Introduction

Tinnitus is a common hearing condition affecting 14% of individuals worldwide [1]. It is the conscious awareness of composite or tonal stimuli for no known corresponding external auditory source [2]. Tinnitus severity varies widely from person to person, with some individuals experiencing mild annoyance and others experiencing severe impacts on their daily lives. Beyond its intrusive auditory nature, tinnitus can have various consequences for an individual’s life, including poor emotional well-being, poor sleep quality, and the ability to communicate effectively in noisy environments [3,4,5]. Various factors, including noise exposure, ear infections, age-related hearing impairment, and ototoxic medication-related side effects, can cause tinnitus [6].

Speech perception in noise refers to an individual’s ability to understand spoken language in background noise. This complex task demands an individual’s auditory and cognitive abilities [7,8,9]. Individuals with tinnitus often experience speech perception deficits despite normal hearing [10, 11]. In contrast, the literature suggests that speech in noise perception is unaffected in individuals with tinnitus and normal hearing [12]. The mixed results of the relationship between tinnitus and speech perception in noise are complex and based on multiple factors. Several mechanisms have been proposed to explain speech processing deficits in individuals with tinnitus. They are reportedly related to peripheral hearing impairment and central auditory processing dysfunction, the influence of cognitive control mechanisms [13], and attention-based networks [14]. These proposed mechanisms are influenced by various factors, such as the degree of hearing loss, tinnitus severity, age, sex, and hearing aid use [12, 15,16,17].

Despite the increasing interest in exploring the effect of tinnitus in various central auditory processing tasks, particularly speech in noise, the existing studies differ significantly in terms of the target stimulus type or masker. The preferred masker stimuli are babble speech, with the kind of babble varying across studies from a single talker [10] to an 8-talker [18, 19]. In terms of target stimuli, studies have used a variety of speech tokens ranging from consonant vowels (CVs) [20, 21] to sentences [4, 12, 22].

The literature suggests that tinnitus may interfere with auditory processing, leading to difficulties in speech perception, especially in noisy environments [23,24,25], but the underlying mechanism remains unclear. The literature suggests complex interactions between tinnitus and other factors, such as age, cognition, and hearing status [26, 27]. Tinnitus rarely presents solitarily; rather, it often coexists with hearing loss. Tinnitus often occurs alongside hearing loss, raising questions about whether tinnitus alone can affect speech perception, whether these effects are combined, or whether hearing loss might mask the effects of tinnitus [23, 28]. Additionally, age-related changes in auditory processing and cognitive function may further contribute to poor speech perception in tinnitus patients [15, 29, 30]. In addition, the variation in material and quantification methods also varies across studies [31, 32]. In the literature, speech in noise has been studied using materials ranging from phonemes to sentences. The cognitive resources required to process sentences or words differ significantly [33]. Similarly, the type of background noise can differentially affect speech in noise perception. It has been demonstrated that different types of maskers can produce varying levels of interference, with speech maskers leading to poorer scores than steady noise [34].

We hypothesize that tinnitus negatively impacts speech in noise perception and that this effect is moderated by various factors, such as type of hearing acuity, age, type of target, and masker. The current systematic review and meta-analysis aimed to comprehensively examine the impact of tinnitus on speech-in noise perception. Through meta-regression, factors such as hearing acuity, age, and target-masker type were quantified to explore the influence of these moderators on the relationship between tinnitus and speech perception in noise performance. Thus, this study offers a thorough understanding of the effects of tinnitus on speech perception in noise.

Methods

This study aimed to systematically review peer-reviewed, published studies employing speech-in-noise perception tests in individuals with tinnitus. The study was preregistered and approved by PROSPERO (ID: CRD42022350779). The current systematic review adhered to the PRISMA guidelines [35].

