The inability to understand speech in the presence of noise continues to be a significant problem for most hearing instrument wearers. This problem is exacerbated as the degree of hearing loss increases. One of the most effective strategies for improved understanding in noise is to provide a better signal-to-noise ratio (S/N). An improved S:N ratio can be accomplished by use of a directional microphone system. For people with severe and profound hearing loss, however, the advantage of an ear-level directional hearing aid may not be sufficient to counteract the deficit created by noisy and reverberant environments. Multiple-microphone (>2) arrays may be a better design alternative. These devices typically have higher directivity and provide greater improvement in the signal-to-noise ratio for individuals with severe to profound hearing loss.
A multiple microphone array was developed as a collaborative effort between Starkey Laboratories and Cardinal Sound Labs, Inc. It is based upon the work of Stanford University Professor Dr. Bernard Widrow and described in the Widrow-Brearly Patent, 'Directional Hearing Aid' (U.S. Patent number 4,751,738). The Radiant Beam Array (RBA) is a broadfire, multi-microphone array incorporating six spatially separated microphones. (Figure 1) The array is worn around the neck with the microphones strategically positioned across a horizontal pattern on the chest. Output from the RBA is transmitted via a wireless link (induction) to hearing instruments or CIC modules incorporating a telecoil.
Figure 1. Picture of the Radiant Beam Array with dimensions.
There are four user adjustable parameters that may be altered by using push button controls on the front of the RBA: memory, volume, kneepoint, and beamwidth. There are four available memory settings that provide frequency shaping to accommodate various listening situations. The volume control allows the overall gain to be adjusted by 20 dB in 2 dB steps. The kneepoint and beamwidth adjustments are the most important parameters because they are used to reduce environmental noise and improve directionality. The compression kneepoint ranges from 60-80 dB SPL in 2 dB steps as illustrated in Figure 2. Raising the compression kneepoint decreases the gain for inputs below the kneepoint and the inverse is true when the kneepoint is lowered.
Figure 2. Typical Input-Output function of the Radiant Beam Array with the green curve representing a 60 dB kneepoint and the red curve representing an 80 dB kneepoint.
The beamwidth as displayed in Figure 3, refers to the pattern of the directional response and may be adjusted to a wide, windy, or sharp pattern. The wide beamwidth is essentially omnidirectional. Adjusting the beamwidth from wide to sharp will emphasize sound in the forward direction 10-30 dB relative to sound originating from behind and to the side of the user.
The measured Directivity Index (DI) of the RBA in the sharp beamwidth pattern ranges from 8.1 to 10.7 dB, depending on the frequency measured, This value is significantly higher than ear-level directional hearing aids that have measured directivity indices typically ranging from 2-6 dB.
Figure 3. Sharp, Windy, and Wide directional beamwidth patterns.
The purpose of this study was to evaluate the performance of the RBA with behind-the ear hearing instruments. The three main objectives of the study were to:
Ten adults with flat or sloping, moderately severe to profound bilateral sensorineural hearing loss participated in the study. Figure 4 displays the mean right and left audiograms for the subjects. Three subjects had asymmetric hearing losses. All subjects were pre-existing hearing instrument users.
Three of the subjects were wearing power BTE hearing aids binaurally. Two subjects were wearing power, directional BTE hearing aids (one monaurally and one binaurally). Two subjects were wearing binaural ITE and ITC instruments respectively and three of the subjects were wearing binaural CIC instruments. All subjects utilizing BTE instruments were fit with the RBA in addition to their existing hearing aids. The subjects utilizing custom hearing aids were fit binaurally with new BTEs in conjunction with the RBA.
