2-Pane Combined
Full Summaries Sorted

Effect of systematic desensitization training on acceptable noise levels in adults with normal hearing sensitivity


Accurate speech perception relies on the ability of the auditory system to perceive the speaker’s voice in the presence of background noise. Even young adults with normal hearing and normal cognitive abilities have difficulty in perceiving speech in the presence of irrelevant background information.[ 1],[ 2] This process is even more challenging for children, geriatrics, and also for individuals with hearing impairment.[ 3],[ 4],[ 5] Speech perception abilities in noise can be quantified in several ways. One of the ways is to measure the speech intelligibility in the presence of background noise, and the other way is to measure the willingness of a person to accept noise while listening to speech. The willingness of a person to accept noise while listening to speech can be measured using the acceptable noise level (ANL) test.[ 6] The ANL was originally termed as tolerated signal-to-noise ratio, and it is determined by estimating the maximum amount of background noise that the listeners can accept while attending to target speech. To determine ANL, the listener is presented with a running passage, and she/he is instructed to adjust the level of the passage to her/his most comfortable level (MCL). Then the background noise is introduced, and the listener adjusts the level of the noise to their maximum acceptable level [background noise level (BNL)] while attending to the running passage. ANL is derived by subtracting the BNL from MCL. Individuals who accept high levels of background noise have small ANLs and individuals who accept low levels of background noise have large ANLs.

ANL is a subjective measure which is not directly comparable to a listener’s objective ability to recognize speech in the presence of noise. Few studies have reported that there was no correlation between ANL and speech recognition ability in noise.[ 7],[ 8] Also ANL is not related to age,[ 6],[ 9] gender,[ 10] hearing sensitivity,[ 9] type of background noise distraction,[ 11] middle ear characteristics,[ 12] cochlear responses, and efferent activity of the medial olivocochlear bundle pathway.[ 12],[ 13] However, ANL is related to the presentation level of the target stimulus.[ 14] Even though the presentation level of the target stimulus has an impact on ANL, it is advisable to measure ANL at the MCL, because ANL is primarily used as a prefitting measure for hearing aid, and hearing aid fitting strategies attempt to amplify speech signal to the MCL (e.g., National Acoustic Laboratories’ nonlinear fitting procedure, version 1 [NAL-NL1], Desired Sensation Level [DSL]). ANL-based model for the prediction of hearing aid use has been the research focus in recent years.[ 15]

For clinical use, ANL has been adapted in different languages such as Swedish,[ 16] Mandarin,[ 17] and Danish.[ 16],[ 18] Few attempts were made to assess the utility of “International Speech Test Signal (ISTS)” as the common test stimulus for different languages.[ 16],[ 19] Brännström et al.[ 16] found it advantageous to use ISTS as the common test stimulus. However, the results of Ho et al.[ 19] are not encouraging. It has been shown that if common instructions are used, ANL can be reliably measured across testers, laboratories, and clinics, because it has high inter tester reliability.[ 20] Nabelek et al.[ 7] also reported that ANL is reliable and does not change with the acclimatization to hearing aids in listeners with impaired hearing, in aided, as well as in un-aided conditions. However, Olsen and Brännström[ 15] stated that instructions given may affect ANL and recommended to consider the coefficient of repeatability and minimal clinically important difference while determining the repeatability of ANLs.

In today’s clinical practice, ANL is one of the essential components in “Functional Communication Assessment (FCA)”[ 21] which is used for checking an individual’s communication abilities before and after hearing aid fitting. ANL is important for its various advantages such as verification of hearing aid features and prediction of hearing aid success. Unaided ANL can predict the hearing aid use with 85% accuracy. Specifically, individuals with low ANL are likely to benefit from the hearing aid more than the individuals with high ANL, who may become unsuccessful hearing aid users.[ 9] On the basis of these results, Nabelek et al.[ 9] suggested that ANL measured prior to hearing aid fitting may serve as a predictor of success of hearing aid fitting user, and it can be used as a counseling tool. Similarly, Ho et al.[ 22] reported that ANL could predict real-world success with hearing aids.

Individuals with high ANL are unlikely to become successful hearing aid users. Hence, attempts have been made to improve ANLs of individuals with the help of digital technologies. Technological solutions available in hearing aids such as directional microphones and digital noise reduction (DNR) algorithms improve the ability of an individual to tolerate background noise.[ 8],[ 23],[ 24],[ 25] Using DNR algorithms along with directional microphones significantly improves ANL when compared to using any one of them.[ 26] However, DNR algorithms have the disadvantage of reducing the intensity of the desired speech along with interfering noise because the desired speech and interfering noiseoften co-exist. Also in most real life situations, the desired speech signal and background noise are not spatially dispersed which makes the directional microphones ineffective. Wide dynamic range compression, one of the essential features in hearing aids, worsens the ANL.[ 24] Hence, there is a need for alternative management options for individuals having high ANL. It is unknown whether auditory training can be used as an alternative approach to improve ANL. Hence, this study was conducted to investigate the effect of systematic desensitization training on ANLs.