Eligibility criteria

The inclusion criteria were based on the Cochrane guidelines for conducting systematic reviews [36]. The review consists of studies consisting of individuals with continuous tinnitus with or without hearing loss. The individuals were adults aged > 18 years, irrespective of sex. The review included studies assessing various objective measures of speech in noise perception tests, irrespective of the type of stimuli and the type of interpretation. All the studies were original research articles, such as randomized controlled trials, observational studies, and cohort studies. Studies comprising a control group, an experimental group or an experimental group alone were included in the review. Other study designs, such as case reports, case series and animal studies and reviews, were excluded. Furthermore, if the studies had intervention as the main objective, only the prespeech in noise perception scores were considered for review. The criteria are depicted in detail in Table 1.

Table 1 Eligibility criteria of the potential studies included
Full size table

Literature search strategy and information sources

Based on the databases, using the keywords "tinnitus" along with "speech in noise perception," a search string was constructed using appropriate Boolean operators (Keywords: (tinnitus [MeSH]) AND ("speech in noise" OR "speech perception in noise" OR "sentence in noise" OR "signal-to-noise ratio" OR "SNR" OR "signal in noise" OR "word in noise" OR "speech recognition in noise"). The keywords generated articles from five databases: PubMed, Scopus, Web of Science, CINAHL, and ProQuest. There were no restrictions on the language of the search strategy. The search was predominantly run through the title and abstract of all articles until January 10, 2023. Using Covidence software, 264 articles were retained after removing duplicates (n = 247) (Fig. 1).

Fig. 1
figure 1

The PRISMA flowchart visually summarizes the systematic review process

Full size image

Screening of articles

Two independent reviewers screened the articles based on title and abstract (Reviewer 1, SM and Reviewer 2, NK). Reviewer 3, KG, resolved the conflicts from this initial screening. Initially, 511 articles were identified through a primary database search. After the removal of duplicates, 264 studies remained for the initial screening. Subsequently, a thorough examination of the titles and abstracts led to the inclusion of 47 studies for full-text analysis. After an additional in-depth review, 32 studies were deemed suitable for inclusion in the systematic review, while articles with a high risk of bias were excluded. For a detailed illustration of the process and the exclusion criteria, refer to Fig. 1, which presents the PRISMA flowchart.

Extraction of data

Data extraction was independently performed by two reviewers (SM and NK), and a data check was performed by a third reviewer (KG). The extracted data included the following: type of study design, place of study, age and gender of participants, mean and standard deviation (SD) values of the speech in noise perception test as the primary outcome variable and other data, such as hearing and tinnitus characteristics. The unreported means and SDs were requested from the corresponding authors or extracted from the figures using the WebPlot digitizer online tool [37].

Risk of bias (RoB) analysis

Quality assessment of the 36 eligible studies was carried out by two independent reviewers (BR and SM), and conflicts were resolved by HP. The Critical Appraisal Skilled Programme (CASP) [38], which is based on case‒control studies, was used for risk analysis. The studies were evaluated based on whether they utilized a thorough aim of the study, methodology, the credibility of the findings, and their relevance from a clinical perspective. Questions such as “did the study address a clearly focused issue?”, “Did the authors use appropriate method to answer the question?”, “Were the controls selected in an appropriate way?”, and many more were used to qualitatively assess the studies. The last column consisted of the overall quality of the studies based on the various domains, and this was considered the final result. Based on these factors, the reviewers were instructed to rate the overall risk as either "High", "Average", or "Low". A third reviewer resolved conflicts when the two results were opposite to each other, and this was considered the final qualitative assessment. Based on the overall input from the reviewers, four articles were excluded (Table 2).

Table 2 Results of risk of bias (RoB) analysis across studies using the Critical Appraisal Skills Programme (CASP) Checklist
Full size table

Data synthesis

The extracted data were synthesized into a narrative form under various categories, including place of the study, age and gender of the participants, characteristics of tinnitus, hearing acuity of the participants, tinnitus assessment, tools used for evaluating the severity and impact of tinnitus and interpretation of the speech in noise perception scores.

