By By Francis Kuk, PhD
The use of sensograms or in-situ thresholds increases the accuracy of the hearing aid fitting and simplifies the task of verifying audibility. Here’s why.
One of the many benefits of utilizing digital signal processing in hearing aids is the realization of applications that would otherwise deem impossible to achieve. One such application is the measurement of in-situ thresholds, or determining the hearing threshold of the wearers with internally generated tones from the hearing aids.
This technique was first introduced in 1996 by Widex A/S along with its introduction of the first digital in-the-ear hearing aid.1 In-situ threshold was commercially coined as sensogram, to reflect hearing thresholds measured by the Widex Senso hearing aids. In recent years, other manufacturers have begun to implement the same application in their select lines of hearing aids as an optional feature. It is reasonable to question why this method of threshold determination is better for fitting hearing aids than using thresholds obtained through conventional means (ie, through headphones or inserts).
What Is In-situ Audiometry (Sensogram) and How Does It Differ from an Audiogram?
The sensogram is the hearing thresholds of a hearing-impaired person measured with his/her Widex hearing aids (and earmolds for BTE) inside the ears (ie, in-situ). The measured values are used by the fitting software to generate the target gain (or input-output characteristics) by the specific hearing aid.
All devices that generate or reproduce sounds (ie, hearing aids, audiometers with headphones/loudspeakers) have to be calibrated in a sound cavity so that the acoustic analog of the electrical voltage generated by the device can be estimated. The same electrical voltage results in different acoustic sound pressure levels (SPL) when it is measured in different sound cavities. In general, for the same voltage source, a large cavity (or coupler) produces a lower SPL and vice versa.
Audiometric thresholds or audiograms are measured either under headphones or with insert earphones. To determine the SPL delivered by the headphones (typically TDH-39, 49, or 50), a coupler that has a volume of 6cc is used to approximate the volume of air seen by a headphone on the average adult ear. On the other hand, if an insert earphone is used as the transducer, a 2cc coupler is used to calibrate the SPL. In both cases, the results are converted into a dB HL notation to reference the hearing level to young adults with normal hearing.
A hearing aid is typically calibrated with a 2cc coupler to approximate the volume of air seen by the hearing aid when it is worn in the ear. Examining the output from the 2cc coupler is adequate when the purpose of the test is to check the electro-acoustic integrity of the hearing aid (ie, quality assurance). A special 2cc coupler, called the 711 coupler, also has the volume and impedance characteristics of the average adult ear so that the coupler output can be approximated to the output seen on an average adult ear.
All coupler measurements are based on closed couplers with a fixed volume. The ear-canal, on the other hand, also can be viewed as a coupler, except that it is different among individuals (thus different impedances and volumes). In addition, when a hearing aid is worn, intentional leakage in the hearing aid through vents, unintentional leakage, impedance of the middle ear, and associated tubing resonances all contribute to changes in the SPL in the ear-canal relative to the coupler reading. Using the hearing loss information measured on a fixed size coupler (such as with the audiogram) to specify in-situ gain may not yield the intended amount of real-world amplification for the wearer.
The sensogram concept was formulated to account for the acoustic changes in the true threshold of hearing-impaired people when wearing Widex hearing aids. The sensogram, as measured through the Widex hearing aids, is calibrated on a 711 coupler so that its output on the average adult ear-canal can be ascertained. However, because the sensogram is measured with the patient wearing the hearing aid, the tubing used in the earmold, the amount of venting or unintentional leakage, the residual volume of the occluded ear-canal, and the impedance difference between the hearing aid receiver and the ear-canal all contribute to result in a different SPL from that measured in a 711 coupler. These differences could result in a different threshold between the audiogram and the sensogram.
Thus, although the audiogram and the sensogram are very similar for the average adult ear when measured with a closed earmold, sensogram thresholds are not identical to audiogram thresholds in other situations.
The purpose of measuring sensogram and audiogram is also different. The purpose of determining audiometric thresholds is for diagnostic and rehabilitation purposes: to establish the presence/absence of a hearing loss and to quantify its severity. The purpose of determining the sensogram is only for gain specification on the hearing aids as they are worn by the wearer under the specific acoustic conditions. This is because the acoustic conditions (tubing length, vent diameter, residual volume) could all contribute to affect the “apparent threshold” of the individual. This threshold could be substantially different from the audiometric thresholds.
