Accessing Frequency-Lowering Technology
Olivia is a 13-year-old girl with a mild to moderately severe sensorineural hearing loss (Fig. 46.1).
She was born 14 weeks prematurely (26 weeks’ gestational age) and weighed 1 lb, 8 oz at birth. She endured an extended stay in the neonatal intensive care unit (NICU), where she received ototoxic antibiotics to treat a life-threatening infection. Because of her excellent medical care at the NICU, Olivia is now a healthy child who excels in a typical classroom setting. She is enrolled in several honors classes and earns excellent grades in school. She also has a very active social life and is very well adjusted from an emotional perspective.
The ototoxic medications given to Olivia in the NICU appear to be the cause of a permanent, mild to moderately severe hearing loss for both ears. Otoacoustic emissions (OAEs) are absent for both of her ears, but previous auditory brain stem response (ABR) assessment showed repeatable neural responses down to levels consistent with her behavioral hearing thresholds. Consequently, the presence of neural dys-synchrony was ruled out. Additionally, ipsilateral acoustic reflexes are present for both ears at 500, 1000, and 2000 Hz at presentation levels of 90 dB HL or lower, indicating intact neural synchrony and normal middle ear function bilaterally.
Olivia has consistently worn high-end, digital, behind-the-ear (BTE) hearing aids for both ears. These hearing aids were coupled to half-shell earmolds with large sound bores and 3 mm horn tubing to provide maximal enhancement of high-frequency output. Probe microphone measures indicated the output of both hearing aids matched (± 3 dB) the desired sensation level (DSL) v5.0 prescriptive targets for children for inputs of 55, 65, and 75 dB SPL. Also, the maximum output to an 85 dB sound pressure level (SPL) swept pure tone did not exceed maximum output levels as recommended by DSL v5.0 for children.
In spite of Olivia's excellent academic and social outcomes, she and her mother did report that some challenges had arisen during the first 3 months of middle school. Specifically, she was having difficulty understanding some of her female teachers during class lectures as well as some of her girlfriends in social situations. Furthermore, Olivia's mother reported that Olivia seemed very fatigued at the end of the school day and was often very irritable at home in the evenings. Olivia noted that she felt like she had to expend a great deal of energy in concentrating on the speech of others in the classroom as well as in social settings.
Pure tone air- and bone-conduction assessment was conducted to rule out a change in Olivia's hearing sensitivity as a cause for her reported difficulties. The results of this assessment were essentially identical to those of previous hearing examinations. Acoustic immittance assessment was also unremarkable, effectively ruling out middle ear dysfunction as a cause for her reported listening problems. Probe microphone measures indicated that the output of both hearing aids was appropriate relative to the DSL v5.0 target, and electroacoustic analysis and a biological listening assessment both ruled out hearing aid dysfunction as a cause for Olivia's difficulties.
Speech recognition assessment and aided threshold testing were conducted in the sound booth. Aided thresholds to warble tone stimuli in the soundfield were below 20 dB HL at octave frequencies from 500 to 2000 Hz but were 40 dB HL at 4000 Hz and 70 dB HL at 6000 Hz. Olivia scored 96% correct on the W-22 monosyllabic word recognition test (presentation level: 50 dB HL) while using both hearing aids. Word recognition (W-22) was also assessed at a presentation level consistent with soft speech (35 dB HL), and Olivia scored 76% correct. Speech recognition in noise was adequate as represented by a 3 dB signal-to-noise ratio (SNR) needed for 50% correct performance in the binaural aided condition on the BKB-SIN test. Persons with normal hearing typically require a 0 to 3 dB SNR for 50% correct performance on the BKB-SIN test, so Olivia's performance in noise was quite good.
Overall, Olivia's performance on “conventional” audiological assessments was satisfactory. Also, the output of her hearing aids matched the DSL v5.0 target within ± 3 dB across the frequency range. The most concerning aspects of the aforementioned results were the poor aided thresholds at 4000 and 6000 Hz and the fair word recognition score at a presentation level consistent with soft speech.
It is likely that Olivia will not have optimal audibility for at least some speech sounds. Specifically, some high-frequency phonemes, such as /s/, /f/, and /th/, may be inaudible for Olivia. This is especially true when the talker is a female or a child. Stelmachowicz et al1 showed that the peak energy for the phoneme /s/ exists beyond 8000 Hz when the talker is a woman or child. Conventional BTE hearing aids fitted to custom earmolds do not offer adequate output beyond 4000 Hz to allow for sufficient audibility of high-frequency phonemes, particularly for female and child talkers.2
Examination of the output of Olivia's hearing aids as measured by probe microphone assessment indicates a suboptimal output for sounds above 4000 Hz. This deficiency is attributed to many factors, including (1) the long transmission line of a BTE (ie, the receiver tubing, the earhook, the earmold tubing, and the earmold sound bore) results in attenuation of high-frequency components of sound as it travels from the hearing aid receiver to the ear canal, (2) available gain before feedback in the high frequencies is typically insufficient to allow for acceptable output beyond 4000 Hz, and (3) many children have small external ear canals, so large horn tubing and sound bores cannot be provided to enhance high-frequency output.
