General Questions About HEARLab
What is the HEARLab® device?
The HEARLab® device is used in the recording and analysis of human electrophysiological data, in the form of the Cortical Auditory Evoked Potential (CAEP). The CAEPs are recorded from an active electrode placed near the vertex with respect to a reference electrode placed near either mastoid, with a ground electrode placed on the forehead. Two tests are available: Aided Cortical Assessment (ACA) and Cortical Threshold Evaluation (CTE). Aided Cortical Assessment (ACA) is intended to assist the audiologist to determine whether speech is audible to people who have been fitted with hearing aids. Cortical Threshold Evaluation (CTE) uses tone bursts to estimate the hearing threshold levels of hearing-impaired adults without requiring the active participation of the person being tested.
What do ACA and CTE stand for?
Aided Cortical Assessment (ACA) and Cortical Threshold Evaluation (CTE).
What is ACA?
Aided Cortical Assessment (ACA) is intended to assist the audiologist to determine whether speech is audible to people who have been fitted with hearing aids. Three speech stimuli with low (/m/), medium (/g/), and high (/t/) frequency emphasis are presented to the patient in the free-field at input levels ranging from 55 to 75 dB SPL. These signals have a presentation time long enough to activate the compression circuits of a hearing aid, making it an effective tool for determining whether the amplification produced by the hearing aid provides a signal that is actually detectable at the cortical level of the patient. The unaided response of the patient can also be measured, allowing the clinician to compare the unaided and the aided cortical responses.
The relationship between presence of a cortical response, as automatically detected by the HEARLab® device, and the audibility of the stimulus eliciting the response, has so far been established only for infants (aged 8 to 30 months) listening to speech sounds presented in the sound field, and for hearing-impaired adults (aged 39 to 82, mean 64) listening to speech sounds presented in the sound field. The acoustic stimuli are presented at 55 dB SPL, 65 dB SPL, 75 dB SPL. The stimulus set comprises 3 specific speech tokens /m/, /g/ and /t/ which were extracted from running speech. Their durations are 30, 20, and 30 ms respectively, and they have dominant power in the bands 200 to 500 Hz, 800 to 1600 Hz, and 2000 to 8000 Hz respectively.
What acoustic stimuli are available for presentation in ACA mode?
Three “natural” speech stimuli, /m/, /t/ and /g/ are available in the ACA mode. They are speech tokens extracted from recordings of running speech. They were chosen for their dominant energy spectrum in different frequency bands; /m/ for the low frequencies, /g/ for the mid frequencies and /t/ for the high frequencies. The aim is to gives some indication to the tester as to whether speech information of low-, mid- and high-frequency bands are independently detectable by the subject at the cortical level.
What presentation levels are available for presentation in ACA mode?
Three presentation levels are available for each of the speech stimuli. These levels are:
- 55 dB SPL, representing speech at soft levels
- 65 dB SPL, representing normal (average conversational) speech levels; and
- 75 dB SPL, representing speech at louder levels
More information can be found in section 7.1 of the HEARLab® manual and the spectral energy of the speech stimuli is shown in Appendix B of the manual.
Why does HEARLab® only use /m/, /g/ and /t/?
These speech sounds have been extensively used in cortical response projects at NAL (Dillon, Golding, Purdy, & Katsch, 2006; Golding, Dillon, Seymour, Purdy, & Katsch, 2006; Golding, Pearce, Seymour, Cooper & Ching, ibid) for the assessment of infants with normal hearing and those fitted with hearing aids. These sounds cover the biggest part of the speech spectrum. In addition, it has been planned to add the very-high frequency sound /s/ in the future as this sound might prove to be useful. The duration of these stimuli have been determined based on experimental evidence gathered at NAL (Golding et al. 2006).
Can ACA be used for those who wear cochlear implants?
Cochlear implants send out an electrical artefact and the current software used by HEARLab® can potentially misinterpret the presence of the artefacts which can lead to a wrong decision about corticals. Despite this, a couple of independent clinical studies seem to indicate that cochlear implant (CI) artefacts do not necessarily preclude the measurement of cortical auditory evoked potentials (CAEPs) in children using HEARLab® (Kosaner 2011, 2012; Raghunandhan 2012). These studies suggest that HEARLab® recordings can be used to track cortical maturation after implantation and to evaluate CI fitting in children. However, as a systematic research study has not been carried out yet to determine the exact influence of CI artefacts on CAEPs recordings – although the study is in progress at NAL – we do recommend to refrain from using the HEARLab® automatic detection system when testing CI clients, and to inspect the waveforms visually ONLY. Please discontinue testing when there is a suspicion of CI artefact presence. It might be possible to reduce the size of the CI artefact by repositioning electrodes. Currently, HEARLab® is certified in Australia, United States, Europe, and the Middle East only for testing unaided or with hearing aids, not with cochlear implants.