Use of the audiometric thresholds to specify gain on a hearing aid runs the risk of not accounting for the individual acoustic variables that may affect the output of a hearing aid. In essence, one runs the risk of a less-than-precise gain setting.
How Reliable Are In-situ Fittings?
Because the sensogram is determined in the same manner as the audiogram (ie, same bracketing procedure as in an audiogram), one would expect the test-retest reliability of the sensogram to be similar to the audiogram. Smith-Olinde et al2 examined test-retest reliability of the sensogram measured at 4 octave frequencies (500, 1000, 2000, and 4000 Hz) in 43 adults using the Widex Diva hearing aid and insert ear-tips as the earmolds. These authors found less than 1 dB mean difference between the test and retest thresholds at each frequency. The authors concluded that the test-retest reliability of the sensogram is equivalent to that of currently accepted audiometric procedures. A similar conclusion was reached by O’Brien et al3 who also demonstrated that in-situ thresholds are reliable and valid.
Advantages of In-situ Fittings
Portability. The sensogram is a vital part in the fitting of the Widex hearing aids. Because the determination of the sensogram involves testing the patient with the hearing aids in the ear using a programming device, such testing can be conducted anywhere as long as it is quiet. When the external device is a portable computer (eg, netbook or notebook) or a dedicated programmer, it has the advantage of portability. The clinician/dispenser can take the programmer and determine the patient’s hearing loss (for the purpose of fitting the aids) anywhere desired. This is especially valuable for patients who are home-bound or in a nursing home.
Another example may be patients with a fluctuating hearing loss (eg, Meniere’s disease) where the hearing sensitivity changes unpredictably. McNeill4 reported patients with Meniere’s disease successfully measuring their own sensogram at home to compensate for their fluctuating hearing losses.
It must be stressed that the sensogram does not replace the audiogram. After all, the sensogram does not allow a determination of the patient’s bone-conducted thresholds, a measure critical to determine if the hearing loss is sensorineural or conductive in nature. However, if the nature of the hearing loss has been confirmed previously, the sensogram can be used as a convenient tool to monitor any changes in hearing sensitivity and aid in fine-tuning of the hearing aids.
Ease in determining real-ear thresholds. It was previously pointed out that the audiometric thresholds are not accurate in specifying hearing aid gain. Rather, the real-ear or in-situ thresholds are needed for an accurate prescription of required gain. To accurately determine the real-ear thresholds, clinicians have to couple the output of the audiometer to the patient’s earmolds, determine thresholds, and note the SPL using a probe-microphone real-ear system when a threshold response is made. This is an elaborate process that is not possible in most clinics. It has been suggested that the use of the real-ear-to-dial difference (REDD) could convert audiometric thresholds into real-ear SPL at thresholds.5 While this is certainly more accurate than simply using audiometric thresholds as the in-situ thresholds, the same study shows significant variability among subjects. The use of REDD allows specification of average in-situ thresholds, but it is not equivalent to measuring the individual real-ear threshold.
On the other hand, the sensogram is a real-ear threshold. The stimulus is generated from a hearing aid worn by the patient. The stimulus is further shaped by all the acoustic conditions (eg, tubing length, vent diameter, impedance, residual volume, etc) that could affect the magnitude of the stimulus within the wearer’s ear-canal. Thus, it is the real-ear threshold of the patient, except that this index can now be determined easily and accurately without any external devices such as a probe-microphone real-ear system.
Precision in specifying hearing aid gain. The availability of the in-situ thresholds allows gain on the hearing aid to be specified more precisely for the patient’s hearing loss.
For example, let’s assume that an adult wearer has an audiometric threshold of 50 dB HL at 500 Hz when measured using insert earphones. Assume this is the “true” hearing loss of the adult (another term for true hearing loss is Equivalent Adult Threshold, EAT). If one were to assign nonlinear gain for such a hearing loss using a hypothetical formula of 100% compensation for soft sounds, 50% compensation for normal sounds, and 10% compensation for loud sounds, one would have assigned 50 dB of gain for soft sounds, 25 dB of gain for normal sounds, and 5 dB of gain for loud sounds using information from the audiometric thresholds. If this is an average individual with the typical ear-canal characteristics wearing an occluding earmold, this could be the real-ear gain received by the patient.
