For many years, functional gain was the only method available to quantify the in situ performance of hearing instruments. Technically, functional gain is defined as the difference in dB between aided and unaided sound-field thresholds as a function of frequency; typically, the goal has been to “shift” thresholds into the range of 20–25 dB HL.
Today, probe-microphone technology can provide a more reliable and efficient method of quantifying the in situ performance of hearing instruments than a functional gain method. In current audiologic practice, however, aided sound-field thresholds alone often are used to determine the appropriateness of a hearing instrument fitting.
LIMITATIONS OF FUNCTIONAL GAIN
Some of the pitfalls and limitations of functional gain were reported as early as 1980 by Macrae and Frazier.1 They noted that when wearers of hearing instruments had regions of normal or near-normal hearing sensitivity, aided thresholds were equal to or poorer than the unaided thresholds in these regions. This phenomenon occurred because of internal hearing instrument noise and/or amplified room noise. That is, even if a high-gain hearing instrument was fitted to an individual with normal hearing, the aided thresholds were no better, and often poorer, than the unaided thresholds (if you don't believe this, try it).
In practice, this means that it would be difficult, if not impossible, to use functional gain measures to assess the adequacy of various venting schemes or electronic modifications to the frequency response when regions of near-normal hearing exist (<25 dB HL). Since this effect was first described, similar findings have been reported by others.2–5
Over the last 20 years, additional criticisms of functional gain have been described.6,7 Stelmachowicz and Lewis pointed out that functional gain is a poor predictor of the gain for typical speech inputs when a hearing instrument is operating non-linearly.6 At that time, most circuits were linear, so the problem tended to occur only for high-gain hearing instruments fitted to individuals with severe-to-profound hearing losses. They demonstrated that, even with aided sound-field thresholds in the range of 20–30 dB HL, conversational speech might be only a few dB above threshold if maximum output was set conservatively (as might be the case for a young child).
Additional criticisms included: poor test-retest reliability,8 limited frequency resolution (five to six frequencies at best), inefficiency, and an inability to provide any information about real-ear maximum output levels. Functional gain also requires valid behavioral thresholds, which are often difficult to obtain with young children. Seewald, Moodie, Sinclair, and Cornelisse provide an excellent review of these issues.9
PROBE-MIC TECHNOLOGY EMERGES
In the early 1980s, probe-microphone technology was introduced as an alternative to functional gain measures. Even with the earliest systems, the advantages of this new test technique were obvious: good frequency resolution, good test-retest reliability (even with children), and the procedure was much faster than functional gain (multiple gain and output curves can be obtained in a few minutes). Early studies using this technique revealed large individual differences in the gain and SPL delivered to the ear using the same hearing instrument.10 Studies also showed age-related differences in the SPL developed in the ear canal for children in the birth to 7-year age range.11–13 These results reinforced the importance of individual real-ear measures of hearing instrument performance when fitting both adults and children.
Over the next decade, the increased sophistication of probe-microphone systems, in conjunction with the development of target-based prescriptive methods, highlighted the importance of estimating (through electroacoustic measures) the audibility of speech over a range of commonly experienced input levels (e.g., soft, average, and loud speech). This issue has become increasingly important as the signal processing schemes used in hearing instruments have become more complex. We now know that: (1) simple test signals cannot predict performance accurately for many advanced signal processing hearing instruments and (2) estimating the audibility of speech at multiple input levels can provide valuable information when fine-tuning these devices to optimize performance.
So why, after more than 20 years of research, do some hearing instrument manufacturers still recommend the use of functional gain to evaluate their (non-linear) hearing instruments? Why do surveys show that many clinicians serving children still rely on aided sound-field thresholds to verify hearing instrument performance?14,15 Why do some training programs in audiology continue to teach this outdated and invalid procedure?
The following example is presented to persuade audiologists, educators, and manufacturers to abandon the time-consuming and less useful clinical procedure known as functional gain (aided audiogram).
Figure 1 shows the unaided audiogram of a patient with a moderate bilateral sensorineural hearing loss. For purposes of illustration, a Phonak P2-AZ hearing instrument has been adjusted to two different settings: linear and non-linear (wide dynamic range compression). Figure 2 shows a series of gain curves for the linear (upper panel) and non-linear (lower panel) conditions. Note that the gain does not change with input level for the linear condition and that the gain with a 50-dB SPL input signal is similar for the linear and non-linear conditions. There can be little argument that these two hearing-instrument configurations should pro-cess incoming speech very differently as the level of speech changes. Figure 3 shows the maximum output (SSPL90), which has been set to be equivalent for these two settings.
Figure 4 shows the sound-field aided audiograms obtained for these two hearing instrument configurations fitted to the right ear of the patient. The aided thresholds are essentially identical! This occurs because the gain for low-level inputs is similar. Aided sound-field thresholds are insensitive to any of the signal processing differences between these two circuits.
Figure 5 shows a series of amplified speech spectra for the two hearing instrument configurations using the Situational Hearing Aid Response Profile (SHARP).16 In each panel, the open circles show right-ear thresholds in dB SPL and the asterisks show the real-ear maximum output of the hearing instrument. The solid and dashed lines depict the amplified long-term speech spectrum and the hatched region indicates the portion of the speech spectrum that is audible.
Three different listening conditions are depicted for each circuit type and in each panel an Aided Audibility Index (AAI) is given (see Stelmachowicz, Kalberer, and Lewis for a discussion of AAI and compression ratio computations17). For the linear hearing instrument, the range of speech is 30 dB (+12 dB/−18 dB) and for the non-linear hearing instrument the range has been compressed.
