Skip to content

LESSON 130

Threshold estimation with ABR

We explain how ear-specific and frequency-specific threshold estimation is done in infants with tone burst ABR. With examples, we cover converting dB nHL thresholds to estimated hearing level (eHL), the steps of the threshold search, and preparing for testing in natural sleep.

Watch on YouTube

Follow along

Hello, and welcome to the AIHEAR Audiology Academy.
In this lesson we will learn how frequency-specific threshold estimation is done in infants with ABR.
We will cover the tone burst threshold search, converting nHL values to estimated hearing level, and preparing for natural sleep testing.
In the first months of life, it is usually not possible to obtain a reliable behavioral threshold from a baby.
A baby who does not pass newborn hearing screening is referred to the audiology clinic for diagnostic assessment.
The aim is not only to answer whether there is a loss, but to estimate thresholds for each ear and frequency.
International recommendations aim for diagnosis usually before three months, and for intervention to begin before six months.
Frequency-specific thresholds form the basis of the estimated audiogram that will be used for hearing aid fitting if needed.
A click is a very brief, broadband stimulus; it produces a strong, clear wave V.
But the click response mainly reflects hearing in the region of about two thousand to four thousand hertz.
So with a click alone, a low-frequency loss or a notch at a particular frequency can be missed.
A tone burst is a short piece of tone lasting a few cycles, with its energy concentrated around a particular frequency.
This allows separate threshold estimates at five hundred, one thousand, two thousand and four thousand hertz.
The resulting picture is a frequency-specific estimate that resembles an audiogram.
We covered the structure of these stimuli in detail in our previous lesson.
The level of ABR stimuli is often given in dB nHL, the normalized hearing level.
The zero point is the average level at which a group of normal-hearing adults can behaviorally hear that brief stimulus.
Brief stimuli are heard at higher levels than a long pure tone, because the ear integrates sound over time.
So the nHL scale is not identical to the dB HL scale of the audiogram.
Also, the threshold found with ABR is usually somewhat higher than the same person's behavioral threshold.
To bridge these two differences, the correction values we will see shortly are used.
Because we do not know how long the baby will sleep, the test order is planned in advance.
In many protocols, two thousand hertz is tested first in each ear, then five hundred hertz.
If time allows, four thousand and one thousand hertz are added.
This way, even if the baby wakes early, we have information about both low and high frequencies in both ears.
The order may vary with the clinic's protocol; what matters is collecting the most critical information first.
In infants, air conduction ABR is usually done with insert earphones.
Each ear is tested separately, and results are reported separately for the right and left ears.
Because the baby's ear canal is small, the same setting produces a higher sound pressure in the canal than in an adult.
Insert earphones also prevent the canal collapse that the pressure of supra-aural earphones can cause.
Interaural attenuation is high with insert earphones, so the need for masking in air conduction is reduced.
Vernix or fluid residue in the canal can falsely raise thresholds, so the canal should be checked before testing.
The threshold search usually starts at a moderate level where a clear response is expected.
When a repeatable wave V is seen, the level is usually lowered in ten decibel steps.
As threshold is approached, five decibel steps are used if needed.
The threshold is the lowest level at which a wave V is replicated in both recordings.
To strengthen this decision, the absence of a response is shown at the level just below, with adequate averaging.
A no-response decision is reliable only if the residual noise in the recording is low enough.
As the level decreases, waves I and III disappear; wave V is the last to disappear.
This is why wave V is followed in a threshold search.
As the level drops, wave V latency increases and its amplitude decreases.
At low frequencies, for example five hundred hertz, the response is broader and comes later.
At these frequencies, the slow negative wave that follows wave V also helps to recognize the response.
So in tone burst recordings, the analysis window is usually widened to twenty to twenty five milliseconds.
The response decision is first made visually, by checking whether two recordings overlap at the same latency.
Residual noise is the background activity left in the recording after averaging; the lower it is, the more reliable the decision.
Many devices also use statistical measures such as Fsp to show numerically whether the response stands out from noise.
Using visual assessment and statistics together reduces the chance of a wrong decision.
If there is no response even at the device's maximum level, the result is reported as no response at the highest level tested.