Can ACA testing be performed monaurally?
Masking is not available for testing in ACA mode. However, the non-test ear may be blocked with an earplug (e.g. foam EAR plug). For aided evaluation, if monaural results are desired the hearing aid in the non-test ear can be left in-situ but switched off.
Can we compare a response at 55 dB SPL on HEARLab® to aided thresholds from an audiogram?
No. A positive response at 55 dB SPL could be the subject’s threshold or suprathreshold so we cannot make any comparisons with the aided audiogram as HEARLab® ACA currently does not allow testing at softer levels than 55 dB SPL.
Why can’t we test speech sounds at lower intensities?
Unfortunately if the stimulus intensity is reduced below 50 dB SPL then we risk the possibility that the stimulus intensity will be too close to the noise floor, especially for the lower frequency sounds /m/ and /g/. This will reduce the cortical response or cause it to disappear entirely. Alternatively you can test with inserts using the CTE module however speech sounds are not available for this module and you cannot perform aided testing this way.
What is CTE?
Cortical Threshold Evaluation (CTE) uses tone bursts to determine at what levels these frequency specific tones are audible to the test person. This assessment can be used for estimating the hearing threshold levels of hearing-impaired adults without requiring the active participation of the person being tested. The relationship between the estimates of hearing threshold obtained on the basis of the CAEPs detected by the HEARLab® device, and behaviourally measured hearing thresholds has so far been established only for adults aged 43 to 89 (mean age 71 years). The tone bursts are 40 ms duration at frequencies of 500, 1000, 1500, 2000 and 4000 Hz. Intensities presented with insert phones cover the range -10 to 110 dB HL in 5 dB steps. Intensities presented with bone conductors cover the range -10 to 70 dB HL in 5 dB steps (60 dB HL for 500 Hz).
Is it possible for speech stimuli to be audible to a person yet the CAEP is absent?
Yes. Several studies have shown that a small proportion of people do not exhibit a strong CAEP even when the stimulus is audible. Clinical studies done with HEARLab® ACA on children and with CTE on adults have also shown results in line with these studies. The HEARLab® system is intended as an adjunct to the assessment of hearing, not as a replacement for other methods of hearing assessment.
What acoustic stimuli are available for presentation in CTE mode?
There is a choice of six tone-burst frequencies in the CTE mode. The envelope of these tone-bursts comprises a rise time and a fall time of 10 ms with a plateau of 30 ms. The duration of these tone-bursts is equivalent to 40 ms. The tone-burst frequencies available are 500, 1000, 1500, 2000, 3000 and 4000 Hz. The calibration procedure employs a continuous tone that has the same peak level as the tone-burst stimuli used during the cortical testing. To compensate for the short duration of the tone bursts, the CTE mode automatically applies a 6 dB correction to the levels determined in the continuous tone calibration.
Can we use CTE to measure behavioural thresholds in infants or children?
The relationship between the estimates of hearing threshold obtained on the basis of the CAEPs detected by the HEARLab device, and behaviourally measured hearing thresholds has so far been established only for adults aged 43 to 89 (mean age 71 years). This does not mean we cannot use it for infants or children but a great deal of caution MUST be exercised when interpreting results. If there is a response = good. No response = caution!
What are the correction factors for CTE testing?
CAEP thresholds estimate behavioural thresholds comparably well as ABR thresholds.
NAL did a study with 34 hearing-impaired adults, and these were the results:
91% of cortical thresholds lie within 15 dB of behavioural thresholds.
Correction factors are all about 5 dB for the different tone bursts (500, 1000, 2000, 4000 Hz). So subtract 5 dB from the cortical threshold to obtain the behavioural threshold. These are average values.
There is some variability in the threshold estimates (similar to ABR). Lower frequencies (500 and 1000 Hz) have a standard deviation of about 8 dB. Higher frequencies (2000 and 4000 Hz) about 14 dB. This is caused by the observation that about 7% of the subjects had elevated CAEP thresholds (more than 20 dB worse than behavioural threshold) at higher frequencies. We don’t know yet why this is caused, but ABR has an analogeous counterpart in the form of ANSD (although with corticals it’s not ANSD that explains it).
I’ve mainly tested with the ACA module. What’s the best clinical approach to start with CTE testing in adults?
Testing takes a while: similar to tone burst ABR one has to record one frequency at a time. Recording times are comparable. It’s about 10 to 15 minutes per frequency. You need to record about 100 epochs per intensity (2 minutes), and on average you need about 5 intensities to obtain enough information to decide on the threshold. So for 4 frequencies per ear, in total it will take about 1.5 to 2 hours with a rather quiet subject.
Uncontrollable movements will be an issue and extend recording time.