Probe Microphone Measurements:
Probe microphone measurements (Frye 6500) were performed for the hearing aid only and the hearing aid plus the RBA. All hearing devices were set in the omnidirectional mode when applicable. Hearing aid only measurements were always taken with the reference microphone positioned at ear level.The RBA was adjusted during the probe microphone evaluation to provide a transparent frequency response for the hearing aid only and the RBA paired with a BTE. Typically, the volume of the hearing aid needed to be turned down by approximately 10 dB once the RBA was fit. The volume was adjusted until the RBA probe microphone response closely matched the hearing aid only response to ensure evaluations of the RBA would be based on directional benefit and not on additional gain. All probe microphone measurements with the RBA were performed with the device in the wide beamwidth pattern and the probe system's reference microphone placed directly over the center RBA microphone.
During the initial visit, the increase in available gain before feedback was evaluated to compare the hearing aid alone to the RBA system. Probe microphone measurements were obtained to determine the point of audible and/or visible feedback. The audiologist listened to the output of the probe microphone via headphones and used subject report to determine the point of audible feedback. Visible feedback was defined as an increase of>3dB at the peak frequency where feedback was present. The same measurements were taken for the RBA working in conjunction with the hearing aid. The volume was increased on the RBA to maximum or until feedback occurred. If the maximum volume on the RBA did not elicit feedback, then an alternate memory with higher gain was selected. If feedback still did not occur, the hearing aid volume was increased either to maximum or until feedback occurred.
Following probe microphone verification, each subject was given an extensive orientation to the RBA system, along with a daily log for documentation of his or her listening experiences. The subjects were requested to use the RBA system for two weeks. All subjects were contacted within the first week of use for follow-up.
Hearing In Noise Test (HINT)
The HINT (Nilsson et al, 1994) was used to assess the directional effects of the RBA on speech reception thresholds for sentences in quiet and noise. This test requires the user to repeat sentences in the presence of speech-weighted noise that is fixed at one level. The HINT stimuli consist of 20 sentences for each condition. The RBA and hearing aids were adjusted to match the settings for the probe microphone measures.
The HINT was performed in the soundfield under the following test conditions: unaided, hearing aid only, and RBA system in wide, windy, and sharp beamwidth mode. Measurements were taken with speech and competing noise both at 0°, speech at 0° and competing noise at 90°, speech at 0° and competing noise at 270° and finally, speech at 0° and competing noise at 180°.
Abbreviated Profile of Hearing Aid Benefit (APHAB):
The Abbreviated Profile of Hearing Aid Benefit (Cox and Alexander, 1995) is a 24-item questionnaire whereby data is analyzed using four subscales. The subscales are: ease of communication (EC), reverberation (RV), background noise (BN) and aversiveness (AV). Aversiveness refers to the unpleasantness of environmental sounds.
The APHAB was administered both pre- and post-fitting. Each subject completed the APHAB for 'without my hearing aid', 'with my current hearing aid', and 'with the RBA system'.
Following the two-week trial period, subjects were asked to complete a 23-item questionnaire assessing the performance of the RBA system in a variety of listening situations. Areas to be evaluated included sound quality, loudness, clothing noise, electromagnetic interference, and overall performance. The final item in the questionnaire allowed subjects to indicate their preference between the RBA system or the hearing aids alone. A 9-item comparison questionnaire was also administered to address this issue. Subjects were asked to directly compare the RBA system to the BTE hearing aids alone and their own hearing aids, if applicable.
Probe microphone evaluation showed an increase in REIG with the RBA versus the hearing aid alone for all subjects. Out of a total of 19 ears, three ears had feedback with the hearing aid only, but no feedback occurred with the RBA and hearing aid both set at maximum volume. Two ears had no feedback with the hearing aid at maximum volume, but feedback did occur with the RBA as a result of the increased gain. The remaining ears demonstrated no feedback in either condition, but there was still an increase in gain with the RBA system, especially in the low to mid frequencies.