Meta-analysis and meta-regression

Comprehensive meta-analysis software (version 4) [39] was used to analyze the influence of tinnitus on speech perception in noise scores, combining the means and SDs of the speech in noise perception scores and assessing heterogeneity. The analysis used the random-effects model to allow the results to be generalized to other studies [40]. Continuous variables were estimated using differences in means with 95% confidence intervals (CIs). The absolute heterogeneity level among the effect estimates was assessed by calculating the between-studies SD and 95% prediction interval (PI). I2 values of < 25%, 25 to 50%, and > 50% represented low, moderate, and high heterogeneity, respectively. Furthermore, to investigate the potential role of moderators such as hearing acuity, age and target type on the influence of speech perception in noise scores among the tinnitus population, a subgroup meta-regression was performed.

Results

Narrative synthesis

Overall, the current systematic review included 32 published studies for qualitative analysis. The overall summary of the reviews, such as demographic information, is presented in Fig. 2.

Fig. 2
figure 2

The total number of studies based on a place of study, b age group, c type of amplification device, and d hearing status

Full size image

Place of study

Twelve studies were conducted in various European countries, including the United Kingdom, Belgium, Netherlands, France, Denmark, Sweden, and Austria. In addition, seven studies were carried out in the United States of America; eight in Asian countries such as Taiwan, Korea, India, and China; four in South America; and one in Australia. Figure 2a depicts an overview of the different places where the study was conducted.

Age and sex

Fourteen studies included participants with tinnitus ranging in age from 18 to 75 years. However, 12 studies included only younger adults (from 18 to 55 years), and six studies included only older adults (from 55 to 75 years). Two studies classified the participants into young and older groups with tinnitus [12, 41]. The proportion of this classification is depicted in Fig. 2b.

Hearing acuity

In most of the studies [number of studies (N) = 17], routine pure-tone audiometry was used, while in a few studies, high-frequency pure-tone audiometry was used (N = 3). The review included the mean threshold of the control group, which was 18.5 dB HL, and that of the tinnitus group, which was 31.6 dB HL. The mean thresholds were calculated as three-tone averages of 500 Hz, 1000 Hz and 2000 Hz bilaterally, except for the study by Tai and Hussain [61], which included right- and left-ear-specific thresholds. Out of the 32 studies, 16 reported a pure tone average of participants less than 25 dB HL, and four studies categorized the hearing thresholds as mild severity. Fifteen studies included participants with more severe hearing loss (ranging from 40 HL to 108.51 dB HL). Figure 2c, d provide information on the total number of studies conducted on hearing status and the type of amplification devices used.

Tinnitus characteristics

As per the inclusion criteria of the current review, all the studies included only participants with continuous tinnitus. The mean tinnitus duration reported in ten studies was 7.04 years (ranging from 1 month to more than 10 years). Fifteen studies reported on individuals with more than 6 months of tinnitus. Three studies included participants with more than 3 months of tinnitus [10, 42, 43]. One study by Van Eynde et al. [44] included five participants with less than 1 month of tinnitus. Eight of the 32 studies included pitch-related information, and eight studies included loudness-based information (Tables 3, 4). The pitch and loudness measures were either included descriptively (tonal, noise-like pitch), the mean of the pitch matched, or the individual pitch of the participants. Tinnitus in both ears was reported in 12 studies, and unilateral tinnitus was reported in 13 studies, of which ten studies reported the laterality of tinnitus. Studies have also classified tinnitus as central (midline or global perception) and unknown laterality [15, 44,45,46]. The tinnitus characteristics are depicted as two separate tabular columns based on the inclusion of the tinnitus group + control group (Table 3) and the tinnitus group alone (Table 4).

Table 3 Tinnitus characteristics of the studies including both tinnitus and control group
Full size table
Table 4 Tinnitus characteristics of studies comprising the tinnitus group alone
Full size table

Tools used for evaluating the severity and impact of tinnitus

The most commonly used questionnaires/scales were the Tinnitus Handicap Inventory (THI) [47] (N = 20), followed by the Tinnitus Functional Index (TFI) [48] (N = 5) and the Tinnitus Questionnaire [49] (n = 4). Studies have also used the Tinnitus Reaction Questionnaire (TRQ) [50] (N = 1) to measure tinnitus severity. However, five studies did not report scales or questionnaires to measure tinnitus [42, 51,52,53,54]. Apart from the commonly used measures, the loudness or annoyance of tinnitus was measured using the tinnitus loudness scale (TLS) [16] and the annoyance rating scale [55]. The analog scale for loudness and annoyance was used in the following studies Gilles et al. [45]; Huang et al. [16]; Liu et al. [56]; and Mondelli et al. [21].