Operation of the CTE module is pretty much identical to the ACA module (aided corticals). You can only select one frequency to test at one time however (500, 1000, 2000, 4000 Hz) and the hearing level (dB) at which you want to test. Select masking noise if appropriate. All the rest is identicial to the ACA.
We recommend the following to reduce testing time:
- Start with only 2 frequencies per ear, like 500 or 2000 Hz, or 1000 and 4000 Hz. Check the other frequencies when you’re finished with these.
- Start with 60 dB HL. If present, go down to 30; if not, go up to 85. If doing 30, and there is a cortical present, go down to 15; otherwise go up to 45. If doing 85, and there is a cortical present, go down to 75; otherwise 100. These are not fixed rules, but they give a good idea how to search quickly for the likely location of the threshold. An example of a ‘decision tree’ can be found at decision tree.png
- Using this ‘decision tree’ method, you can actually define a range you’d be satisfied with, e.g. do I only use 10 dB steps? If so, you’ll be much quicker (but a bit coarser), e.g. you might find no response for 60, but a response at 80, hence the hearing threshold lies somewhere in between 60 and 80 dB HL. If you’re happy with this, proceed to other frequencies which need the rest of your time.
- As you’ll do multiple intensities for the same frequency, you’ll quickly start to detect a pattern in both p-value and cortical waveforms. Especially with hearing-impaired subjects you’ll see that if you’re far above threshold (like 20 dB) the cortical will be huge and p-values will quickly reach 0.001 or smaller (we would suggest to then stop the measurement prematurely). Conversely, if you had a barely significant p-value (e.g. p in the 0.02 to 0.05 range), you very likely will be close to threshold anyway. We’re working on a way for HEARLab® to make those decisions automatically.
- Re the above,we also warn that when you get a barely significant cortical at e.g. 60 dB HL, it might be a false detection and you might divert to the ‘wrong’ side of the tree. If you keep on getting non-detected corticals, it might be a good idea to just try a point high above 60 dB HL (without damaging the subject’s hearing of course).
- If you’re uncomfortable with the decision tree, standard audiometry patterns work too, like starting at 85 dB HL, then if you get a significant cortical dropping 30 dB and so on. It may take a bit longer, but might show the relationship between intensity and CAEP amplitude better.
Where can I find a step-by-step list for cortical testing?
Where can I find information on CAEP to give to parents or clients?
The file below serves as information for the parent/guardian when testing with CAEPs:
How do I report HEARLab® results to other professionals?
Some report examples can be found here:
The compressed file includes reports for:
- after aided testing
- a possible CI candidate
- complex cases
- 2 other cases
Is it currently possible to import sound stimuli from external source/wav files. The equipment seems to have the physical capability with line in/out sockets?
The current software does not use sound stimuli from external sources and therefore it is currently not possible to import sound stimuli from external source.
Note on HEARLab® product concept:
The HEARLab® product concept is to develop a versatile device (equipment hardware) that can be configured and controlled by a software module that presents acoustic stimuli, acquires responses, analyses, presents and saves results in the manner required for the test. Different software applications will implement different audiological tests. The hardware external input facility is provisioned for future applications that may make use of external sources. The software released so far does not.
Why is the low pass filter up to 4 kHz when usually cortical filter only goes up to 100 Hz so that muscle movement has less impact on the measurement?
Referring to the product concept above, the acquisition electronics set the circuit low pass at 4 kHz to allow for development of ABR software applications. For cortical applications, band limiting appropriate for cortical measurements is capably handled by software.
Will sleep affect the cortical response?
Yes, the state of alertness of the person being tested is critical in cortical testing. Sleep is known to affect the amplitudes and latencies of cortical responses (Hall, 2007). CAEP responses will become longer and the amplitudes larger, but they also seem to be less reliably measurable, both in adults as in infants.
NAL has to do much more research about the effect of sleep on cortical testing on HEARLab®, but based on some earlier published papers available in literature, it seems possible to test both awake and sleeping subjects. However, caution should be taken that the subject doesn’t get into the REM sleep / active sleep / deep sleep, which basically removes all auditory cortical activity.
The subject shouldn’t change state during the recording, as then different response morphologies will overlap and cancel each other out in the average of the waveform.
Where can I find more information on the scientific background of CAEPs and HEARLab®?
See Chapter 8 (p 75) of the HEARLab® manual.
Can I use HEARLab® on infants (or adults) fitted with bone conductor devices?
Yes. This is no different to normal hearing aid testing. If the sound is audible enough, then there should be a response.
Are there any safety issues associated with HEARLab®?
Malfunction of the device, including the software, is unlikely to cause any injury to the patient. The device has been evaluated and tested to only cause minor discomfort to the patient, like temporary reddening of the skin at the places where the electrodes are attached. This discomfort should disappear within 24 hours. There have been no reports of any serious harm caused from the use of HEARLab® and its electrode application. See section 1.3.5 of the HEARLab® manual for more information.