The importance of a lack of access of speech energy above 4000 Hz should not be understated. For instance, it is known that over 25% of cues responsible for the perception of speech exist at or above 3000 Hz [American National Standards Institute (ANSI) S3.5, 2007]. Furthermore, phonemes with peak energy above 4000 Hz are also common. For example, it is estimated that the phoneme /s/ is the third most frequently occurring phoneme. A great deal of linguistic information is present in high-frequency phonemes. For instance, the phoneme /s/ denotes plurality, verb tense, and passiveness.
Finally, it is more likely that high-frequency consonants will be inaudible when the overall level of the input signal is consistent with soft-spoken speech. Again, examination of Olivia's hearing aid output as measured by probe microphone assessment indicates particularly inadequate aided output beyond 4000 Hz for low-level speech inputs. This fact is likely reflected in the deterioration in word recognition observed between assessments conducted at presentation levels consistent with average versus soft-level speech.
As previously stated, the aided thresholds at 4000 and 6000 Hz and the low-level word recognition score both reflected difficulties that Olivia may experience. Neither, however, fully indicated her ability to hear phonemes with peak energy above 4000 Hz. As a result, a novel word recognition assessment, referred to as the University of Western Ontario (UWO) Plural test,3 was administered while Olivia wore her own hearing aids in the soundfield. In short, the UWO Plural test features 15 simple words that are familiar to the typical pre-school-age child (eg, cat, book, shoe, etc.). These words are presented in the singular and plural form (30 words total), so correct recognition of words in the plural form is dependent upon the child's ability to hear the /s/ in the final position. The words of the Plural test are spoken by a female talker, so the peak energy in the final /s/ peaks beyond 6000 Hz. Two full lists of the UWO Plural test were presented from a loudspeaker in the soundfield at a presentation level of 50 dB A, while Olivia wore both of her hearing aids. She scored 60% correct on this test, with most of her errors consisting of a deletion of the /s/ in the final position. She also occasionally substituted high-frequency phonemes in the initial position (eg, she responded with rock for the word sock). The UWO Plural test fully exemplifies the difficulty Olivia experiences with detection and recognition of high-frequency phonemes.
Several solutions have been posed to address insufficient audibility for high-frequency speech sounds when using a BTE hearing aid. For instance, hearing aid manufacturers have developed BTE hearing aids with reported coupler bandwidths that extend out toward 10,000 Hz. Unfortunately, the aforementioned long transmission line of the BTE hearing aid still provides attenuation to the signal in the high-frequency range. Also, the available real-ear gain before feedback above 3000 Hz is often insufficient to allow for optimal output in the high frequencies.
Another potential solution is a BTE hearing aid with receiver-in-the-ear technology. This option has been used very successfully with adults, and because the receiver is positioned in the external ear canal, there is no signal attenuation related to the transmission line. This technology, however, is often unsuitable for young children because their small external ear canals will not accommodate the presence of a receiver. This limitation precluded the use of this type of approach with Olivia.
Finally, some contemporary hearing aids offer frequency-lowering technology, which shifts high-frequency inputs to a lower-frequency range where sufficient audibility is presumably more likely. It was determined that Olivia should be fitted with Phonak Nios BTE hearing aids (Phonak AG, Stäfa, Switzerland), which possess a frequency-lowering technology known as nonlinear frequency compression (NLFC). NLFC compresses all sounds above a clinician-programmable, cutoff frequency so that high-frequency sounds are shifted to a lower frequency. NLFC was initially intended for use with hearing aid wearers with severe to profound high-frequency hearing loss, but it was decided that its application might provide better access to high-frequency sounds for Olivia.