The gain increase provided by the RBA varied greatly across frequencies and subjects. Calibrating the probe microphone system was difficult due to the body-baffle effect. Therefore, quantifying the additional gain provided by the RBA was problematic. According to Nichols et al. (1947) body-baffle can result in a decrease of up to 20 dB in the mid to high frequency range (800- 2000 Hz). The decrease in gain is caused by wearing the hearing aid microphone on the body or specifically in this case, at the center of the chest. The external reference microphone of the probe system was placed over the center microphone of the RBA and leveled. The reference microphone was active during the measurements; that is, the output of the loudspeaker was adjusted in real-time by the level measured at the reference microphone This placement resulted in a lower measured SPL with the largest effect occurring at 2000 Hz. Therefore, the measured increase in gain provided by the RBA was lower than what was expected. Further investigation following the clinical trials, determined that the body baffle effect could be minimized by tilting the reference microphone slightly forward.
Importantly, as pointed out by Nichols and colleagues (1947), body-baffle effects are simply a measurement phenomenon. In a diffuse soundfield or a reverberant environment, the body-baffle effects diminish resulting in a negligible effect on speech recognition. The speech recognition performance and the subject's assessment of sound quality confirm this conclusion.
Data is presented for the six subjects able to complete the HINT. None of the subjects could complete the HINT in the unaided condition. Two subjects could not complete the aided 0° azimuth condition. Due to the degree of hearing loss combined with their low discrimination scores, four subjects were unable to accurately repeat the first sentence of the HINT in any of the test conditions. Therefore, HINT scores could not be measured for these four subjects.
Figures 5, 6, and 7 show the HINT scores for the subjects' BTE and for the RBA in sharp mode at 180, ±90 and 0 degrees. All subjects performed the best with the RBA in the sharp beamwidth mode across all test conditions. The Windy beamwidth mode also resulted in better S:R scores than any of the omnidirectional schemes. Poorer performance was observed in the omnidirectional mode for the hearing aid alone and the RBA in wide mode. As would be expected, a smaller difference in performance was seen between the hearing aids and RBA in the 0° azimuth condition. The greatest improvement in the SNR scores occurred with speech at 0° azimuth and competing noise at 180° azimuth with the RBA in the sharp beamwidth pattern where the mean was -7.5 dB. At first glance, the HINT results may not appear to be significant. However, the magnitude of the improvement is impressive considering that all subjects had moderately severe to profound hearing losses bilaterally. Figure 8 shows the mean HINT results at 180 degrees for each aided configuration.
Nine subjects showed benefit across at least one APHAB subscale demonstrating subjective improvement when wearing the RBA system as compared to their own hearing aids. Individual APHAB benefit scores are plotted in figure 9, showing the perceived improvement of the RBA as compared to the hearing aids only.
Seven subjects had APHAB scores showing a significant preference for the RBA. Specifically, five subjects had benefit scores of 5 points or greater across the three communication subscales (EC, RV, and BN); one subject had one communication sub-scale (EV) score of 23; one subject had an aversiveness (AV) score of 44, demonstrating improvement over the hearing aid alone. According to the APHAB guidelines (Cox, 1997), this data indicates a 'real difference' between the RBA system and the subjects' own hearing aids. Only one subject (subject 10) did not show preference for the RBA across any of the APHAB subscales.
Figure 10 shows the mean APHAB benefit score for the RBA as compared to the unaided condition.
Figures 11 and 12 show the mean APHAB benefit scores for the RBA system as compared to the hearing aid only. The APHAB mean benefit scores for the RBA system exceeded the scores for the subjects' own hearing instruments across the group. The most significant improvement was for the background noise communication subscale with a mean score of 12.9 and the reverberation subscale with a mean score of 10.6.
