., 2014).
While most of current data -including results from this study -is evident of SIN deficits in individuals with tinnitus, it is worth mentioning that not all studies produced similar results (Zeng et al., 2020;Oosterloo et al., 2020;Bureš et al., 2019).Oosterloo et al. (2020) conducted an extensive and comprehensive study on population-based data from the Netherlands (i.e., Rotterdam Study) in which they detected reduced SIN performance exclusively among individuals afflicted with tinnitus and N. Sommerhalder, P. Neff, Z. Bureš et al. concurrent hearing impairment, while controlling for general cognitive abilities.Although the results of this study may indicate high validity given the sample size, it is worth considering methodological differences, whether in the statistical approach or in the method to measure SIN.The SIN task employed in the former study was a digit-in-noise paradigm, in contrast to the utilization of a more ecologically valid SIN task involving everyday sentences in our study.These differences could very well explain why the results of the former study and our results do not converge.
Collectively, our study provides further evidence to solidify the existing knowledge regarding the deleterious impact of tinnitus on speech comprehension under certain conditions.These findings emphasize the importance of considering tinnitus as a multidimensional condition that affects both auditory and cognitive processes.

Cognition
Third, with regards to cognition, Stroop performance resulted in a difference between the two groups, with the tinnitus group showing slower reaction times.This finding is consistent with previous literature, as several previous studies found a similar difference in the Stroop task between the two groups (e.g., Araneda et al., 2018;Gonendik et al., 2021).Therefore, our results could affirm findings of these previous studies.This slightly poorer interference control has been interpreted as an impairment of central executive control in individuals with tinnitus (Araneda et al., 2015).Other scholars go even further and suggest, that this deficit in top-down cognitive control is crucially involved in the maintenance of tinnitus (Searchfield and Goodey, 2010).This is supported by Araneda et al. (2018) who could show alterations in brain responses in the ventromedial prefrontal cortex and the dorsolateral prefrontal cortex in individuals with tinnitus in the Stroop task.In their experiment, the involvement of the aforementioned brain areas was linked to the efficiency of top-down cognitive control in tinnitus individuals.In accordance with Searchfield and Goodey (2010) they hypothesized, that this deficit in top-down cognitive control could at least partly explain why individuals with tinnitus have problems suppressing the tinnitus sensation.
In contrast, the two groups performed equal in the Flanker task.This is in line with e.g., Jensen et al. (2021), but contrary to e.g., Heeren et al. (2014) and Sherlock and Brungart (2021), who found a difference between the tinnitus and the control group, with the tinnitus group performing inferiorly in the Flanker task.The literature for the Flanker task does not seem as unanimous as the literature for the Stroop task: While both the Stroop task and the Flanker task tap into similar cognitive domains, they have yielded inconsistent findings in relation to tinnitus-related cognitive performance.This might be attributed to the tasks themselves, as it has been shown that they are statistically not related (Rouder and Haaf, 2019).Moreover, the inclusion of the Flanker task in this study was a logical choice due to its widespread application in neuropsychological assessment batteries for research purposes, its frequent pairing with the Stroop test, and its application in tinnitus research.Consequently, its incorporation into this study was deemed appropriate despite the lack of significant findings.
In the current study, it was observed that the groups did not exhibit differences in the response inhibition task.This outcome contradicts our initial hypothesis and deviates from the results reported by Trevis et al. (2016), who reported a poorer performance among individuals with tinnitus.However, our results are consistent with Leong et al. (2020), supporting the notion that there are no group disparities in this particular task.The incorporation of the stop-signal task into our current study is substantiated by its wide recognition as the gold-standard measure for evaluating response inhibition, as previously established in the literature (Logan et al., 1997;Verbruggen and Logan, 2008).It is noteworthy that our task differed from previously employed tasks in other studies, primarily due to the inclusion of an auditory instead of a visual stop signal.
Regarding working memory, the two groups performed equally well in the 2-back task.Additionally, we exploratively checked the 1-back and the 3-back condition but found no difference either.This is contrary to Trevis et al. (2016), who reported poorer performance of the tinnitus group in a 2-back task.However, our results are in line with Waechter et al. (2019Waechter et al. ( , 2021) ) who also did not find group differences in the n-back task.This might be due to the task as it does not only require working memory but also high levels of selective attention (Diamond, 2013).Nagaraj et al. (2020) used a span task6 to measure working memory and found clear differences between tinnitus and control group, with the tinnitus group performing inferiorly.Rossiter et al. (2006) also reported similar results applying an auditory working memory task.Overall, evidence exists that suggests deficits in working memory in tinnitus individuals.However, we here present results that do not support deficits in a dual-n-back task in individuals with tinnitus.Taken together, the cognitive results of our study are indicative of a diminished executive cognitive control among individuals with tinnitus which might be at interplay with observed SIN performance deficits.However, based on our data, it remains inconclusive whether tinnitus leads to worse cognitive performance or if diminished executive control contributes to the development of tinnitus.

