Comparison between frequency transposition and frequency compression hearing aids

Background High-frequency hearing loss is one of the challenges for accurate hearing. One method of amplification toward improved detection and discrimination of high frequencies is through lowering of high frequencies to a lower-frequency hearing region. Frequency compression (FC) or frequency transposition (FT) can be used for individuals with different configurations of hearing impairment, such as steeply sloping and high-frequency hearing loss (containing high-frequency dead regions). Objective The purpose of this review was to compare the effects of FC and FT on the speech recognition abilities of people with hearing impairment. Patients and methods This study is a systematic review of previous investigations carried out to test the efficacy of frequency-lowering algorithms (i.e. FC or FT) in improving the detection of high frequencies. The concept of FC and FT depends on enhancement of speech recognition in patients suffering from high-frequency hearing loss. In this review, compression shifting and frequency transposition were compared on the basis of improvements in the audibility of high frequency sound to improve the outstanding and acceptance level of speech. Conclusion It can be concluded from this review that both FC and FT are useful in people with high-frequency hearing loss. FC has a potential role in recognition of monosyllabic words, consonants, and sentences in the presence of background noise. In contrast, FT facilitates detection of fricatives that ultimately leads to improvement in the discrimination of consonants. Throughout the literature, only one study has compared both FT and FC using various speech tests to evaluate the performance of different hearing instruments. Therefore, further studies addressing the benefits of both modalities of hearing aids with a more standardized outcome measure for both adult and pediatric patients are required.


Introduction
High-frequency sounds are crucial for communicating the semantic aspects of speech, such as plurals and possessives. Most conventional hearing aid users have access to low-frequency information; however, they may have limited access to high frequencies, which leads to missing out on high-frequency speech consonants and thus a signifi cant impact on speech understanding, especially in the presence of background noise [1]. In previous studies, it has been demonstrated that highfrequency hearing sensitivity declines with age even in the absence of specifi c pathology [2]. Sensorineural hearing impairment is usually associated with damage to the outer hair cells in the cochlea; sometimes the inner hair cells can be damaged in certain frequency-specifi c regions in the cochlea [dead regions (DRs)], which can be defi ned by the range of characteristic frequencies of the inner hair cells and the neurons adjacent to the DR. Th is region cannot be detected by a typical audiogram, but it can be diagnosed by using methods involving psychophysical tuning curves and threshold equalizing noise testing. Th e treatment of patients with DRs related hearing impairement depends on which frequencies are eff ected as using hearing aids with suitable amplifi cation is suitable for those with low-frequency hearing loss. For high-frequency DRs, hearing aids with frequency transposition (FT) or frequency compression (FC) can be used [3].
Most people with sloping hearing losses and having high-frequency thresholds of about 70 dB and above are unable to extract information from high-frequency parts of the speech. Hearing ability deteriorates if these people are provided with hearing aids that amplify speech frequency range as it leads to reduction in the ability of those individuals to recover useful information from high-frequency parts of the speech signal [2]. One of the solutions to overcome