Let’s say this individual uses a large vent in the hearing aid, and the sensogram threshold is measured to be 60 dB HL instead. That is, there is a 10 dB vent effect. If one had used the 50 dB HL to specify gain on the hearing aid, the resulting gain from the hearing aid as measured in the ear-canal, with the 10 dB vent effect subtracted off, would have become 40 dB gain for soft sounds, 15 dB gain for normal sounds, and -5 (or 0) dB gain for loud sounds. That is, the real-life gain is 10 dB less than what we had hope for (unless we verify with real-ear measurement and correct it afterwards).
On the other hand, if we had simply taken the 60 dB HL sensogram thresholds and specify coupler gain based on that, one would have 60 dB gain for soft sounds, 30 dB gain for normal sounds, and 6 dB gain for loud sounds. And taking away the 10 dB gain from the vent effect (when the hearing aid is worn), the resulting real-ear gain would be 50 dB gain for soft sounds, 20 dB for normal, and -4 (or 0) dB for loud sounds. The gain is closer to using the audiogram thresholds to specify gain in a closed earmold; however, it is not the same as the closed earmold. This effect has been reported by Keidser et al6 when examining the in-situ thresholds measured with a manufacturer’s in-situ audiometry module. The authors stated that appropriate corrections should be included when in-situ audiometry is used to measure vented earmold/hearing aids.
The sensogram (and fitting module) adopted by Widex includes a vent effect compensation algorithm in its sensogram determination. This patented algorithm is called Assessment of In-Situ Acoustics (AISA).7 In this algorithm, the actual vent effect is automatically estimated during the feedback test, the result of which is then added to the sensogram to estimate the true hearing loss and to compensate for the vent effect in the automatic fitting.
Using the previous example, when the sensogram is measured along with AISA activated, the sensogram will still be 60 dB. But internally, within the Widex hearing aids, it will calculate the true thresholds to be 50 dB (60 – 10) and base its gain specification on this threshold. Thus, a desired real-ear gain of 50 dB for soft, 25 dB for normal, and 5 dB for loud will be targeted. However, because of the 10 dB vent effect, 10 dB of the hearing aid gain will be dissipated because of this vent effect. Thus, the hearing aid will internally be set to a coupler gain target of 60 dB (50 + 10) for soft, 35 dB (25 +10) for normal, and 15 dB (10 +5) for loud sounds. And when the amplified sounds exit the hearing aids, the 10 dB vent effect will result in a real-ear gain of 50 dB for soft sounds (60-10), 25 dB for normal (35-10), and 5 dB for soft sounds (15-10). The various scenarios are summarized in Table 1.
|Sensogram||Sensogram + AISA|
|Desired gain||50, 25, 5||50, 25, 5||60, 30, 6||(50, 25, 5)|
|HA gain setting||50, 25, 5||50, 25, 5||60, 30, 6||(50, 25, 5) +10
60, 35, 15
|Real ear gain||50, 25, 5||40, 15, -5||50, 20, -4||50, 25, 5|
|Table 1. Comparing real-ear gain using different methods to estimate thresholds.
Thus, one can see that, if one simply uses the audiogram thresholds to specify gain on a hearing aid, the resulting gain would be different from the target when the hearing aid/earmold is vented. This is not a problem if one verifies with real-ear measurement and adjusts the gain settings on the hearing aid accordingly.
On the other hand, the real-ear gain resulting from the sensogram is closer to the desired gain even in a vented earmold. The accuracy of the sensogram is further improved when it is coupled with the AISA feature. This allows achievement of the desired target gain in any earmold conditions without the need of verification and adjustment. Systems that do not have this feature may be limited in accuracy of gain specification. Thus, in-situ thresholds improve the accuracy and ease of gain specification.
Ease and accuracy of predicting real-ear performance. One of the key ingredients to a successful fitting is completion of verification and validation. Kochkin8 showed that hearing aid satisfaction is inversely related to the number of visits paid to the dispensers. What is especially noteworthy is that those who verify and validate their fittings experienced on average 1.2 fewer visits than those who do not verify and validate. This means dispensers who spend time performing verification will save on average at least 1 hour per patient than those who do not verify. This could have significant financial and logistic impact on a private practice.
On the other hand, there is cost involved in verification. In addition to the cost of acquiring the equipment, the time involved in performing the task, the required skill level of the dispensers, the variability in the measurement, and the questionable validity of the measure in predicting real-life satisfaction prevented many dispensers from using real-ear measurement to verify the goodness of their fittings.
