記住我
Objective: To study the Influence of Auditory Training on acceptable noise level (ANL) scores in elderly persons with hearing impairment. Design: Quasi-experimental study design. Study sample: A total of 20 bilateral mild to moderately severe sensorineural hearing loss participants with “high” ANL scores were taken into the study and randomly allocated to experimental and control groups. In the time frame, the experimental group provided 12 sessions of speech in noise training with a hearing aid and the baseline measures were repeated in both groups. Results: The Acceptable noise level and Speech in Noise scores significantly improved post-training only in the experimental group. They also showed a significant difference “Client Oriented Scale of Improvement (COSI)” scale in the domain “Conversation in Noise”. Conclusions: Acceptable noise level is susceptible to training similar to that of speech in noise score. It provides hope to the individuals who are poor candidates to the hearing aids.
Keywords: Acoustic stimulation, hearing aids, hearing impairment, questionnaire, signal to noise ratio, speech perception
How to cite this article:
Hearing loss is the most common sensory deficit among the aged population. Age-related hearing loss (ARHL) is known as presbycusis[1] and is characterized by inaudibility, difficulty to discriminate and identify sounds, inability to sustain a conversation in challenging situations such as speech babble or background noise.[2] Assessment of speech perception in noise and acceptance of background noise level is crucial in such population that plays a vital role in diagnosis, rehabilitation as well as prognostic indicators. Speech in noise and acceptable noise levels are one example of that.
Speech in noise is a monoaural low redundancy speech test developed by Killion et al.[3] It identifies the individual’s neural processing abilities related to speech perception particularly in the aged population as well as hearing aid candidacy evaluation. Literature reports that individuals with reduced Speech in noise performance are considered to be poor hearing aid beneficiaries and candidates.[4],[5] Speech in noise test also assists in predicting an individual’s performance in the background noises, assessing the required signal-to-noise ratios, selection of hearing aid technology, and hearing aid programming. Speech in noise is also used in the auditory training program to improve the performance of hearing aid users.[6] Sweetow and Sabes[6] reported that training on a speech-in-babble task, the participants performed better in a Speech in Noise task compared to their pre-training, which might bring some changes in hearing aid benefits.
Speech perception is of prime importance in hearing aid acceptance, usage specifically in the background noise. Individuals with hearing impairment either have to identify the signal in the presence of background noise or withstand the maximum noise for better performance. In these aspects, Nabelek et al.[7] developed the acceptable noise level (ANL) test that intended to address the individual’s performance in the background noise and predicts the hearing aid benefit by categorizing score into three levels: low (<7 dB), mid (7–13 dB), and high (>13 dB). Individuals with “low” ANLs are likely to become successful hearing aid user, whereas an individual who has “high” ANLs are unlikely to be successful hearing aid users. People with “mid” ANLs may or may not be successful hearing aid users.[8]
ANL test differs from Speech in Noise test mainly in terms of mechanism and it is least sensitive to certain audiological and non-audiological factors such as age, gender, loudness tolerance, hearing aid experience,[7],[9] degree of hearing loss,[7],[10],[11] unaided and aided conditions,[8],[12] types of noises,[7] monaural versus binaural performance,[13] medications, attention deficits, speech presentation level,[10],[14],[15] and technologies involved in hearing aid[12],[16],[17] except music, where it was better.[18]
With regard to training, unlike Speech in Noise, ANL lacks research evidence of its influence. There exists one recent study[19] that attempted systematic desensitization training on acceptable noise levels that facilitated an individual’s ability in accepting a higher level of background noise. However, a systematic desensitization program is different from that of the formal auditory training procedure. Hence, there arises a need to study whether auditory training can influence ANL score and change its sensitivity. Further, a study in this domain can change the audiologist’s perspectives on ANL, hearing aid selection, and counseling. Finally, the research findings can be a gateway to future research.
MethodsResearch design
A quasi-experimental, pre- and post-test study design was used for the present study. Ethical clearance was obtained from the Institutional Ethics Committee, Kasturba Hospital of the Manipal Academy of Higher Education, Manipal with reference number IEC 203/2018. In addition, the study was registered at Clinical Trial Registration India with the following reference number: (CTRI/2018/07/015099). All the participants signed the written consent form before taking part in the study.