Materials and Methods


In the current study, 38 individuals (33–females and five − males) with normal hearing participated. The observational study design was adopted for the current study. All the participants were in the age range 18–25 years, and the mean age of the sample group was 22 years. Informed consent was obtained from the participants prior to their participation in the study. None of the participants had any history of recent otological or gross neurological deficits. All the participants had air and bone conduction thresholds ≤15 dB HL at octave audiometric frequencies (250, 500, 1000, 2000, 4000, and 8000 Hz). The participants had normal middle ear function as determined determined by ‘A’ type tympanogram with the presence of ipsilateral and contralateral acoustic stapedial reflexes at 500, 1000, 2000 and 4000 Hz. The study was conducted in accordance with the Declaration of Helsinki 1975.[ 27] This study was approved by the institutional ethics committee.


A digital recording of running male discourse was used as the target stimulus. The discourse used was a Kannada language story having a duration of 120 s. The story was subjected to familiarity rating by proficient Kannada speakers and then considered for the study. Four talker Kannada speech babble (two males and two females) was used as the competing stimulus. The target speech stimulus and the competing stimuli were processed to have the same root mean square value for the entire stimulus duration (120 s) and delivered to participants through a calibrated clinical audiometer (GSI-61; Grason-Stadler Inc., Minnesota). Both the stimuli were routed to a loudspeaker located at 45° azimuth on the right side.


Estimation of acceptable noise level

The procedure used to obtain ANL was similar to that of Nabelek et al.[ 6] Baseline ANL was measured for all the 38 participants. Prior to testing, each participant was given verbal instructions describing the experiment and expected task. Note that the instructions used in the current study are different from the original ANL test wherein instructions were presented both verbally and in written format. However, the participants were interviewed before the test to ensure that they understood the instructions clearly. The target stimulus was routed through channel one of the audiometer, and the competing stimulus was routed through channel two and output of both the channels were given to the loudspeaker located on the right. The intensity dial of channel one was kept constant, and intensity dial of channel two was varied to present the competing stimulus.

First, the baseline ANL was measured for all participants. To obtain ANL, MCL was established by presenting the running discourse. The discourse was presented at 0 dB HL and increased in 10 dB steps until the participant wanted the intensity of the signal to be reduced. Then, the intensity of the discourse was varied in 2 dB steps. Participants were instructed signal by indicating thumbs up or thumbs down for increasing or decreasing the stimulus level until the loudness of the discourse was most comfortable. Next, the participant’s BNLs were established. The running discourse was continuously presented at MCL while the competing stimulus, that is, the speech babble was adjusted. The same procedure that was used to find the MCL was followed here, except that the participant had to indicate the level at which they felt tired or tensed because of the babble. The maximum noise level that was most acceptable for participants without becoming tensed and tired while following the story was considered as the BNL.[ 6] Three trials of MCL and BNL were measured, and for each of them, ANL was determined by subtracting the BNL from MCL (ANL = MCL−BNL). The mean ANL was computed by averaging the ANL results of all three trials.


On the basis of the baseline ANL, participants were categorized into groups with low ANL, mid ANL, and high ANL. Those having ANL of less than 5.9 dB were categorized into the group having low ANL (n = 12), and participants with ANL greater than 10.4 dB were categorized into the group having high ANL (n = 13). Participants with ANL of ≥5.9 dB and ≤10.4 dB were in the group having mid ANL (n = 13). Systematic desensitization procedure was used for training. Participants with high ANL attended one session of training per day, and each session lasted for 20–30 min. Each participant attended ten sessions of training over a period of 2 weeks. For training, standardized sentences taken from QuickSIN[ 28] were presented at the MCL, and speech babble was presented simultaneously 2 dB below the ANL of the participant. The target stimulus was routed to channel one of the audiometer, and speech babble was routed to channel two of the audiometer, and the output of both channels was given to the loudspeaker located on the right side. The speech babble was presented continuously. If the participant was comfortable listening to sentences in the presence of the speech babble, the intensity of the babble was increased in 2 dB steps and at each step, four sentences were presented. The participants were instructed to repeat the sentences verbally. Target sentences were presented after obtaining verbal responses for the previous sentence from the participant. Feedback was given regarding the correctness of the response. Silence duration of 1 s was placed at the beginning of the sentence. The level of speech babble was increased until (i) the participant reached the point where any additional increase would produce discomfort; (ii) if the participant felt tensed or tired because of the speech babble (iii) or until the participant was unable to understand speech. During the training session, the participants were instructed to ignore the noise and concentrate on the target signal. Encouraging the participants to repeat the target would ensure that, they had concentrated on the target signal. The participants were counseled to not bother about the noise as far as understanding of the target signal is not affected.

