Written by Travis M. Moore
Last edited 21-Jul-2020
We're now leaving the cochlea behind and turning our attention to more a more central structure: the brainstem. Recording from the auditory brainstem involves a test called, you guessed it, the auditory brainstem response (ABR). You will see below that the auditory nerve starts things off, which provides a nice segue from ECochG. From there, we are able to follow the pathway of sound as it travels up through the various brainstem nuclei. So put simply, the ABR is a test of neural synchrony in the auditory nerve/brainstem. It can tell us whether or not those structures are intact, but not about true "hearing." Regardless, the ABR is still very useful, and is the most common evoked potential test in the hearing clinic. It is used to screen the lower auditory system of newborns, and can even be used to estimate behavioral hearing thresholds for fitting hearing aids in young children (or other populations unable to respond behaviorally). We'll talk more about this test and its clinical applications below.
The click-evoked ABR produces seven peaks, referred to as waves (see Figure 1). The naming convention is "Wave" + a Roman numeral. So the first wave is reported as Wave I, the fourth wave is reported as Wave IV, and so on. And all of these potentials should occur with 10 ms! It doesn't take sound long to reach the primary auditory cortex (i.e., the brain) once it has entered our ears. Out of the seven waves, we only actually know with any degree of certainty where the first two waves come from: the distal and proximal portions of the auditory nerve (AN). In the ABR's case, the distal AN is near the cochlea (further from the midline), and the proximal AN is closer to the brainstem (closer to the midline). The wave from the distal AN reflects the synchronized depolarization of neurons in response to sound that we have been talking about all along. It is the same activity seen as the compound action potential (CAP) of the electrocochleogram, but referred to as "Wave I" when recorded via ABR.
The proximal portion of the AN likely produces Wave II. This might be surprising, given that we have already established we only measure one action potential in response to a stimulus. (Recall the different conduction speeds of the axons wash out any further synchronous response visible to surface electrodes.) Wave II is not an exception to this rule; instead, we see a second response because there is a stationary dipole near the brainstem (i.e., another opportunity to see voltage at the scalp). This dipole is created by the surrounding anatomy. Once the AN leaves the cochlea, where does it go? It's helpful to have a specific image in your mind, so the anatomy we're talking about is presented in Figure 2.
After the AN leaves the cochlea, it travels through the
internal auditory meatus (IAM), with which should be familiar
from the auditory anatomy module (not yet available online). The IAM
is a bony canal that houses the facial nerve, cochlear nerve,
and the superior and inferior vestibular nerves. All those nerves
travel together through the IAM, which is made of very dense bone
(very resistant to electrical fields). However, when the
nerves emerge from the IAM, they enter the fluid-filled subarachnoid
space, and we are able to see the stationary voltage change from
transition from resistant to conductive material (i.e., IAM to fluid).
This stationary dipole is very likely wave II of the
Okay, wave III likely comes mostly from the cochlear
We said above that the ABR is the most commonly used evoked potentials test in the hearing clinic. One of the reasons it's so popular is because it's extremely easy to record. Unlike the finicky ECochG, the ABR in a normal-hearing patient should jump out at you (more or less). Interpreting a click ABR involves labelling waves I, III and V. (Remember we don't see all the waves when using more frequency-specific stimuli.) So why don't we label waves II, IV, VI and VII? We don't label waves VI or VII because they are very hard to get. Wave II is small, but you can oftentimes see it. It's just not as stable or diagnostically useful as I, III and V. Wave IV typically rides on the positive-going slope of wave V, but is also small and variable.
Picton (2010) offers a good strategy for picking ABR peaks. He recommends starting with wave I. To do so, find the most positive peak after 1 ms, before the waveform crosses into the negative area of the plot. Wave V is next, and is the positive peak after 5 ms that occurs immediately before the big negative dip in the waveform. With waves I and V identified/labeled, wave III is simply the largest positive peak between them. Just like that, you have labelled the ABR!