Comparison between frequency transposition and frequency compression hearing aids
Mohammed Alnahwi, Zeinab A. AlQudehy Background High-frequency hearing loss is one of the challenges for accurate hearing. One method of ampli cation toward improved detection and discrimination of high frequencies is through lowering of high frequencies to a lower-frequency hearing region. Frequency compression (FC) or frequency transposition (FT) can be used for individuals with different con gurations of hearing impairment, such as steeply sloping and high-frequency hearing loss (containing high-frequency dead regions). this problem is bringing down the high-FCs of the speech to lower-frequency regions where the person is more able to analyze sounds. Th ere are two methods of lowering the frequency of high-frequency signals: FT and FC.
Th e basic principle of FT involves transferring the high-frequency sound to lower frequency by adding the processed signal (transposed) to unprocessed signal in the lower frequency [2]. A number of techniques are available for FT. Th e fi rst technique was used in a commercial hearing aid (Oticon TP72, UK) that consists of two channels: frequencies below 3 kHz were amplifi ed as in a conventional device, and high frequencies within 4-8 kHz were converted into low-frequency noise below 1.5 kHz by a nonlinear modulator [4]. Inability to preserve the spectral shape of the incoming signal is the main drawback of this technique [4]. Th e second transposing technique divides the incoming signal into low and high pass bands, with crossover at a frequency of 4 kHz. Th e incoming information from the high pass band is subtracted by 4 kHz from each frequency and then mixed with the low pass frequency band. However, these transposition schemes can preserve some but not all of the information of the high-frequency spectral shape. Th ese schemes have been developed by Velmans and Marcuson [5] in the frequency RE cording device. More recently, a transposing speech vocoder has been developed to shift down the speech cues, such as level, pitch, and tone. In this method, speech is fi ltered into a bank of adjacent narrow bands and the level within each band is detected [6]. Th e transposing speech vocoder works on the principle of using levels detected from high-frequency bands to modulate low-frequency bands of noise or low-frequency tones ( Fig. 1). Th e transposition algorithm is a method of minimizing the problem of overlapping spectra and was discovered by Robinson et al. [7], who estimated the edge of the DR to decide the starting point of the transposition. Low-FCs were amplifi ed to ensure that sound remained unaff ected by transposition. Transposition was applied only if signifi cantly highfrequency sound components were present for the DR band to just above the edge of the DR and overlapped with the speech that has been low pass fi ltered. Th e latest technique for transposition used the audiogram to determine the point of starting frequency for transposition. Th e start frequency for each patient was calculated individually with the sloping hearing loss greater than 10 dB/octave (within frequencies from 0.5-4 kHz). Th e baseline performance was measured above the frequency of 1.6 kHz and above the 70 dB threshold. Th e frequencies below the starting frequency are amplifi ed without transposition. One octave above the starting DR frequency was analyzed by narrow range hearing aid with the highest intensity being transposed linearly and overlapped above the starting frequency. Th is technique was applied in the transposing device (Widex Inteos, Denemark), which determines the effi cacy of the audibility extender program [8,9]. Th e transposition processing usually does not aff ect the low-frequency information. As a result, it could produce more natural sound quality and preserve a harmonic relationship between FCs for the high-frequency-transposed region [10]. However, the disadvantage of this processing is the overlap between the transposed high-frequency information and the lower frequency. Moreover, unconditional transposition could add high-frequency noise and mask the useful low-frequency information. As a result, many new approaches attempt to avoid this problem by using the conditional transposition (transposition active when the input signals mostly consist of highfrequency information).
Th e basic technique of FC is decreasing the bandwidth for the output signals. Furthermore, it can be linear (also known as proportional shifting) or nonlinear (nonproportional shifting). Th e linear frequency shifting brings down all energy peaks (frequencies) to lower frequencies by a compression factor. For example, if the FC ratio is 2, then sounds at 6000 Hz will be shifted to 3000 Hz and those at 3000 Hz will be shifted to 1500 Hz and so on. Th is method will preserve the spectral shape information for the frequency component. In this method the constant ratio can off er a constant relationship between frequency component peaks and the important cues  for recognition of the vowels in speech [11]. However, the drawback of this method is that it provides an unnatural sound quality due to low pitch of speech. For instance, a female voice will be more like a male voice. Th is method is not used in commercial hearing aid devices because of unnatural sound quality. Nonlinear FC decreases the bandwidth for the input speech signals (Fig. 2) by increasing the amount of frequency lowering for high frequencies [12]. Th e advantage of these schemes is that they avoid the overlap between the high and mid frequencies (shifted and unshifted signals). However, the harmonic ratio for the high frequencies is compressed and will not be preserved. Th erefore, speech perception may be aff ected when the cutoff frequency of compression is moved to lower frequencies. Th e current commercial device, which incorporates nonlinear FC, is Phonak Naida (Sonava Group, Switzerland).
Th is study reviews published papers to compare the diff erent methods of FC and FT by analyzing the improvement seen on speech tests in patients with high-frequency hearing loss.