As we saw in earlier lessons, a good threshold recording starts with low and balanced electrode impedance.
When the baby moves or sucks, muscle artifact increases; the recording is paused until it settles.
Artifact rejection is kept on, so sweeps with very large voltages are not included in the average.
Near threshold the response is smaller, so the number of sweeps is increased and residual noise is checked to be low enough.
If recording quality is poor, the threshold may be overestimated; this must be stated in the report.
A clinical tip: if you get no response at any frequency with tone bursts, do not end the session right away.
Record with a high-level click, separately in condensation and rarefaction polarity.
An early response that inverts with polarity may be a cochlear microphonic and raises the suspicion of auditory neuropathy.
We will see in later lessons how this response is told apart from stimulus artifact, and auditory neuropathy in detail.
The estimated hearing level, or eHL, is the ABR threshold converted to the audiogram scale.
The calculation is simple: subtract the correction value for that frequency from the threshold found in nHL.
Correction values vary with the device, stimulus and protocol; published tables also differ from each other.
In many tables, the correction is larger at low frequencies and smaller at high frequencies.
Some protocols apply a smaller correction as the degree of hearing loss increases.
Now let us do the calculation with an entirely made-up infant recording.
In this example the 500 Hz threshold is 50 dB nHL; with a correction of about 15 decibels, the estimate is 35 dB eHL.
For 1000 Hz, about 10 decibels is subtracted from the 45 dB nHL threshold, giving 35 dB eHL.
At two thousand and four thousand hertz the correction is smaller; in this example about five decibels was used.
The result is an estimated loss profile in the mild to moderate range, increasing toward the high frequencies.
The correction values here are only examples; in the clinic, always use the table of your own protocol.
Estimated thresholds are reasonable estimates, not exact values that replace the behavioral threshold.
To tell whether a loss is conductive or sensorineural, bone conduction ABR is needed; we will see it in the next lesson.
In infants under six months, the middle ear is assessed with one thousand hertz probe tone tympanometry.
OAE results are also part of the test battery and give additional information about cochlear function.
When the baby becomes able to do behavioral testing, the estimated thresholds are confirmed with visual reinforcement audiometry.
Because the ABR is easily affected by muscle activity, the baby needs to be in deep, calm sleep.
In the first few months of life, the test can often be done in natural sleep, without medication.
The appointment is planned around the baby's usual sleep and feeding times.
The family is asked to keep the baby awake as much as possible before arriving and to feed the baby at the clinic.
Electrodes and insert earphones are placed before the baby falls asleep, while the baby is held by a parent.
A dim, quiet and warm room helps the baby stay asleep.
As the baby grows, usually after six months, it becomes harder to get a long, calm recording in natural sleep.
Sedation may then be needed; the decision and its administration are the physician's responsibility.
The audiologist's role is to prepare in advance a test plan that makes the best use of the sedation time.
In some children, the test can also be planned during general anesthesia given for another reason.
An ABR recording from an awake, moving baby is usually not reliable for threshold estimation.
The report states the stimulus used and the frequencies tested for each ear.
Both the nHL thresholds and the eHL estimates calculated from them are written separately.
The correction table used is also stated, so that another center can interpret the result correctly.
The baby's sleep state, whether sedation was used, and the recording quality are also added to the report.
Frequencies that could not be completed are clearly listed, and a follow-up appointment is planned.
The click mainly reflects the high frequencies; tone bursts are used for frequency-specific information.
The threshold is the lowest level at which a repeatable wave V is seen.
The nHL threshold is converted to estimated hearing level, or eHL, with a frequency-specific correction.
In young infants, testing is done in natural sleep whenever possible.
Test yourself: if the 500 Hz ABR threshold is 50 dB nHL and the correction is 15 decibels, what is the estimated threshold?
The correct answer is A: 35 dB eHL.
The correction value is subtracted from the nHL threshold, not added.
In the next lesson we will examine the latency-intensity function and bone conduction ABR, which lets us tell the type of loss.
Subscribe to the channel so you never miss a new lesson. See you next time!

Keywords

Click a keyword to search all lessons for that term.

Prerequisites

LESSON 128 Clicks, tone bursts and chirps

Comments

Questions and comments appear after approval.

No comments yet. Be the first to comment.