However, verification does not need to be time-consuming or expensive. Cox9,10 proposed some simple verification tools that do not require expensive setup and expert skill level. One of the tools is the use of sound-field aided thresholds to estimate the softest sound that wearers can hear with their nonlinear hearing aids. While the aided threshold does not provide information on supra-threshold processing and has measurement issues by itself, it does provide a powerful index of audibility. After all, there will not be any intelligibility when there is no audibility.
Aided Sound-field Thresholds
The measurement of the sensogram simplifies the attainment of the aided threshold. Indeed, with the sensogram information, the Widex fitting software can predict within 5 dB of the measured aided sound-field thresholds. This is because all Widex hearing aids specify gain based on the sensogram thresholds. Because the sensogram is measured in-situ with no additional acoustic conditions to affect its magnitude (eg, receiver, tubing, etc), gain applied to the in-situ thresholds would directly compensate for the thresholds.
The extent of the compensation is specified by the in-situ input-output curve. One can easily calculate the predicted aided thresholds by adding gain to the sensogram thresholds (which is essentially the unaided thresholds) to obtain the predicted aided sound-field thresholds. This is automatically performed by the fitting software.
Kuk et al11 performed a study to estimate the correspondence between the predicted aided sound-field thresholds and the measured aided sound-field thresholds in adult hearing aid wearers. The difference between the predicted and the measured aided thresholds for each subject at different frequencies is shown in Figure 1. For the majority of subjects, the deviation between the measured and predicted aided thresholds was within 5 dB—the magnitude of variability in typical audiometric threshold measurements. This suggests that the predicted aided thresholds are very similar to the measured aided SF thresholds.
Using the predicted aided thresholds saves clinicians valuable clinical time when the aided thresholds cannot be determined because of difficulties from equipment and room setup, or uncooperative patients. This is an especially valuable tool to estimate aided audibility in pediatric patients.
The author firmly believes in the merits of verification and Widex is committed to ensuring that all fittings are verified properly. However, judging by the fact that only about one-third of dispensers routinely perform real-ear measurement (REM),8 it is apparent that most hearing aid clinics do not have time to implement REM. Thus, Widex created the SoundTracker (ST) display in the Compass software as a simulation of the hearing aid output in the average individual ear.12 The hearing aid output shown on the ST is also calibrated on the 711 coupler.
Figure 2 shows the SoundTracker display. The individual bars are the individual channels. The display also shows the sensogram, the average output, the input level (in lighter color) and the gain (in darker color). The average output is also shown (green shaded area). The difference between the output level and the sensogram (marked by red arrow) is the sensation level (SL). The individual channel and overall rms levels of the input and the output are also available with the appropriate cursor placement. The display also can be expressed in dB HL or dB SPL notations.
While one cannot expect a direct correspondence in output level between the SoundTracker and the REM (because of the difference in output between a 711 coupler and the individual ear-canal), one can expect a direct correspondence in the level difference between the sensogram and the ST output (or sensation level, SL711), and the SL between the sensogram and the REM output (SLREM). This is because both the ST aided output and the sensogram were determined using the 711 coupler, while the REM uses the individual’s ear-canal as the cavity/coupler during the in-situ threshold and real-ear aided response measurements.
As the equations below show, the SL measured with the SoundTracker and the REM device should be identical.
Equation 1: SL711 = HA Output(711) – Sensogram
Equation 2: SLHL = HA OutputHL – Sensogram
Equation 2 can also be expressed as:
SLRE = (HA Output + RECD) – (Sensogram+RECD) = HA Output(711)–Sensogram(711)= SL711
We conducted an internal validation study to examine the sensation level shown on the SoundTracker and their correspondence with the SL shown on the Audioscan Verifit and the Frye 6500 system. We used the ISTS (International Speech Test Signal) on the Audioscan and the composite signal on the Frye as the test stimuli. In both cases, we measured the real-ear dB SPL of the sensogram tones in the subjects’ ears at thresholds. Then we presented the continuous ISTS signal from the Verifit system (and composite signal from the Frye) and measured its real-ear output in the subject’s ears. Meanwhile, we also noted the output of the hearing aid as displayed on the ST on Compass. The SL was calculated by taking the dB difference between the sensogram and the hearing aid output to the external stimuli. The results, as expressed in dB SL, are shown in Figure 3.