Of the 32 studies included in the review, 13 included only a tinnitus group and did not include a control group. These studies reported decreased speech-in-noise scores, aligning with the main objective of their research, which was to assess the effectiveness of interventions such as hearing aids, sound generators, cochlear implants, and auditory training. Six studies included intervention studies consisting of pre- and postspeech performance in noise. Similarly, four studies explored the effectiveness of hearing aids with or without sound generators in treating tinnitus. Other studies have explored speech-in-noise performance and correlated the scores with tinnitus characteristics and audiometric thresholds [44, 46, 52, 57].

Speech in noise measurement

The studies included a variety of target and masker combinations. Of these, the most commonly used targets were sentence materials (N = 24), followed by words and phonemes/syllables [20, 21, 23, 41, 42, 44, 57, 58] (N = 8). Speech-shaped noise was the most commonly used noise type (n = 10) [4, 16, 24, 51, 53, 59], and multitalker babble noise was most commonly used (n = 11) [15, 18, 19, 46, 55,56,57, 60,61,62]. The other types of noise used in speech in noise perception tests are spatial noise, competing signals, and amplitude-modulated noise [10, 12, 20, 45, 52]. A detailed representation of the target and masker characteristics is presented in Table 5.

Table 5 Characteristics of various speech-in-noise tests
Full size table

Meta analyses

A total of 13 studies were included in the meta-analysis based on the study's primary aim and the data availability.

Meta-analysis: speech in noise perception test

Since the overall studies had a variety of interpretations of speech in noise tests, further meta-analyses were divided into three categories based on the way the speech in the noise perception test was scored: (a) SNR50, (b) SNR loss and (c) raw mean scores. The meta-analysis of eight studies with an SNR50 comprising 1045 individuals with tinnitus and 3532 controls yielded a significant effect on speech performance in noise. Using the random-effects model, speech-in-noise perception scores among individuals with tinnitus had a standard mean difference of 2.85 (95% CI 1.68–4.02, p < 0.01) (Fig. 3). The figure depicts the difference in means, standard errors and variance of individual studies, along with the forest plot representation of the same studies. The relative weight (%) represents the sample size. Furthermore, Egger’s regression test, which is the relationship between the effect size estimates and their precision (standard error), showed p = 0.11, which indicated no publication bias. The same funnel plot is depicted in Fig. 4.

Fig. 3
figure 3

The meta-analysis of speech perception in the noise test was interpreted as the SNR50. A total of eight studies were arranged according to the relative weight and depicted along with the respective forest plots

Full size image
Fig. 4
figure 4

Funnel plot of studies that included speech-in-noise perception scores interpreted using SNR50

Full size image

In contrast, the results of the meta-analysis of studies using SNR loss and raw scores as outcome measures were less pronounced. For the raw mean score values, two studies were included in the meta-analysis [16, 20], which indicated a significant mean difference of − 8.747 (95% CI − 16.646 to − 0.847, p = 0.03) (Fig. 5a). Similarly, for SNR loss, two studies were included in the meta-analysis [17, 61], and the mean difference was 0.188 (95% CI − 0.360 to 0.737, p = 0.50), indicating no significant effect of speech on the noise perception score (Fig. 5b).

Fig. 5
figure 5

Meta-analysis of speech perception in the noise test interpreted as raw mean scores (above) and SNR loss (below). A total of four studies are computed and depicted along with the respective forest plots

Full size image

Overall, the meta-analysis results suggest that SNR50 scores may be poorer in those with tinnitus than in a control group without tinnitus. Speech in noise perception tests that calculate an SNR loss or a raw percent correct score may be less sensitive to these group differences.