Research rationale, aims, and objectives
A number of studies have been conducted to determine the eff ectiveness of hearing aids incorporated with diff erent frequency-lowering algorithms. With the development of a range of frequency-lowering techniques, researchers have conducted a number of experiments to test the effi ciency of hearing instruments installed with diff erent frequency-lowering algorithms. We conducted this systemic review to ascertain the adequacy, effi ciency, and eff ectiveness of the most eff ective hearing aid algorithms for people with highfrequency hearing impairment, and compared the eff ectiveness of FC and FT in hearing aids by assessing the performance of patients with high-frequency hearing loss in speech tests by reviewing previous research fi ndings. Th e fi ndings of this study can be used as a secondary data source or as reference in other similar studies.
To meet our objective, we focused on the following research questions: (1) What are FC and FT? (2) How do these methods improve the speech recognition performance of individuals with highfrequency hearing impairment? (3) Is there any diff erence between these methods (FC/FT) in improving speech perception in individuals with high-frequency hearing loss, if studied before in the literature?

Patients and methods
A comprehensive systematic review was conducted and was based on an open, preidentifi ed, and reproducible method. Th e search was conducted at the Centre for Reviews and Dissemination, University of York (2008), and started with 'Database of Abstracts of Reviews of Eff ects ( DARE) and the Cochrane Database of Systematic Reviews ( CDSR)' (University of York, 2008). In each study, the main points were discovered, assessed, and summarized. Collectively, all of the selected studies were put together and a generalized conclusion after analysis was fi nalized. According to the Centre for Reviews and Dissemination ( CRD, University of York, 2008), it is important to determine the PICO elements for any systematic review. PICO stands for Participants/population, Interventions, Comparators and Outcomes. A PICO structure was used to frame the objective questions and to ensure that the researcher obtained the maximum studies related to the primary study by using two or more search terms in one of the four parameters [13] as demonstrated in Tables 1 and 2.

Study selection and data sources
Th e research method adopted in this study is the systematic review of existing literature on related  topics. Selected studies were based on qualitative research methodology. Data were collected from the following databases: Medline, OVID, PubMed, Embase, CINAHL, and ScienceDirect. Th e search used the following keywords: (1) Amplifi cation device (2)  Th is classifi cation aims to assess the strength of evidence for the benefi ts of intervention or treatment, such as hearing aids with FT or FC. Data in this study were analyzed by means of meta-analysis. Follow-up of references in relevant publications was undertaken so that all relevant publications were included. Th e studies collected from the databases were classifi ed and coded according to the topic, techniques used, number of subjects, results, and research methods, as shown in Fig. 3. In this review we included all nonrandomized studies, prospective cohort studies, and case-controlled series published in English language only from 1990 onward. Our study population included children and adults with high-frequency hearing loss. Th e studies included in our review were those with defi nite outcome terms such as speech perception score or signifi cant diff erence. Studies on other types of hearing loss, such as conductive hearing loss or fl at sensory neural hearing loss, were excluded from this review.

Results
Th is section will give a brief description of the included studies (Tables 2 and 3). Table 2 summarizes all studies addressing FT and Table 3 presents those addressing FC.

Frequency transposition in patients with high-frequency hearing loss
Two out of fi ve papers, that is 40%, showed overall improvement in speech recognition from FT in patients with diff erent degrees of high-frequency hearing loss, this improvement was due to training given for the device [15,25]. However, in three of fi ve (60%) studies no improvement was seen in the speech recognition test on using diff erent techniques for the transposition hearing aid [6,14,16]. In the study by Rees and Velmans [14] an improvement of 8% was reported with FT as compared with conventional hearing aids and consonant detection score (40%).
Robinson et al. [17] reported a small improvement of 20% in the detection of aff ricates in seven patients. However, the overall results showed that there is no signifi cant improvement. Kuk et al. [16] showed improvement in FT-enhanced fricative detection by 43%. Th e combined analysis (using pass/fail speech score criterion for any score above 60% in patients, indicating signifi cant improvement) of these studies indicated that FT improved aff ricate detection, consonant recognition, and high-frequency fricative production in both adults and children suff ering from high-frequency and sensorineural hearing loss. For instance, 10 children showed signifi cant improvement in consonant recognition from 18 to 69% and in vowel identifi cation from 56 to 90% after auditory training.
Percentage of patients who bene ted from frequency transposition (FT) and frequency compression (FC).