Post-training evaluation

ANLs were measured for all participants at the end of the training period. The procedure used for measuring ANL was the same as the one used to establish the baseline ANL. Target and background noise used was the same as that of pretraining evaluation, that is, Kannada male discourse and four talker babble.


ANL data obtained from the participants were tested for normal distribution using Kolmogorov–Smirnov (K-S) test of normality. Results revealed that data are normally distributed (K-S = 0.103, P > 0.05). Histogram representing ANL values across participants is depicted in [Figure 1].{ Figure 1}

To classify ANL as low <33rd percentile criteria was considered, for mid ANL 33rd to 66th percentile criteria was considered and for high ANL >66th percentile was considered. The values for 33rd and 66th percentile were 5.9 and 10.4, respectively. The classification of ANL is represented in [Table 1]. Nabelek et al.[ 6] derived the criterion for classification for ANL which differs slightly from the above derived criterion. According to them, low ANL is from 0 to 6, mid ANL is from 6 to 14, and high ANL is above 14. These ANL values represent the ANLs in pretest session.{ Table 1}

In the current study, participants with high ANL were trained using systematic desensitization procedure. At the end of the training period, ANL was measured again for participants of all the three groups, and these ANL values represent the ANLs in post-test session. Repeated measure of analysis of variance was performed to investigate whether the ANL measured in post-test session was different from ANL measured in pretest session for all the three groups. For the analysis, group (low, mid and high ANL) served as the between subject variable and test sessions (pretest and post-test sessions) served as the within subject variable. Results revealed that there was a significant main effect of test session on ANL (F( 1, 35) = 142.56, P < 0.05) and the interaction between test sessions and the group was also significant (F( 2, 35) = 147.02, P < 0.05).

Pair wise comparison was made using separate paired-t tests to determine the difference in ANL between pretest and post-test sessions in all the three groups. Analysis revealed no significant difference between pretest and post-test ANL scores in low ANL group (t11 = −0.56, P > 0.05) and also in mid ANL group (t12 = 0.0, P > 0.05). However, there was a significant difference observed between pretest and post-test ANL scores in the high ANL group (t12 = 12.5, P < 0.05). Mean pretest and post-test ANLs of the three groups are given in the form of a bar graph in [Figure 2]. It can be observed from the figure that ANL values have reduced after training in high ANL group.{ Figure 2}

From the above bar graph, it can be observed that mean ANL score change following training was 9.95 dB. Extrapolating this result to individual clients would be difficult because the mean value can be affected by extreme values in the sample and, thus, might compromise accuracy. Hence a finer look at the obtained results was rendered to see the variation in ANL scores across the test and retest sessions of individual participants. The amount of ANL change varied from participant to participant. Change in ANL for each participant after training is given in [Figure 3]. It can be observed from the figure that almost all the individuals in the high ANL group have shown improvement after training.{ Figure 3}


The present study aimed at investigating the effect of systematic desensitization training on ANL. The statistical analysis revealed that there was a significant improvement in ANLs of participants with high ANL after training. These findings suggest that systematic desensitizationtraining had a positive effect on individual’s ability to tolerate the background noise. The desensitization training used in the current study could have resulted in neurophysiological changes in the auditory system similar to the effect of desensitization training in hyperacusis. Repeated and gradual exposure of sound reduces the aversive effect of sound in individuals with hyperacusis[ 29],[ 30] by reducing the response of the autonomic nervous system and limbic system. It is possible that similar to individuals with hyperacusis, even in individuals with high ANL, increasing the level of background noise may be associated with negative sensation due to the activation of the autonomic nervous system and limbic system. Because of the negative feeling such as frustration and irritation attached to the background noise, individuals with ANL may not be accepting a high level of background noise while listening to speech. Desensitization training would have reduced the aversive effect of background noise, thereby, making the individuals accept more amount of background noise.

In the current study, the training was given only for 2 weeks, and such a short-term training is proven to be useful. As per our knowledge, there is meager research information available pertaining to the effect of short-term training on ANL. However, there have been few investigations on how a short-term auditory training impacts speech-in-noise perception.[ 31],[ 32],[ 33] In none of these studies, the training period exceeded 2 weeks, and such a short-term auditory training had a positive effect on speech perception in noise. One important observation in these studies is that transfer of training effect to untrained materials was limited. Learning resulting from such auditory training is specific to the trained speech materials and the parameters of the background noise such as signal-to-noise ratio and noise types.[ 33],[ 34] In the current study, speech babble was used for the training and also for the assessment of training efficacy. Nabelek et al.[ 6] reported that ANL is independent of noise type. Hence it may be expected that training effect may be transferred to other types of background noise as well. However, it would be better if future studies can verify the transfer effect of training to other background noise as well.