Meta-regression

It is well known that factors such as age, hearing acuity and target type affect speech perception in noise scores in addition to tinnitus alone. Therefore, a meta-regression was conducted to investigate the potential role of moderators in speech perception and noise performance.

The covariates used for meta-regression were hearing acuity, age and target-masker type. Q statistics were used to analyze the fit of moderators in one model. Neither the combined model including hearing acuity, age and target type (i.e., intercept) was significant (1.572 ± 1.815; 95% CI − 1.986 to 5.130, p = 0.386), nor did it individually show a significant effect on speech in noise performance (hearing thresholds: − 0.005 ± 0.019, 95% CI − 0.032 to 0.044, p = 0.760; age: − 0.0008 ± 0.041, 95% CI − 0.082 to 0.081, p = 0.985; target type: − 0.786 ± 1.784, 95% CI − 4.283 to 2.711, p = 0.659) (Table 6). This finding implies that hearing acuity and age do not significantly influence speech-in-noise perception scores in tinnitus patients. Based on the goodness of fit, the heterogeneity was high, with significant differences among the moderators (Tau2 = 1.739, I2 = 93.04%, p = 0.000, R2 = 0.00).

Table 6 Meta-regression random-effects model: test of the model using three moderators (hearing thresholds, age and target type)
Full size table

The results are further depicted using a scatterplot visualizing the meta-regression results of the effect size (Hedges's g) on hearing threholds (Fig. 6a), age (Fig. 6b) and taregt type (Fig. 6c).

Fig. 6
figure 6

The scatterplot includes a regression line (intercept) and confidence intervals under two conditions: (upper left) depiction based on hearing thresholds (HT), (upper right) depiction of age changes, and (lower) changes in target type. HT hearing thresholds, TT target type

Full size image

Discussion

Speech perception in noise tests is among the complex auditory tasks and plays a vital role in understanding everyday listening situations. A fair number of studies have explored speech-in-noise measures in tinnitus populations. This review aimed to understand the effect of continuous subjective tinnitus on speech perception in noise scores in general and concerning certain factors.

The overall impact of tinnitus on speech perception in noise

The overall meta-analysis results suggest that continuous tinnitus significantly negatively impacts speech-in-noise performance. There could be a few possible mechanisms leading to this impact.

Hearing loss: First, the primary cause of tinnitus is cochlear damage. This reduces audibility, causing a broadly tuned response of the basilar membrane and negatively affecting temporal resolution and frequency selectivity [63, 64].

Another major cause is the presence of hidden hearing loss. SNHL, especially in aging ears, is often accompanied by tinnitus and hyperacusis. It has been suggested that loss of sensory outflow from the auditory periphery causes a compensatory increase in “central gain”, which underlies these perceptual anomalies, such as tinnitus [65]. In the current review, studies have highlighted that although normal audiometric thresholds exist, increased hearing thresholds at high frequencies (> 2000 Hz) imply damage to inner hair cells or cochlear nerve terminals. The literature reports that acoustic overexposure can lead to rapid, irreversible loss of cochlear nerve terminals, eventually causing slow degeneration of spiral ganglion cells despite the full recovery of thresholds and no loss of OHCs [45].

Central masking: The speech in the noise test is generally attributed to a top-down phenomenon. Hence, another major factor leading to poor speech in noise performance is the involvement of the central auditory system (CAS) rather than the periphery. The literature reports that tinnitus affects the primary system as a "central masker" to interrupt speech in noise performance [58]. It has been reported that tinnitus occurs depending on the different plastic changes in the CAS [45].

Auditory segregation: The speech-in-noise test requires an individual to segregate speech in the presence of noise. The current literature shows that individuals with tinnitus are less able to use segregation cues regarding talker or masker differences [56]. The segregation ability better represents the release from informational masking and the role of central processing difficulties.