The same verification assessments that were used to evaluate Olivia's performance with her original set of hearing aids were used to evaluate the appropriateness of her new aids. However, a special adjunct examination was conducted during probe microphone assessment. Specifically, the Audioscan Verifit hearing aid analyzer (Audioscan, Dorchester, ON) was used to measure the real-ear-to-coupler difference (RECD) for each ear using the same foam eartip used to measure Olivia's hearing thresholds. This step was performed to determine her exact threshold in sound pressure level (SPL) at the eardrum. Then probe microphone measures were conducted, and the output of each hearing aid was matched (± 2 dB) to the desired sensation level (DSL) v5.0 target for the Audioscan Verifit “standard speech signal” presented at 55, 65, and 75 dB SPL. Probe microphone measures were also utilized to ensure that the maximum output of the hearing aid did not exceed the DSL v5.0 target for a 90 dB swept pure tone. The aforementioned measures were conducted with NLFC disabled.
Next, we used the Frequency Lowering Verification test of the Audioscan Verifit to ensure audibility for inputs out to 6300 Hz. This test is designed to demonstrate the restoration of audibility of high-frequency sounds that a hearing aid has shifted to lower frequencies. The Audioscan Frequency Lowering Verification test signal was developed by applying 30 dB of low-pass filtering with a cutoff frequency of 1000 Hz to the standard speech signal used in the Audioscan Speech Mapping fitting platform, except for a ⅓-octave band-pass region centered on either 4000, 5000, or 6300 Hz. For a hearing aid without frequency-lowering technology, there should be a ⅓-octave wide peak in the amplified output at the specified frequency of the test signal (eg, a ⅓-octave band centered at 6300 Hz for the 6300 Frequency Lowering Verification stimulus). To reflect the degree of audibility for high-frequency signals, the aided output for 6300 Hz Frequency Lowering Verification signal (presented at 65 dB SPL for this study) was measured with NLFC disabled and enabled. As seen in Fig. 46.2, the objective of this test was to verify improvement in the audibility of high-frequency speech components upon activation of NLFC.
The unaided thresholds (bottom line) in the example in Fig. 46.2 reveal that high-frequency components (eg, near 6300 Hz) of average conversational speech will be inaudible with conventional amplification (the #1 line indicates the real-ear output to the 6300 Hz Frequency Lowering Verification stimulus presented at 65 dB SPL with NLFC disabled in the hearing aid). However, when the NLFC is enabled, substantial improvements in the audibility of high-frequency speech components are expected. The example in Fig. 46.2 illustrates that NLFC has the potential to provide adequate audibility of speech inputs up to 6300 Hz (the #2 line indicates the real-ear output to the 6300 Hz Frequency Lowering Verification stimulus presented at 65 dB SPL with NLFC enabled in the hearing aid).
It was likely that the fatigue that Olivia experienced at the end of a long day was attributable to the lack of audibility for high-frequency sounds. Because of her inability to hear important high-frequency phonemes, she had to expend a considerable amount of energy concentrating so that she could “fill in the gaps” in instances in which she did not hear every aspect of the speech signal.
Furthermore, it was expected that better audibility of high-frequency phonemes would allow her to more effectively monitor her own voice. Consequently, it may improve the articulation of her own speech.
Olivia received substantial benefit from the use of hearing aids with nonlinear frequency compression. Her aided thresholds improved to 20 dB HL at 4000 Hz, 25 dB HL at 6000 Hz, and 20 dB HL at 8000 Hz. She achieved a score of 98% correct on the UWO Plural test, and she required a 1 dB SNR for 50% correct on the BKB-SIN test (as compared with +3 dB with her previous set of hearing aids). Furthermore, she scored 92% correct on the W-22 word recognition test when presented at a level of 35 dB HL. Additionally, her standard score on the Goldman-Fristoe Test of Articulation improved from 67 with the use of her previous set of hearing aids to 94 after 6 months of use with hearing aids with nonlinear frequency compression.
Olivia was very pleased with the benefit she received from the use of hearing aids with nonlinear frequency compression. She noted that she felt like she could finally understand people when they spoke in the classroom, rather than having to guess at what they said. Her mother also noted that Olivia's behavior was much better in the evenings after a long day of school, and she also noted that it did not seem as if Olivia had to use as much effort when listening at home and in social situations. Quite obviously, the dramatic improvement observed in audiological outcomes as well as speech and language outcomes justify the benefit of nonlinear frequency compression for a child with a moderate to moderately severe, high- frequency hearing loss. As such, hearing aids with nonlinear frequency compression should be considered not only for children with severe to profound high-frequency hearing loss but also for children with mild to moderate hearing loss. Full-time, consistent access to all sounds across the speech frequency range at all presentation levels is critical for optimal listening abilities, as well as for speech, language, academic, and social development.
1. Stelmachowicz PG, Pittman AL, Hoover BM, Lewis DE, Moeller MP. The importance of high-frequency audibility in the speech and language development of children with hearing loss. Arch Otolaryngol Head Neck Surg 2004;130(5):556–562