Participants
Participants were 17 males and three females whose age ranged from 55 to 73 years with Mean[X] =62.8, Standard deviation [SD] = 4.08 years. All the participants had bilateral symmetrical mild to moderately severe sensorineural hearing impairment and baseline ANL score >15 dBHL. These participants were fitted with hearing aids and randomly assigned to the control group that received no treatment and the experimental group that underwent listening training. The experimental (trained) group was composed of 10 participants (8 males and 2 female) with Mean age [X] =61.6, Standard deviation [SD] =4.08 years and control group (9 males and 1 female) with Mean age [X] = 62.1, Standard deviation [SD] = 4.58 years. Both the experimental and control group participants satisfied the following inclusion criteria: all were native Kannada speakers and proficient speakers of Kannada language, degree of loss not exceeding moderate category, normal middle ear functioning, no neurological, motor, and mental health anomalies or associated syndromes that may hinder understanding and realization of the proposed tasks. Also, all participants passed the Montreal Cognitive Assessment screening test with scores of above 26 indicating intact cognitive abilities. [Table 1] displays the demographic details of all the participants. The experimental group received 12 sessions of listening (Speech in Noise) training between the pre- and post-training sessions with each session lasting 45 min. The control group was tested at the time interval equivalent to the experimental group, that is, 12 days interval between baseline testing and testing 2.
Table 1 Displays the demographic details, speech identification scores, and tympanometric results of the participantsProcedure
The experiment took place in two phases. The first phase involved the measurement of participants’ hearing thresholds, baseline ANL, and Speech in Noise (Kannada) score, and randomizing the participants into control and experimental group; the second phase involved the auditory (listening) training and post-training ANL and Speech in Noise evaluations. All the evaluations were conducted in an acoustically treated two-room setup with the ambient noise levels maintained well within the permissible levels .
Phase 1
Hearing evaluation
We initially examined all the participants’ ear canals for the wax and any foreign body that can influence the test result. Later, using Madsen Austra clinical audiometer and adapting modified Hughson–Westlake procedure,[20] the participants hearing thresholds were established in the audiometric frequencies [(AC 250 Hz to 8 kHz) and (BC 250 Hz to 4 kHz)]. Subsequently, the speech recognition threshold (SRT) and speech discrimination (SDS) scores were obtained using Kannada spondee words[21] and monosyllabic word list[22] successively at 20 and 40 dBSL. Finally, the UCL level was determined. Using Grason–Stadler (GSI) Tympstar version 2, the participant’s middle ear and acoustic reflexes were obtained. Followed by participants’ speech-in-noise and ANL scores were established. [Table 1] displays participants’ speech audiometric and tympanometric results of right and left ears and [Table 2] displays the participants’ pure tone audiometric test results.
Table 2 Displays the participants’ right and left ear pure tone audiometric thresholds in terms of its mean and standard deviationSpeech in Noise test
To establish Speech in Noise scores, we used material “Speech in Noise sentences in Kannada” developed by Avinash et al.[23] It consisted of a total of seven sentence lists and each list has seven sentences with five keywords in each of the sentences. From the list, six sentences were presented to the participants at their most comfortable level in the presence of speech-weighted noise by varying the signal-to-nose ratio from 25 to 0 dB randomly. In response, the participants were asked to repeat as they hear the sentences. Every correct keyword repetition assigned a score of 1 and SNR was calculated using the formula: SNR Loss = 25.5 − total number of words correct (dB).
Acceptable noise level scores
The participants were made to sit in the free field setup with a 0-degree azimuth to the speaker. The procedure began by determining the MCL by varying the cold running speech from soft to loud and vice versa until they find a suitable point that is most comfortable to them. While continuing the cold running speech at the MCL, the speech weighted background noise level (BNL) was introduced and adjusted in 5 dB steps until the participant is willing to “put up with” the background noise while listening to and following the speech. Finally, BNL minus MCL determines the ANL scores.