Cognitive deficits: Previous literature has shown that tinnitus affects understanding speech in noise due to deficiencies in central processing and/or attention or working memory [28, 66]. Tinnitus affects cognitive abilities in a way that drives cognitive resources away from auditory inputs, making speech-in-noise performance difficult [67]. Previous literature endorses this through details of cognitive tests performed using both behavioral and electrophysiological tests [68]. It is highlighted in the literature that cognitive spare capacity (short-term storage and information processing) tends to decrease for higher-level speech processing in adverse listening conditions [69, 70]. Furthermore, for individuals with tinnitus and normal hearing, this is explained using the information-degradation hypothesis, where tinnitus, as a distracting sound, makes it more effective for speech-in-noise recognition, which results in the depletion of cognitive resources. The current literature associates the impact of poor speech-in-noise performance with auditory attention (mainly selective attention) [4, 17, 20].

In the overall review, the study by Zeng et al. [12] showed no significant difference in speech-in-noise performance between tinnitus patients and controls, even when hearing loss was controlled for. The authors suggested that tinnitus does not interfere with speech-to-noise performance, mainly due to spectral differences. The tinnitus individuals had more difficulty perceiving female talkers than male talkers owing to the high-pitch nature of tinnitus. Similarly, Bures et al. [15] did not observe a significant score change for individuals with tinnitus or individuals in the control group. The authors reported that understanding of speech is more affected by age than tinnitus. If both tinnitus and aging coincide, the effect of age is greater and may mask the effect of tinnitus. However, as the studies did not find a significant difference in speech-to-noise perception scores, they suggest a trend toward poor speech perception in noise tests among individuals with tinnitus.

An overall review and subgroup analysis were performed based on hearing acuity and age to understand various moderating factors for the major impact of speech-in-noise performance in individuals with tinnitus.

Age and speech perception in noise

Age is a critical factor that independently influences speech perception in noise [54]. The aging process leads to changes in the auditory system and increases vulnerability to noise in the background [71]. Most studies in the current review included individuals with tinnitus (approximately 18–75 years) in a more comprehensive age range. According to the meta-regression analysis, age may not play a crucial role in speech and noise perception in tinnitus patients. Previous studies have suggested that the effect of age on older adults could mask the effect of tinnitus, assuming that the influence of tinnitus on speech perception in noise could be smaller than the effect of age on older adults [15]. However, the current subgroup meta-regression results suggest the opposite, where the addition of age as a factor did not significantly alter speech perception in noise. Since no experimental evidence systematically controls for the effect of age at present, it would be interesting to explore this issue in future studies.

Hearing acuity and speech in noise perception

It is well known that hearing loss is the most common and independent factor that influences speech perception in noise [72, 73]. Reduced hearing is mainly due to damage to peripheral hearing structures, causing difficulty in discriminating incoming sounds to comprehend speech in noise [74, 75]. In the current review, many studies categorized tinnitus participants based on hearing acuity according to their audiometric thresholds. Hence, a subgroup meta-regression was performed to determine the effect of age. Based on the results, hearing ability may not be the main factor contributing to difficulty understanding speech in noisy environments. Since individuals with tinnitus show poor speech perception in noise scores irrespective of hearing loss, it is possible that tinnitus has a mechanism in higher centers that can result in poor performance concerning speech in noisy environments. This may be due to a central cause, such as a disorder in auditory processing or top-down cognitive deficits.

The literature also reports the inclusion of high-frequency audiometry to determine the influence of speech perception on noise performance. The authors reported that individuals with tinnitus and hearing loss (based on increased high-frequency thresholds) had significantly poorer speech perception based on noise scores [23]. This finding suggested that there might be a different mechanism underlying the pathophysiological mechanism of tinnitus and associated hearing loss. Since there are few studies controlling for high-frequency hearing loss at present, future studies can explore these factors in combination with tests assessing higher centers.

Speech-in-noise test: measurement and interpretation

The speech-in-noise test is a complex auditory test used to evaluate an individual's ability to understand speech in a noisy background. The speech-in-noise perception test included various target and masker combinations for the real-world simulation. In the current review, most studies used sentences as the target and speech-shaped noise (n = 12) and multitalker babble (n = 11). Sentences are a highly ecologically valid form of stimulus, as they mimic daily communication. They also present various contextual and linguistic cues to make the speech-in-noise perception test more sensitive for assessment. It has been reported that sentences combined with noise for an assessment can better capture the perceptual demands of everyday communication [76]. Furthermore, noise types such as speech-shaped noise and multitalker babble similarly mimic real-world conditions and create informational masking [77].