Sweetow and Sabes[ 35] recommended the use of FCA prior to prescribing amplification in which, ANL is one of the essential components along with speech intelligibility measure. Taylor and Bernstein[ 36] developed a tool called “red flag matrix” following the recommendation by Sweetow and Sabes.[ 35] Red flag matrix was developed by utilizing the published normative results of QuickSIN[ 37] and ANL studies.[ 9] On the basis of the ANL and QuickSIN score, an individual can be placed in any one of the four quadrants of red flag matrix. Results of the current study have important implication in treating individuals falling in the second quadrant of red flag matrix which represents near normal QuickSIN scores and elevated ANL. Individuals falling in this quadrant are “at-risk” for experiencing annoyance difficulties, which may lead to nonuse of hearing aids.[ 36] Special features such as directional microphones and DNR algorithms may help the listeners to overcome the annoyance difficulties. Use of directional microphone improves the ANL in individuals with mid-range ANL thereby making them full-time hearing aid users.[ 23] Similarly, Lowery and Plyler[ 26] also reported that directional microphones could significantly improve the ANL. Directional microphones decrease the amplitude of the sounds that originate from the sides and back of the listener and retain the intensity of the signals that arrive from the front. Hence while using hearing aids with directional microphones, the user would have to face the direction of the sound source. This is not always easy or possible, and also directional microphones will be ineffective if background noise and target speech are not spatially dispersed. Moreover, the directionality in hearing aids may not provide many benefits to hearing impaired individuals with high ANL values.[ 23] Mueller et al.[ 8] investigated the effect of DNR on the ANL of individuals with mild to moderate sensorineural hearing loss fitted with binaural amplification. ANLs were found to be significantly lower for the DNR “on” listening condition than the unaided and DNR “off” listening conditions. This finding indicates that the use of the DNR algorithm significantly improves ANL scores. This result is further supported by Lowery and Plyler.[ 26] Hearing aids that include features such as DNR can decrease the amplitude of background noise. However, the signal of interest may occur along with the background noise, and the DNR algorithm may reduce the intensity of both the desired and the undesired signal if they share the same spectrum. Since both directional microphones and DNR algorithms have limitations, the desensitization training can be an alternate choice for individual who are falling in the second quadrant of “red flag matrix.” Other than technology, a pharmacological line of management for high ANL has also been investigated. Freyaldenhoven et al.[ 38] studied the effect of stimulant medication on ANL in fifteen normal hearing individuals with attention deficit hyperactive disorder (ADHD)/attention deficit disorder. Methylphenidate was the primary medication used to treat ADHD which is a central nervous system stimulant. The results demonstrated that the medication significantly increased the listeners’ willingness to accept background noise (i.e., decreased ANLs). This not only suggests that acceptance of noise can be manipulated with the use of pharmacological intervention but is a further indication that ANL is influenced by central auditory structures. Though pharmacological intervention has proven to improve ANLs in individuals with ADHD, there has been no research performed on the usage of these medications for individuals who do not have attention deficits. Hence, it is cautioned to conclude this remark on the case-to-case basis, because these medications cannot be used on the general population.

Systematic desensitization training technique to perceptually improve acceptance of noise could provide real world benefit to hearing aid users. The results of the present study showed significant improvements in ANLs in individuals with normal hearing after systematic desensitization training. Research performed in the past have reported that ANL is not related to hearing sensitivity.[ 6] Hence the improvement in ANLs reported by the present study provides an implication to hearing impaired population as well. However, replicating this study on hearing impaired individuals with high ANL would provide more evidence for therapeutic utility.


The present study aimed to investigate the effect of systematic desensitization training on ANL. The results revealed that there was a significant improvement in ANL scores in participants with high ANL after they were trained using the systematic desensitization procedure. However, there was no change in ANL scores in participants with low and mid ANL. From the present study, it can be inferred that the training can facilitate ANL. The research findings related to hearing aid features such as directionality of microphones and DNR algorithms provide improvements in ANL but they come with limitations. Hence, systematic desensitization procedure can be an alternate choice. Participants of the current study were individuals with normal hearing and the results obtained in this study need to be explored in the clinical population as well.

DMU Timestamp: November 09, 2018 23:10

0 comments, 0 areas
add area
add comment
change display
add comment

Quickstart: Commenting and Sharing

How to Comment
  • Click icons on the left to see existing comments.
  • Desktop/Laptop: double-click any text, highlight a section of an image, or add a comment while a video is playing to start a new conversation.
    Tablet/Phone: single click then click on the "Start One" link (look right or below).
  • Click "Reply" on a comment to join the conversation.
How to Share Documents
  1. "Upload" a new document.
  2. "Invite" others to it.

Logging in, please wait... Blue_on_grey_spinner