For the lowest speech input level (average conversation at 4 meters), the AAI value is slightly higher for the non-linear circuit even though the gain was equivalent for a 50-dB SPL input. This occurs because compression has rendered the lower level components of speech more audible for the non-linear circuit.
For average conversation, the majority of the spectrum and the peaks of speech are above threshold at most frequencies for both circuits, so the AAI values are similar.
For a shout, the linear condition is in saturation (denoted by the shaded region) and the AAI is reduced accordingly (0.77). Saturation is prevented in the non-linear condition, resulting in a relatively high AAI value (0.93).
This type of information is not attainable from the aided audiogram. In fact, when one compares hearing instruments, the aided thresholds will always be best for the circuit with the greatest gain for low-level input signals or the lowest compression threshold. That information alone, however, does not provide a complete picture of how the device amplifies speech. Understanding how the range of typical speech inputs fits within each individual's dynamic range can help guide decisions such as the number of compression bands, compression threshold, and compression ratio. Similar displays of audibility can be obtained by using DSL[i/o]18 or NAL-NL1.19
ROLE of FUNCTIONAL GAIN
So, is there a place for functional gain or the aided audiogram? Yes. Aided thresholds are the best way to document performance for bone-conduction instruments, frequency-transposition devices, cochlear implants, and for cases where the unaided thresholds may be vibrotactile.
Aided and unaided speech reception or speech awareness thresholds can also be useful to demonstrate the benefit of amplification to parents of young children or to rule out the possibility of non-organic hearing loss, neurologic conditions, or auditory neuropathy. These latter pathologies often can be addressed quickly by estimating the aided threshold for speech. In general, the aided gain for speech should agree reasonably well with the 2-cc coupler gain in the speech frequency region (after correcting for real-ear-to-coupler differences). Large discrepancies would alert the clinician to the possibility of problems other than a peripheral hearing loss.
We have come a long way since 1980. For most fittings with modern hearing instrument technology, we now have more reliable, valid, and efficient electroacoustic alternatives to the aided audiogram for verifying the appropriateness of fittings.
1. Macrae J, Frazier G: An investigation of variables affecting aided thresholds. Austral J Audiol 1980;2:56–62.
2. Balfour PB, Hawkins DB: Aided masked thresholds: Case of deception. JAAA 1993;4:272–274.
3. Rines D, Stelmachowicz P, Gorga M: An alternate method for determining functional gain of hearing aids. J Sp Hear Res 1984;27:627–633.
4. Macrae J: Invalid aided thresholds. Hear Instr 1982a;33:20–22.
5. Macrae J: The validity of aided thresholds. Austral J Audiol 1982b;4:48–54.
6. Stelmachowicz PG, Lewis DE: Some theoretical considerations concerning the relation between functional gain and insertion gain. J Sp Hear Res 1988;31:491–496.
7. Seewald RC, Hudson SP, Gagne JP, Zelisko DL: Comparison of two methods for estimating the sensation level of amplified speech. Ear Hear 1992;3:142–149.
8. Humes LE, Kirn EU: The reliability of functional gain. J Sp Hear Dis 1990;55:193–197.
9. Seewald RC, Moodie KS, Sinclair S, Cornelisse LE: Traditional and theoretical approaches to selecting amplification for infants and young children. In Bess FH, Gravel JS, Tharpe AM, eds., Amplification for Children with Auditory Deficits. Nashville, TN: Bill Wilkerson Center Press, 1996: 161–191.
10. Wetzell C, Harford ER: Predictability of real ear hearing aid performance from coupler measurements. Ear Hear 1983;4:237–242.
11. Barlow NL, Auslander MC, Rines D, Stelmachowicz PG: Probe-tube microphone measures in hearing-impaired children and adults. Ear Hear 1988;9:243–247.
12. Feigin JA, Kopun JG, Stelmachowicz PG, Gorga MP: Probe-tube microphone measures of ear-canal sound pressure levels in infants and children. Ear Hear 1989;10:254–258.
13. Westwood GF, Bamford JM: Probe-tube microphone measures with very young infants: Real ear to coupler differences and longitudinal changes in real ear unaided response. Ear Hear 1995;16:263–273.
14. Hedley-Williams A, Tharpe AM, Bess FH: Fitting hearing aids in the pediatric population: A survey of practice procedures. In Bess FH, Gravel JS, Tharpe AM, eds., Amplification for Children with Auditory Deficits. Nashville, TN: Bill Wilkerson Center Press, 1996: 107–122.
15. Tharpe AM, Fino-Szumski MS, Bess FH: Survey of hearing aid fitting practices for children with multiple impairments. AJA 2001;10:32–40.
17. Stelmachowicz PG, Kalberer A, Lewis DE: Situational hearing aid response profile (SHARP). In Bess FH, Gravel JS, Tharpe AM, eds., Amplification for Children with Auditory Deficits. Nashville, TN: Bill Wilkerson Center Press, 1996: 193–213.
18. Seewald RC, Cornelisse LE, Ramji KV, et al.: DSL Version 4.0 for Windows, a software implementation of the Desired Sensation Level (DSL [i/o]) Method for fitting linear gain and wide-dynamic-range compression hearing instruments. London, Ontario: Hearing Healthcare Research Unit, University of Western Ontario, 1996.
19. Dillon H: NAL-NL1: A new procedure for fitting non-linear hearing aids. Hear J 1999;52(4):10–16.
© 2002 Lippincott Williams & Wilkins, Inc.