Hearing instrument selection and fitting
All the participants were fitted with WDRC (Wide Dynamic Range Compression) hearing aids of Starkey, Oticon, Bernafon, and Signia Company starting from four channels and above. Hearing aids were programmed according to the audiogram and activated all the necessary features enhancing signal-to-noise ratios. Later, insertion gain was performed using Audioscan Verifit 2 real-ear measurement instrument with the target gain placed close to 55, 65, and 80 dB input levels. Further, the participants were subjected to a listening check, where the listeners were assessed for the comfort of listening to sounds outside laboratory settings and the quality of listening. Later, all the participants filled the Client Oriented Scale of Improvement (COSI) developed by NAL (1997). It consists of five domains such as conversation in quiet, conversation in the noise, television, telephone, and increase in social contact. The participants were instructed to rate the hearing aids experience in terms of worse, no difference, slightly better, better, and much better wearing a hearing aid. Finally, the participants were made to choose the hearing that benefitted most and was comfortable.
Participant’s allocation
Chit-pull system was used to allocate participants into the control and the experimental groups. In this procedure, a number of chits representing the experimental and the control group were first placed in a container. All the selected participants drew chit from the container and were assigned to the group represented in the chit. Further, the experimental group underwent auditory (listening) training in phase 2 whereas the control group received no remediation.
Phase 2
Auditory (listening) training
The auditory training task involved listening in speech-in-noise. All the participants received 12 sessions, 1 hour per day, within 2 weeks.
Stimuli used
We used monosyllables, general words, and sentences for training. The general words list contained the commonly used words in day-to-day life and words picked from newspaper/story books/magazines). The sentences consisted of sentences developed for the quick speech-in-noise test by Avinash et al.[23] It had a seven-sentence list. Each list had seven sentences, making a total of 49 sentences. For training, we considered sentences from lists 2 to 7.
Signal-to-noise ratio
We used SNR varying from 0 to 25 dB.
Procedure
The participants wore their hearing aid and sat in a free-field setup, a double-wall sound-treated room. Via audiometer, we streamed the signal, and a speech-weighted noise through the speaker with SNR varied from 25 to 0 dB with 5 dB step size. In initial sessions, participants received training with monosyllables, words, and sentences with favorable SNRs to equally compete, and as the training progressed, we randomized the SNRs and stimuli. The training session did not follow any performance criteria to progress to more difficult SNR. In addition to training, all the participants received a home assignment to use the hearing aid in adverse listening situations at home and work environment.
After completing the training sessions, we repeated participants’ ANL and SPIN scores and COSI questionnaires to compare baseline. Similarly, the control group also received ANL and Speech in Noise tests and COSI with equivalent time intervals in a set time frame.
Descriptively the mean and standard deviation for the ANL and Speech in Noise raw scores were calculated. Repeated measures of ANOVA were used to analyze (1) Within-subject effect (comparison between pre and post), (2) Between the subject effect (between the groups), and also to check (3) interaction effect.
ResultsAcceptable Noise Level Scores
[Table 3] shows the mean and standard deviation of ANL and SPIN scores in the unaided and aided conditions. Paired t-test showed no significant difference between unaided and aided ANL scores with P-value = 0.08 whereas the unaided and aided Speech in Noise scores showed a statistically significant difference with P-value < 0.05.
Table 3 Unaided and aided Acceptable Noise Level and Speech in Noise Scores (SNR loss)[Table 4] presents the aided pre- and post-mean ANL scores post-intervention. To test within the subject effects on ANL scores, repeated measures of ANOVA was adopted. Mauchly’s test indicated that the assumption of sphericity was violated (P < 0.05), therefore degrees of freedom were corrected using Greenhouse–Geisser estimates of sphericity (ε = 0.001). The experimental group showed a significant difference between pre- and post-ANL scores with value F (1.0, 90.0) = 14.400, P = 0.001. Comparison between the subjects, the pre- and post-ANL scores revealed that the mean differences of ANL scores were not statistically significant (F (1, 62.500) = 2.153, P = 0.160). Overall, the change in ANL scores from pre to post is different in both experimental and control groups is significant, F (1, 122.50) =19.600, P ≤ 0.05.