Deficient central mechanism in speech-in-noise performance and executive functioning in individuals with chronic subjective tinnitus
Our present study generated novel insights into the specific auditory and cognitive deficits experienced by individuals with chronic subjective tinnitus.Our results support the hypothesis that a deficient central mechanism, whether auditory or cognitive, or both at interplay, may underlie the observed impairments in SIN and GS performance and executive functioning among individuals with chronic subjective tinnitus, which requires some elucidation.
In more detail, the impaired GS performance among individuals with tinnitus suggests that the deficit extends beyond SIN comprehension.GS tasks require active integration and interpretation of fragmented speech stimuli, relying on top-down control mechanisms and temporal processing abilities (Moradi et al., 2014b,a).The compromised central (neural) mechanisms in individuals with tinnitus may impede their ability to effectively use contextual cues and inhibit irrelevant information, resulting in poorer performance in the task and in similar real life situations.Following this, it can be reasoned that the presence of tinnitus may lead to increased attentional demands and thus reduce the availability of cognitive resources.This concept would align with the principles of load theory (Lavie, 2010).Load theory, grounded in the premise of finite neural resources for processing stimuli, distinguishes between perceptual load, which encompasses neural resources dedicated to sensory stimulus processing, and cognitive load, which involves a central resource engaged in cognitive functions.The incorporation of load theory might be useful in the context of tinnitus and could extend traditional tinnitus models, e.g., with Khan and Husain (2020) proposing a model, which integrates load theory and conventional tinnitus frameworks.Their model suggests that tinnitus likely exerts an influence on both perceptual and cognitive load.Our findings of central mechanism deficiencies in tinnitus seem to reinforce the relevance of the load theory in the tinnitus context.
The study did not find differences between individuals with tinnitus and the control group in SRT or elementary auditory perception.This outcome suggests that the basic auditory processing abilities, including sound detection and discrimination, remain relatively preserved in individuals with chronic subjective tinnitus.These results are consistent with previous research indicating that threshold auditory function is possibly, but not necessarily, impaired in individuals with tinnitus (Boyen et al., 2015;Jagoda et al., 2018;Zeng et al., 2020).In line with our hypotheses and previous literature, we therefore conclude that the absence of differences in these tests underscores the assumption that deficient central mechanisms rather than peripheral auditory dysfunction are responsible for impaired speech comprehension in individuals with tinnitus.