Figure 3
Overall bene t from frequency transposition in patients with high-frequency hearing loss Data were collected from fi ve papers showed that 17 patients (39%) benefi t from the intervention. Th is percentage is for patients in general (adults and children). For pediatric age group, the benefi cial outcome was achieved in 7 out of 10 children (70%). Table 4 illustrates comparative data for FT studies (number of patients benefi ted and speech score).

Frequency compression with high-frequency hearing loss
Five (71.5%) of the seven patients who underwent training on device usage showed improvement from FC [19,[21][22][23][24]. In contrast, two (28.5%) of seven papers reported reduced performance for speech recognition [18,20McDermottHJ. Frequencycompression outcomes in listeners with steeply sloping audiograms. International Journal of Audiology2006; 45:619-629."]. Th e review of FC schemes showed overall improvement in the sentence recognition ability of patients in the presence of background noise and plural recognition, compared with conventional hearing aids.

The overall outcome for frequency compression
Forty-fi ve (53%) of 85 patients benefi ted from FC compared with the conventional hearing aid. In addition, the patients showed greater benefi t when the compression ratio was decreased [26]. Patients trained on device usage and pediatric patients achieved better  speech performance, which was indicated in one paper that compared between children and adults with a benefi t percentage of 91% [27][28][29][30]. Table 5 illustrates comparative data for FC studies (number of patients benefi ted and speech score).

Comparison outcome between frequency transposition and frequency compression
As there was only one study that compared between FC and FT [27] and it was a very old review, it was not possible to carry out the meta-analysis for all 12 studies. Th is leaves a potential gap in the methodological references and a question of how to compare the outcome between FT and FC when there is no recent paper that has conducted a direct comparative study between the two. To overcome these problems, it is recommended to look at the percentage of benefi t from the diff erent speech tests in patients with high-frequency hearing loss using amplifi cation with transposition and the percentage of benefi t for the same conditions using amplifi cation with FC. Figure  4 demonstrates the percentage of pediatric and adult patients showing benefi t from FT and FC and refl ects the diff erences between them (39% FT and 53% FC).
Although the achieved improvement in hearing level depends on other factors such as suffi cient training time, the presence of moderate high-frequency hearing loss with normal to mild low frequency and shifting the specifi c area (DR) the percentage of benefi t among children is greater than that among adults for both FT and FC. A high percentage of children using FC showed benefi t (91%) compared with children who benefi ted from FT (70%).

Summary of the results for the quality of the review
In this systemic review, we used level III and higher evidence, such as prospective cohort studies and casecontrolled series, to review all studies in the literature addressing both FT and FC. A slight change was allowed in sample size, study design, objective, and outcome measures (speech test) across all studies. Th is heterogeneity led to inappropriate use of metaanalysis, which conventionally needs strict similarities between the studies for specifi c parameters. As a result, the review applies metaethnography of the data and a qualitative form to integrate data synthesis [28]. Th e fi ndings of each study are tabulated and summarized in Tables 2 and 3. Th e relative feature of metaethnography is that it provides a general idea of what the evidence indicates, without prejudice to the results of each primary study. Regarding the inclusion and exclusion criteria, the review has a big collection of good-quality papers evaluating the performance of the transposition and FC and applies data for the knowledge base. A total 12 studies (level III studies, nonrandomized studies, prospective cohort studies, and case-controlled series) met the inclusion criteria. Most of the 12 studies were selected according to the inclusion criteria. Th ere   were seven papers on FT and fi ve studies on FC.
Th ere was only one paper that aimed to make a direct comparison between FC and FT [27], but it was not included in our review as it was using older techniques that produced poor natural quality of sound.