Speech in Noise (Kannada) Scores
[Table 5] presents the pre- and post-evaluation Speech in Noise (Kannada) scores of the experimental group and control groups. It is observed that the post-training SNR scores improved from 14.20 ± 4.58 to 9.30 ± 2.53 in the experimental group showing an improvement of 4.9 dB. In the control group, the values changed from 14.80 ± 2.34 to 14.40 with an improvement of only 0.4 dB.
Table 5 Aided pre and post-training mean Speech in Noise (Kannada) scores (SNR loss)Repeated measures of ANOVA were adapted to check the significant differences and Mauchly’s test indicated that the assumption of sphericity had been violated (P < 0.05), therefore degrees of freedom were corrected using Greenhouse–Geisser estimates of sphericity (ε = 0.000). The results show that there was a significant difference between pre- and post-speech in noise (Kannada) scores in the experimental group, F (1.0, 70.225) = 80.770, P ≤ 0.005). Between the subjects, the pre and post-scores revealed no significant differences with value (F (1, 81.225) =3.807, p=.067). Though the Speech in Noise (Kannada) scores differed from pre to post in both groups, in the experimental group the results are significant with value F (1, 50.625) = 58.227, P ≤ 0.05).
Subjective satisfactory quality rating using COSI questionnaire
[Table 6] shows the pre- and post-subjective satisfactory scores, across the domains indicating “Change” versus “No Change” were analyzed for the significance. The conversation in noise domain, in the experimental group, 80% of them reported changes and 20% no changes in performance. Similarly, in the control group, only 20% reported changes. The odds ratio with 95% confidence interval showed a significant difference with a value of 16 (1.77, 143.15). Hence, the experimental group participants are 16 times more likely to improve their ability to converse in noise than when compared to the control group. No statistically significant changes were observed in other domains of COSI.
Table 6 Pre- and post-subjective satisfactory score using COSI questionnaire Discussion ANL scores with and without hearing aid
ANL is a test that predicts hearing aid candidacy. The individuals with a higher ANL score (>13 dBHL) are less likely to benefit from hearing aid.[8] We witnessed a higher ANL score (15 dBHL) by all the participants both in unaided and aided conditions, further exhibiting no significant difference in performance between the groups. The higher ANL scores are attributed to participants’ age and duration of hearing loss influencing peripheral and central auditory mechanisms. The above findings suggested all the participants are poorer candidates for hearing aids, which did not change upon getting a hearing aid.[8],[12] However, the studies differed in the participant’s age range, which showed a negative relationship. Additionally, research also evidences the uncorrelation among age, unaided, and aided scores.[8],[24] Overall, higher the ANL scores, poorer the hearing aid performance, especially in noisy situations irrespective of participant’s age and hearing aids, which calls for a different line of approach such as auditory training.[25]
Effect of auditory training (Speech in noise training) on acceptable noise level scores
Auditory training is an intervention method that facilitates the residual auditory skill to enhance the communication ability of an individual with hearing impairment.[26] Speech in Noise training is one of the training components in auditory training adapted to enhance speech perception in noise. In the current study, the experimental group received Speech in Noise training and showed a significant reduction of 6 to 7 dB in ANL score compared to the control group, despite the previous literature demonstrating no correlation between speech recognition in noise abilities on ANL.[27] The above findings suggest that listening to speech in the presence of noise at variable SNR has a positive, influence in tolerating higher background noise levels. The findings are in agreement with Gordon-Hickey and Morlas[28] who observed a significant relationship between the signal-to-noise ratio and ANL.
The possible reason for Speech in Noise training interacting with ANL can be attributed to “OPERA” hypothesis, which proposes adaptive plasticity in speech processing network: anatomical overlap in the brain networks that process an acoustic feature used in both music and speech and attention activity that engage the neural network are associated with focussed attention.[29] Though the current study observed a significant difference in ANL scores by 6–7 dB post-speech in noise training, the scores still fell short by 3 dB to meet candidacy criteria. Research reports, to observe adequate benefits from auditory training, the duration of training should be provided at least two to three times per week for 5 to 15 weeks.[30] Hence, this might be one of the possibilities of falling short in scores.