Limitations
Several limitations should be considered when interpreting the findings of this study.Most importantly, with 25 participants in the tinnitus and control group, respectively, our study could be considered underpowered.Given the probing and partly exploratory nature of our design, including the manifold of to a certain extent novel tests and the absence of any directly comparable former study design, calculation of meaningful a priori power and/or sample size estimates is rendered impossible at this point.We therefore can not rule out any missed effect in our results presented in this paper in the absence of any adequate a priori power analyses.Yet, looking at our hypotheses and results, as well as previous studies, our results may also be valid and thus allowed for further interpretation.
Beyond that, we acknowledge that our methodology introduces the possibility of order effects and participant fatigue, both of which could exert an influence on participant performance.Although we did not implement explicit randomization of the test sequence, we did implement measures to alleviate these potential sources of bias.Specifically, participants were provided with short breaks after the cognitive tasks and in-between the audiometric tasks to mitigate fatigue.It is important to note that our study cohort consisted exclusively of individuals who were both young and in good health; none of them reported experiencing fatigue or stress during the two-hour testing session.Further, there were no notable differences in tasks (besides the mentioned differences in SIN, GS, Stroop task), not even without correcting for multiple comparison.
Another noteworthy consideration pertains to the characteristics of tinnitus severity of participants.In our study, we observed that the tinnitus severity, as evaluated by the THI, predominantly falls within the 'mild' category.This observation is pertinent for both our clinical (in-house) and non-clinical (out-house) study populations, where severity/distress scores are considerably low compared to comparable study populations in other countries.Given this distinctive nature of our local study populations, we can not outrule any cultural bias on self-reported tinnitus severity.Furthermore, it is plausible that relatively low severity levels could account for the lack of discernible relations between THI and other variables in our secondary analyses.
In general, it is important to acknowledge the inherent subjectivity associated with psychoacoustic measurements, as they rely on the decision thresholds of the individual undergoing testing.While six of the audiometric tests in our study were designed to be user-independent, five of the tests relied on the judgment of the examiner.However, to minimize potential errors stemming from inter-examiner variability, all measurements were conducted by the same investigator.
An additional aspect subject to criticism pertains to the lack of matching between our auditory stimuli and the specific frequency of tinnitus (e.g., in the AM or FM).It is plausible that the observed outcomes could have varied between the groups had we accounted for this factor.
Moreover, it should be acknowledged that the tinnitometry technique employed in this study may have introduced an additional source of error.It is important to note that some participants experienced difficulties in accurately matching their tinnitus percept.Tinnitus matching remains a highly challenging task, with no consensus regarding a standardized method or best practice to date.Moving forward, it is suggested that a combination of multiple matching procedures could enhance the reliability and validity of tinnitus matching while also reducing completion time (Neff et al., 2019).
Methodological shortcomings were also observed in the cognitive tests employed.For instance, the Stroop task, which is typically assessed verbally, was performed on a computer in our study.This introduced potential source of error due to motor response variability.Additionally, it is notable that the Stroop and Flanker tasks were expected to measure similar cognitive components but yielded divergent results.This discrepancy may be attributed to the simplicity of the Flanker task, which might not adequately capture deficits in individuals with tinnitus.Consequently, we argue that the Stroop task is more appropriate for measuring interference susceptibility in individuals with tinnitus.Alternatively, it is plausible that these two tasks do not actually measure the same underlying constructs, as demonstrated by Rouder and Haaf (2019) who found minimal statistical correlation between the Stroop and Flanker tasks in their analyses.
Another point of criticism pertains to the selection of the stop-signal task, which proved to be challenging for participants and despite clear instructions, participants did not consistently adhere to the intended protocol, introducing potential biases in the results.Furthermore, the choice of the dual-n-back task may not have been optimal, as the number of trials per condition were relatively low.In contrast to the psychoacoustic measurements, cognitive tests were expected to be relatively user-independent.
Additionally, it is important to acknowledge that psychometric questionnaires inherently possess a degree of subjectivity and are susceptible to social desirability biases on the part of the participants.
A further point of note involves behavioral results and tinnitusspecific questionnaires.For the Tinnitus Functional Index (TFI), previous research has demonstrated that its auditory subscale does not significantly contribute to the overall assessment of tinnitus-related functional impairment (Fackrell et al., 2016).While not directly comparable, given the absence of the TFI questionnaire and/or an auditory subscale in the THI in our test battery, our data produced no correlations between THI scores and any of the functional test scores.
Further, our study did not incorporate measurement instruments for assessing depression or fatigue, which could potentially serve as confounding variables (Langguth et al., 2011).
Moreover, personality traits could be considered as a potential influence on the study outcomes.Wöstmann et al. (2021) recently reported an association between neuroticism and SIN performance.
Future studies should further explore these relationships by incorporating comprehensive psychometric personality assessment.

