Measuring an individual’s ability to hear speech in quiet or in the presence of background noise has long been a key component of a comprehensive audiological assessment. Moving beyond the pure-tone audiogram allows us to gain insights that extend far beyond threshold information and provide us with greater understanding of the patient’s functional hearing ability in the real world.
As our profession advances towards precision diagnostics and the use of increasingly sophisticated amplification technologies, it is essential to continually evaluate the tools available to us and their relevance within modern clinical practice. In the area of speech testing specifically, several important developments have emerged. These advancements largely fall into three domains:
- Objective techniques to measure aided access to speech sounds.
- Automated methods of measuring the speech recognition threshold (SRT) and word recognition scores (WRS).
- Non-language-based tools to predict hearing-in-noise ability.
We will look at each of these areas and examine where the research is taking us.
Objective measures of measuring aided access to speech sounds
As screening for hearing loss has become more efficient, clinicians are increasingly able to diagnose and fit amplification for children at younger ages. For infants between approximately three and seven months of age, validation of hearing aid fittings often relies solely on subjective observations and caregiver-completed questionnaires. As a result, measurement of access to speech sounds is typically delayed until the child is developmentally capable of completing an age-appropriate speech test. Similar challenges arise in older children and adults with complex needs, for whom behavioural speech measures may also be unreliable or infeasible.
Figure 1: An example of an aided cortical test being performed.
There is an emerging shift in clinical practice aimed at reducing disparities in access to validated speech measures, with many clinics now incorporating aided cortical testing into their validation protocols. This approach uses brief, synthetic speech stimuli such as the ManU-IRU stimuli set which are delivered through a loudspeaker while cortical auditory responses are recorded using an evoked potentials system (Figure 1) [1]. The procedure employs an electrode montage identical to that used for an auditory brainstem response (ABR) test, and the patient remains awake throughout.
This method of recording access to speech sounds has recently been validated through a large-scale clinical feasibility study conducted by the University of Manchester in terms of the ‘Ladies in the Van’ project. Findings from this work demonstrated that aided cortical testing using the ManU-IRU stimuli is a sensitive measure of which speech sounds can be detected by a child wearing hearing aids, providing clinicians with valuable additional information to support the management of hearing loss [2].
Automated methods of measuring access to speech
Although the benefits of traditional behavioural speech testing are well established, several barriers continue to limit its routine use in some clinical settings. Time constraints remain one of the most significant obstacles. To address this challenge, researchers have investigated whether portions of the speech assessment can be completed independently by the patient, allowing clinicians to focus on other tasks while the test is administered, and later devote time to interpretation and counselling.
One solution to this challenge is the Automated Method for Testing Auditory Sensitivity (AMTAS). AMTAS is capable of performing air- and bone conduction audiometry, masking, speech reception thresholds and suprathreshold word-recognition scores. Unlike traditional clinician led speech testing, AMTAS employs an adaptive multiple choice format in which patients select the word they hear from four options, and the system incorporates several quality control indicators to ensure reliable and consistent results (Figure 2).
Figure 2: Example of test screen from the AMTAS speech test.
AMTAS is well researched and its methodology has been described in multiple peer-reviewed publications [3]. As service delivery becomes increasingly complex and waiting lists grow, it is likely that automated methods of testing will become more widely adopted, allowing clinicians to focus on other tasks and therefore improve clinical efficiency.
Non-language-based assessments of hearing-in-noise ability
The final challenge in speech audiometry addressed in this article concerns access to validated speech materials. Standardised speech tests are not available in all languages, and this limitation is even more pronounced for speech in noise assessments. In the absence of validated materials, clinicians often rely on live voice methods, which introduce substantial variability. Furthermore, when the clinician does not share the patient’s language, speech testing may be omitted entirely.
To address these barriers, researchers have developed language-independent approaches based on spectro temporal modulation (STM) detection to predict hearing-in-noise ability [4]. The most clinically established method derived from this work is the Audible Contrast Threshold (ACT™) test.
ACT estimates speech in noise ability using a task that is both rapid (under three minutes) and completely language independent [5]. Instead of recognising words, the patient listens to a noise stimulus that is modulated in both the spectral and temporal domains. When the listener detects the modulation, the modulation depth is progressively reduced following an adaptive procedure until the minimum detectable level is reached (Figure 3).
Figure 3: Example of the ACT test screen after a completed test run.
This threshold, known as the ACT value and expressed in dB nCL, provides a robust predictor of speech in noise performance and can be used to guide the configuration of advanced hearing aid features. As such, the ACT test offers a more inclusive and clinically efficient alternative to traditional speech in noise assessments and is becoming widely used in audiology clinics worldwide.
Summary
Speech testing remains a critical component of the audiological test battery, yet the complexity of traditional methods has limited access for many patients. The advancements outlined above offer opportunities to expand the availability of validated speech measures, ultimately supporting more accurate diagnosis and improved rehabilitation outcomes.
References
1. Stone MA, Visram A, Harte JM, Munro KJ. A Set of Time-and-Frequency-Localized Short-Duration Speech-Like Stimuli for Assessing Hearing-Aid Performance via Cortical Auditory-Evoked Potentials. Trends Hear 2019;23:2331216519885568.
2. Visram AS, Stone MA, Purdy SC, at al. Aided Cortical Auditory Evoked Potentials in Infants With Frequency-Specific Synthetic Speech Stimuli: Sensitivity, Repeatability, and Feasibility. Ear Hear 2023;44(5):1157–72.
3. Margolis RH, Glasberg BR, Creeke S, Moore BC. AMTAS: automated method for testing auditory sensitivity: validation studies. Int Journal Audiol 2010;49(3):185–94.
4. Zaar J, Simonsen LB, Dau T, Laugesen S. Toward a clinically viable spectro-temporal modulation test for predicting supra-threshold speech reception in hearing-impaired listeners. Hear Res 2023;427:108650.
5. Zaar J, Simonsen LB, Sanchez-Lopez R, Laugesen S. The Audible Contrast Threshold (ACT) test: A clinical spectro-temporal modulation detection test. Hear Res 2024;453: 109103.
Declaration of competing interests: LM is employed by Interacoustics A/S.
Click here to read ‘Audible Contrast Threshold – a new test to guide setting help-in-noise features in hearing aids’ by Leigh Martin.