Speech in noise (Kannada)
QuickSIN is an excellent tool for demonstrating hearing aid benefits.[31] In the current study, the aided Speech in Noise (Kannada) scores were less when compared to the unaided score.[8] Nebelek et al. also observed similar findings in the group of Full-time users, Part-time users, and Non-users—the mean unaided and aided Speech in Noise (Kannada) scores were different in all the groups. The result is also in line with the literature[5],[32] where they noted Speech in Noise (Kannada) scores were significantly better in the aided condition when compared to the unaided condition.
Effect of auditory training (speech in noise training) on Speech in Noise (Kannada) scores
The auditory training yielded positive results to the experimental group by significantly improving the speech in noise score from 14.20 ± 4.58 to 9.30 ± 2.53 in comparison to the control group. Research reports, obtaining 1 dB improvement in SNR has been estimated to result in 6% to 8% improvement in speech identification scores,[33],[34] the present study witnessed 5 to 6 dB improvement in SNR loss. The findings are in consensus with the literature,[35],[36],[37],[38],[39] wherein they reported significant improvement with training in the presence of noise among older adults with hearing impairment. Similarly, Rao et al.[40] and Anderson et al.[35] also reported Speech in Noise training improved HINT and QuickSIN scores. In summary, exercising listening speech in the presence of background noise is of great benefit to individuals with hearing impairment, especially for those who complain of it.
Subjective satisfactory quality rating using COSI
COSI is a clinical tool that documents the client’s improvements in hearing ability post-hearing aid fitting.[41] In the current study, it was adapted to measure satisfactory scores pre- and post-speech in noise training with a hearing aid. Post-training the participants reported improvement in all of the domains, however only the “Conversation in Noise” showed a significant difference in performance. Hence, speech in noise training showed a good correlation with improvement in “Conversation in Noise” domain of COSI.[42]
Even though there was a significant improvement in the “Conversation in Noise” domain, it did not significantly change the “Social Contact.” Social contact varies from individual to individual, and it depends on their lifestyle, which might have influenced the result of social contact. Social contact is also influenced by individual assertiveness; hence, working on building assertiveness along with auditory closure might influence the finding.
The current study findings are also in line with a self-reported questionnaire indicating the efficacy of the training among older adults with hearing impairment.[6] A study by Gil and Iorio[36] evaluated the subjective effectiveness based on formal auditory training using APHAB with various domains including ease of communication, background noise, aversiveness of sound, and reverberation condition. The outcome uncovered that there was a pattern toward statistical significance for the experimental group in the reverberation and background noise domain, whereas there was no difference noted in the control group. Hence, they concluded that formal auditory training has a significant effect on the subjective satisfactory score.
Strength
This is the first kind of study in the clinical population and it provides evidence that ANL is susceptible to training.
Clinical implication
The auditory training has an influence on ANL.Speech in noise training; a component of auditory training can change the ANL candidacy.Limitation
Not adapting cross-check principles mainly in terms of objective evaluation (electrophysiological tests).Use of only auditory closure activities for training.Follow-up evaluation to test the sustained ability of training provided.Future directions
To evaluate the effect of other standardized auditory training/ music training on ANL in elderly persons with hearing impairment.Generation of objective evidence to evaluate the relation between ANL and auditory training. ConclusionTo conclude, Speech in Noise training at favorable and unfavorable SNR showed a positive influence on ANL, Speech in Noise (Kannada), and COSI (Conversation in Noise). Though both tests differ in their mechanism, training one can influence the other. However, further research is warranted to justify the findings more objectively. Further, hearing aid fitting and a few sessions of auditory training can be of greater support for the hearing impaired population, especially the older adults.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
Correspondence Address:
Greeshma R
Ayyanath (H), Palliyara, Trikkur (P. O.), Thrissur 680306
India
Kishan M Mohan
Department of Speech and Hearing, Manipal College of Health Professionals, MAHE, Manipal 576104
India
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/nah.nah_5_22
留言 (0)