Future directions
Subsequent studies should address several important considerations highlighted in the previous section.Most importantly, there is an urgent need for more extensive investigations into the relationships between speech comprehension and cognitive abilities in individuals with tinnitus, as the existing literature in this area is considerably limited.To advance systematically in this research avenue, it is thus imperative, as a first step, to conduct a substantial number of comprehensive behavioral studies that involve collecting extensive psychometric data.Additionally, potential confounding variables such as depression, fatigue, and hyperacusis should be taken into account in more extensive research designs.In terms of speech comprehension research, it is advisable to not only incorporate highly controlled stimuli but also naturalistic stimuli to provide a comprehensive understanding of the phenomenon.Our study has taken an initial step in this direction by producing feasible, promising and innovative findings.Moreover, to gain deeper insights, it would be valuable to further collect and explore neurophysiological data to learn more about central auditory processing in individuals with chronic subjective tinnitus.Investigating potential changes over time with these neuroscientific methods could be a fruitful avenue, potentially in longitudinal studies following the onset of tinnitus.

Conclusion
Our present study builds upon previous research on the relationship between tinnitus, auditory and speech processing, and cognitive functions, and expands current knowledge by using a comprehensive behavioral and audiometric approach.We show that tinnitus affects SIN performance in two separate test procedures, including a novel complex speech comprehension task.The groups did not differ in the SRT and in the elementary auditory perception tasks.These combined results may point to a SIN performance deficit of higher cognitive rather than basal auditory levels in individuals with chronic tinnitus.The results of the cognitive tests may support this notion, in that we revealed significant deficits in interference susceptibility, specifically in the Stroop task.This impairment of central executive control could be related to the inability of individuals with tinnitus to suppress the perception of tinnitus.Our study therefore reflects what tinnitus sufferers report from everyday life and further highlights the importance of SIN research in tinnitus.Understanding the underlying mechanisms and functional impairments associated with tinnitus-related speech comprehension deficits is of utmost interest (not only for those who are concerned but also for clinicians and therapists) and can moreover inform the development of targeted interventions aimed at improving communication and quality of life for individuals with chronic subjective tinnitus.

Fig. 1 .
Fig. 1.Mean hearing thresholds from pure-tone audiometry of the control and the tinnitus group including tinnitus pitch and loudness.Hearing threshold: The colored ribbons indicate one standard deviation interval.Tinnitus matching: The petrol diamonds indicate individual tinnitus pitch and loudness estimates from the matching procedure.

Fig. 2 .
Fig. 2. Relationship between hearing and SIN, GS, and Stroop performance for each group.A SIN SNR and PTA are shown and it can be seen that the two groups differ from each other.B GS values (duty cycle) and PTA are displayed, both on a rank-based scale (1 to 50).C Stroop effect and PTA are displayed.TI = tinnitus group, CG = control group.

Fig. 3 .
Fig. 3. Visualization of the partial correlations in a network model.Blue connections represent positive correlations, red connections represent negative correlations.The plot shows positive correlations between age and PTA as well as tinnitus loudness (from matching), tinnitus duration and education, THI and self-rated tinnitus loudness, tinnitus loudness and tinnitus frequency (both from matching), and between GS and SIN.Most factors are at least partly interrelated.AGE = age, SEX = sex, EDU = education, THI = tinnitus handicap inventory(Newman et al., 1996), DUR = tinnitus duration, SRL = self-rated tinnitus loudness, LDS = matched tinnitus loudness, FRQ = matched tinnitus pitch, ST = Stroop effect.

Table 1 Audiometric
Procedures.Each individual audiometric test and the respective parameters and procedures are listed in the table.

Table 2
Cognitive Tasks.Each individual task and the respective parameters and procedures are listed in the table.