It is obvious that the average SL difference between the ST and the REM (Frye and Verifit) for most frequencies is within 1 dB. Indeed, even when examining the individual data, one sees between 1-2 dB variations for 80% of the measurements—an acceptable magnitude of difference given the reliability of REM is within 2 dB. This suggests that, while the absolute dB SPL displayed on the ST may not be the same SPL in the wearer’s ear, the SL displayed on the ST is identical to that shown on a REM. The direct correspondence between the SL noted on the ST and the REM suggests that audibility information displayed on both systems is identical.
The use of this information has been discussed in a previous paper on the use of the SoundTracker.12 This information is available because in-situ thresholds or sensograms are measured; the information would not have been possible if audiogram thresholds, or even corrected audiogram thresholds, were used.
Clinically, the close correspondence between the SL shown on the ST and REM suggests that, when REM is not available or cannot be easily performed, the ST may provide a convenient alternative to estimating audibility as long as sensograms are determined. The knowledge of this correspondence in SL may be especially valuable in pediatric fittings.
However, it must be stressed that the ST is not intended to replace REM. It is not an actual real-ear measure and the absolute SPL that is displayed may be different across wearers. However, if the purpose is to estimate how much audibility is achieved (ie, SL), the ST may provide an expedient alternative to REM.
The use of sensogram or in-situ thresholds increases the accuracy of the hearing aid fitting and simplifies the task of verifying audibility. This could save dispensing professionals time and resources, and could possibly reduce the number of return visits while ensuring, if not increasing, the patient’s satisfaction with the hearing aid fitting. If the infrequent use of verification tools is attributed to time and cost concerns, the sensogram or in-situ threshold can be an easy, convenient, and effective approach to ensure compliance and hearing aid satisfaction. ?
Francis Kuk, PhD, is vice president of clinical research at Widex USA and executive director of the Widex Office of Research in Clinical Amplification (ORCA-USA), Lisle, Ill. CORRESPONDENCE can be addressed to [email protected].
1. Ludvigsen C, Topholm J. Fitting a wide range compression hearing instrument using real-ear threshold data: a new strategy. In: Kochkin S, Strom KE, eds. High Performance Hearing Solutions, Vol II. Hearing Review. 1997;[Suppl]4(11):37-39. Available at: http://www.betterhearing.org/hia/high.cfm
2. Smith-Olinde L, Nicholson N, Chivers C, Highley P, Williams D. Test-retest reliability of in situ unaided thresholds in adults. Am J Audiol. 2006;15(1):75-80.
3. O’Brien A, Keidser G, Yeend I, Hartley L, Dillon H. Validity and reliability of in-situ air conduction thresholds measured through hearing aids coupled to closed and open instant-fit tips. Int J Audiol. 2010;49(12):868-876.
4. McNeill C. A hearing aid system for fluctuating hearing loss due to Meniere’s Disease: a case study. Aust N Z J Audiol. 2005;27(1):78-84.
5. Munro K, Lazenby A. Use of the ‘real-ear to dial difference’ to derive real-ear SPL from hearing level obtained with insert earphones. Br J Audiol. 2001;35(5):297-306.
6. Keidser G, Yeend I, O’Brien A, Hartley L. Using in-situ audiometry more effectively: how low-frequency leakage can affect prescribed gain and perception. Hearing Review. 2011;18(3):12-16.
7. Kuk F, Nordahn M. Where an accurate fitting begins: Assessment of In-Situ Acoustics (AISA). Hearing Review. 2006;13(7):60-67.
8. Kochkin S. MarkeTrak VIII: Reducing patient visits through verification and validation. Hearing Review. 2011;18(6):10-15.
9. Cox R. Verification and what to do until your probe-mic system arrives. Hear Jour. 2009;62(9):10-16.
10. Cox R. Verification and what to do until your probe-mic system arrives. Hear Jour. 2009;62(10):10-14.
11. Kuk F, Ludvigsen C, Sonne M, Voss T. Using in-situ thresholds to predict aided sound-field thresholds. Hearing Review. 2003;10(5):46-50.
12. Kuk F, Damsgaard A, Bulow M, Ludvigsen C. Using digital hearing aids to visualize real-life effects of signal processing. Hear Jour. 2004;57(4):40-49.