The daily log comments were very positive. Some examples are listed below:'Kids at the park playing, I can hear the crack at the ball, cool!' (Male, age 44, binaural power BTE user and uses sign language as predominant method of communication)
'Noticed that I can hear a pencil rolling off my desk, new and strange noises.' (Female, age 40, binaural power BTE user and uses sign language as predominant method of communication)
'I wore the array with the hearing aids, talking with my mother and went to the show. I caught every word, worked like a charm.' (Male, Age 70, binaural ITE user)
'At work, heard people laughing way down the hallway, I have never heard from such a great distance before!' (Female, Age 47, monaural power BTE user)
Overall perceived benefit with the RBA was very good. Half of the subjects reported that they would 'Always' choose the RBA system over their own hearing aids. The subjects that chose 'No Preference or 'Never' were asked to explain their answer. Their responses were mainly influenced by cosmetic factors (half of the subjects entered the study with CIC or ITC instruments) and getting the RBA properly adjusted. Of particular interest were the results of the comparison questionnaire. A higher rating was reported for use of the RBA in moderate and loud background noise by half of the subjects with four subjects having 'No Preference'. Half of the subjects also reported a preference for the RBA when in small groups or listening to a speaker in a large group.
Overall there was higher satisfaction with the RBA in more difficult listening situations. It should be noted that even though subjects felt they were hearing better, most of the subjects opted against wearing the device on a full-time basis and in some cases even a part-time basis due to its size.
As stated previously, one of our objectives was to evaluate the performance of the RBA when worn under clothing. Four subjects in the group reported that they 'Always' wore the RBA under clothing. Three subjects reported that they 'Sometimes' wore the RBA under clothing. Results revealed clothing noise to be problematic some of the time. Half of the subjects were 'Sometimes' able to make adjustments to decrease or eliminate clothing noise. However, for the five subjects that either were 'Neutral' or would 'Never' choose the RBA over their own hearing aids, clothing noise was not one of the contributing factors. Clothing noise was also not a persistent problem on the daily log sheets. A bigger problem appeared to be learning to adjust the RBA while wearing the device under clothing because the user was unable to see the controlling push buttons.
The environmental electromagnetic interference questionnaire was administered with the sound quality questionnaire. The results show that most subjects had mild to moderate interference in various environments. The overall effect appeared to be minimal. Subjects did report that the interference varied depending on the volume setting of the hearing aid. Therefore, it is recommended to adjust the hearing aid volume or gain as low as possible and set the RBA volume as high as possible and/or use a memory with the least amount of attenuation to match target and remain at a comfortable level.
The RBA is an effective directional solution for people with moderately severe to profound hearing loss and provides an advantage for speech understanding in noisy and reverberant environments inclusive of moderate to loud background noise.
1. The RBA directional hearing system improved the HINT S:R score in noise as compared to the unaided and hearing aid alone condition, providing an advantage in understanding speech in the presence of background noise.
2. The overall impression of the RBA was very good, but the user must be willing to forgo more cosmetically appealing smaller hearing instruments in order to receive the many acoustic benefits provided by the RBA.
3. APHAB scores demonstrated over half the subjects perceived significant improvement wearing the RBA versus their own hearing aids. All but one subject had improved benefit scores for at least one of the APHAB subscales. The background noise communication subscale received the highest mean benefit score.
4. Increased insertion gain provided by the RBA may resolve persistent feedback issues with the hearing aid and/or provide additional gain when needed.
5. Product orientation for the user controls is critical for success, especially for the beam width and kneepoint adjustments.
6. The ability to customize the fitting for each individual should greatly contribute to the success of the product. The programming software currently is in development.
Cox R, Alexander G. (1995) The Abbreviated Profile of Hearing Aid Benefit (APHAB). Ear and Hearing, 16(2): 176-186.
Cox R. (1997) Administration and application of the APHAB. Hearing Journal, 50(4): 34-45.
Nilsson M, Soli SD, Sullivan J. (1994). Development of a hearing in noise test for the measurement of speech reception threshold. Journal of the Acoustical Society of America, 95:1085-1099.
Nichols RH, Marquis RJ, Wiklund WG, Filler AS, Hudgins CV, Peterson AS. (1947) The influence of body-baffle effects on the performance of hearing aids. Journal of the Acoustical Society of America, 19(6): 943-951.