Discussion
In theory, shifting frequency hearing aids (such as FT and FC) have been reported to compensate the loss of audibility due to high-frequency loss (sloping threshold) and to lead to improvement in the speech score. When the clinician recommends FT or FC devices, they should be aware of many factors, such as the frequency range, which can support useful auditory information, and the sharpness of the sloping audiogram. Th ese devices using FT or FC achieve better results with patients who have moderate high-frequency hearing loss with normal to mild low-frequency hearing as compared with patients with profound hearing loss. Stelmachowicz et al. [30] reported improvement in fricative perception, which is important for children, for learning grammatical rules and articulation. Th is improvement was higher in patients with moderate high-frequency loss under a variety of low pass fi lter conditions [29]. In the same study, they found that the patient needed to extend the bandwidth up to 9 kHz when there was exposure to female or child speakers. With the introduction of digital signal processing, a number of researchers have tested the effi cacy of new frequency-lowering algorithms, such as FC and FT, with careful sample selection and training of patients and have reported benefi ts among adults and children with hearing loss [8,15]. Th e results of these studies indicate benefi ts of using FC or FT-based hearing aids, which is consistent with the fi ndings of previous researchers. It is evident from the fi ndings that children benefi t more than adults because of greater neural plasticity. Th e auditory sensations produced by these systems are initially unfamiliar when perceived by the cochlea. Th e cochlea is not tuned to hear higher frequency sounds delivered at lower frequency places along the basilar membrane. Th erefore, it is benefi cial to provide adaptation or training to users, regardless of the type of system used, to help users get acclimatized to these devices. Th e hierarchy of evidence level in Fig. 5

Training
Providing auditory training to patients with a prescribed hearing device is more eff ective for speech detection as compared with not providing training; this is supported by the fi ndings of many studies [8,15,18,19]. Th ere are two main reasons for that. First, patients adjust to the quality of frequency-lowered sounds because it will seem unnatural or outlandish at fi rst. Second, patients learn to discriminate or detect lowered signals; for instance, fricative sounds like /s/ and / / will be fi rst perceived as noise or unwanted sound rather than as speech.
As mentioned earlier, there are many studies reporting benefi t from training with both FC and FT. Posen et al. [31] showed suitable performance after 15 h of training in a speech test with transposition. Moreover, another study compared the performance of nine children who had severe high-frequency hearing loss with and without training, and reported good performance after training in the speech recognition test [32]. Th e overall results of training are encouraging. However, the training provided in previous studies Hierarchy of evidence obtains from the systematic review. was in speech tasks used in the specifi c test. Th erefore, the improvement could be a result of adapting to the specifi c speech task rather than due to shifting frequency (FC/FT Th e validity of our study results can be questioned as conclusions are derived from a combination of disparate studies/randomized controlled trials, having widely heterogeneous samples, small sample sizes, varied methods for application of the intervention, and possible selection and observation bias. Th e study suff ers from lack of availability of high-quality randomized trials due to incomplete information about the patient's clinical history, hearing characteristics, and baseline information about individual hearing loss. Also, the study of patients varied largely from study to study, as a result of which it is diffi cult to analyze their fi ndings on a common platform and may result in bias. In addition, only studies published in English were included in the review, as a result of which several papers with relevant fi ndings published in other languages were excluded. Th ese limitations can be overcome by testing these frequency-lowering algorithms on a larger sample with minimum variation in the type of hearing loss, patient's age, and experience of using the device. More scientifi c databases can be searched for a more refi ned search.
More research is required to compare the effi cacy of FC and FT techniques. Th ere is no study in this systematic review that directly compares the two algorithms. Th erefore, studies that carry out comparison between the two techniques in the same research with common participants and common fi tting strategies are required.

Conclusion
Past literature indicated that both FC and FT are useful in people with high-frequency hearing loss. Th e fi ndings of past studies have revealed that FC has a potential role in the recognition of monosyllabic words, consonants, and sentences in noise, whereas FT has a role in fricative detection that ultimately leads to improved discrimination of consonants, but does not negatively aff ect vowel recognition. Only one study has compared both FT and FC using various speech tests such as the vowel-consonant-vowel test, the nonsense syllable test), etc., to evaluate the performance of diff erent hearing instruments. Th erefore, further studies addressing the benefi ts of both hearing aids and standardizing the outcome measure for both adults and children are needed.