In the literature, three methods were used for speech in noise perception assessment: SNR 50, SNR loss and raw speech in noise perception scores. The current review consists of most studies skewed toward the SNR 50 as an outcome measure for assessing speech in noise perception and shows a significant negative impact of speech-in-noise scores. SNR50 represents the characteristics of the psychometric function at the SNR level that allows a subject to correctly identify 50% of the signal, i.e., the 50% speech reception threshold (SRT50) and the function's slope [78]. This detailed interpretation of speech in the noise perception test increases the sensitivity and reliability of the results.

The other methods used in the literature are SNR loss and raw score loss. Two studies each reported SNR loss and raw scores [16, 17, 20, 61]. The SNR loss represents the difference in the signal-to-noise ratio between two optimal conditions (with one reference value) in noise, quantifying the degree to which speech perception in noise is affected, and the raw scores are the overall average of the scores per item in the test. SNR loss resulted in no significant impact of speech on noise performance. This may be due to the limited number of studies included in the meta-analysis.

While the SNR loss and SNR 50 are valuable metrics, they differ in a few ways. SNR loss provides a global view of an individual's speech perception affected by noise, allowing for performance comparisons across norms. The SNR50, on the other hand, provides a specific reference point where speech intelligibility is affected for at least 50% of the responses, making it a critical SNR. It also considers parameters such as the presentation level of the target, the number of correct responses, and the total number of items tested per level, ensuring the accuracy of the test. Based on the above variations, the overall significant negative impact of speech perception on noise performance may be most suitable when interpreted using the SNR50.

Tinnitus characteristics and speech perception in noise

Based on the literature, most of the studies used the THI to assess the severity and impact of tinnitus, followed by the TFI, TRQ and TQ. In the literature, the THI is the most common tool used to assess tinnitus severity. This could be due to the development of the assessment tool much earlier than other tools and the ease of administration, as only three response options are included for each question. However, the literature suggests that the THI questionnaire has not been prospectively evaluated for assessing responsiveness and lacks a comparison of the effectiveness of the outcome [79].

The other commonly used tool in the literature is the TFI (n = 6), the most recently developed tool [48]. It is a tool used to measure the severity and impact of tinnitus and is based on the WHO International Classification of Functioning, Disability and Health (FCF) framework. Consisting of eight subscales, the TFI comprehensively assesses multiple dimensions of tinnitus-related impact. Although the current review consists of a few articles using the TFI, despite its recent development, future studies can include the TFI to measure tinnitus severity and impact to evaluate individuals with tinnitus and group them according to specific problems based on the subscales. This interpretation will have a substantial impact on both assessment and treatment measures.

Tinnitus severity is a fundamental aspect of tinnitus evaluation and provides valuable insights into the subjective perception of the loudness and annoyance of tinnitus. The current literature includes most studies of individuals with minimal to mild tinnitus severity (starting from a score of 15 based on the THI). The literature suggests a lack of correlation between tinnitus severity and speech in noise perception [15, 61]. However, there is no definite hypothesis to explain the lack of correlation except for the heterogeneity of the study sample.

Conclusion

The main limitation of the current review is the heterogeneity of the tinnitus population. It is challenging to homogenize the groups of participants based on various factors, such as etiology, comorbidities, and psychological factors. Furthermore, the lower sample size in various studies makes it challenging to generalize the findings. Despite these limitations, the current review integrates the findings of each study, addressing mainly the influence of tinnitus on speech in noise perception and other parameters.

According to the current review, tinnitus has an overall effect on speech-in-noise performance, and multiple reasons (causes) are hypothesized for this phenomenon. The literature reports multiple possibilities and hypotheses to account for poor speech recognition scores in noise, but most studies have not proven this. Future studies should explore the correlation between reduced speech perception in noise scores and other factors that may influence this correlation.