The ABR is so consistent that scientists have identified specific latencies and amplitudes for the waves, based on testing large groups of research participants. These normative data act like a benchmark for "normal" function. After labeling the waveform, your next step is to compare the peak latencies you identified with the normative data. It's highly unlikely you'll see latencies that are earlier than the norms; that would mean the patient's brainstem is faster than normal. Instead, if there is a problem, the peak latencies will likely be slower than the norms (indicating impaired brainstem function). You'll need to make sure to use the standard deviations provided with the normative data (normal is always a range). Do the same for the amplitudes. Table 1 below provides some example normative data that may or may not be appropriate for your clinic's use. We'll talk about why there is not a single set of normative data more below.
|Mean Latency (ms)||1.69||2.78||3.77||4.97||5.63|
|Mean Amplitude μV||0.30||0.17||0.34||0.06||0.61|
|Mean Both Genders||2.07||1.86||3.95|
The ABR just wouldn't be a real audiological test if it didn't have its own brand of dB! Unfortunately, yes, we must learn about two new types of decibels, but only one takes any real explaining. Before we dive in, let's take a moment to remember what dB are. If you need a refresher, look over the Bels and Decibels module. In short, dB are really ratios. Remember why? Check out the formula for dB SPL below:
You'll notice that we're taking the logarithm (to make the number smaller) of a ratio. Decibels are always in reference to something. In the case of dB SPL, we are comparing our measured sound pressure (p) to an agreed-upon reference sound pressure (pr). What about dB HL? (HL = hearing level.) We're just using another reference; in this case, we comparing dB SPL to the average hearing threshold of a group of normal-hearing adults. In fact, the letters after "dB" tell you what the reference is. That's how there can be so many types of dB.
dB nHL is no different - it just means we are using a new reference in the bottom of our ratio. Note we're using a modified type of "HL." The "n" stands for "normalized," so "dB normalized hearing level." We've seen above that "HL" refers to the average thresholds of a group of normal-hearing adults using pure tone audiometry. We use dB HL all the time, so why mess with a good thing?
The short answer is: "psychoacoustics." A somewhat longer answer is that psychoacoustics tells us that the longer a sound is presented, the more noticeable it is, up to around 200 - 500 ms, depending on the frequency and presentation level. Figure 4 demonstrates that behavioral thresholds decrease (get better) the longer audiometric frequencies are presented until around 200 ms. Take a look at the calibration standards for audiometers (ANSI S3.6) and you'll see pulsed tones have to be on for at least 225 ms. That's not an accident. The audiometer presents tones for at least long enough for the auditory system to register the sound fully. We don't want to say a patient has poor thresholds just because the sound was presented too quickly for a patient to have a fair chance of hearing it.
But what about click stimuli for recording the ABR? The duration is 100 μs! That's nowhere near 225 ms! So now we see why the ABR needs its own type of dB based on a reference that involves extremely short stimuli. So dB nHL provides a reference of average behavioral thresholds based on very short stimuli used to record the ABR (instead of pure tones).
From the explanation of dB nHL above, it should be no surprise that the lowest stimulus level that produces an ABR threshold is higher than the level of a pure tone that produces a behavioral threshold. In other words, thresholds are higher (worse) when measured via ABR compared to behavioral thresholds. In theory, there should always be a difference between dB nHL thresholds and dB HL thresholds. Enter dB eHL: "dB equalized hearing level." The dB eHL reference is simply based on a correction factor that is applied to account for the difference between dB nHL (ABR) and dB HL (behavioral) thresholds. As you might have guessed, dB eHL must be measured for every clinic individually.
Can you think of why you need clinic-specific correction factors? It's because ABR thresholds depend on the parameters used to record them, for example: filtering, presentation level, electrode montage, patient population, and more! That means you'll get ABR thresholds specific to your clinic, based on whatever parameters your clinic uses.
The procedure is easy enough. First, measure behavioral thresholds for 10 - 20 normal-hearing individuals (dB HL). Next, measure their ABR thresholds (dB nHL). Last, subtract dB HL values from the dB nHL values and you have your correction factors. From then on, when you record ABR thresholds in dB nHL, just subtract the correction factor to get dB eHL, which should be pretty close to a patient's behavioral threshold. For example, "20 dB eHL" tells other clinicians that you measured an electrophysiological threshold, but adjusted it to approximate a behavioral threshold.
For instance, let's say you test 15 normal-hearing people and get an average threshold at 1000 Hz of 5 dB HL. Next, you collect ABR thresholds from the same group using a 1000-Hz stimulus (discussed later) and the average threshold is 30 dB nHL. Your CF would be 30 dB nHL - 5 dB HL = 25. Now you're ready to see patients. Your first patient has an ABR threshold of 40 dB nHL. But that doesn't mean much to anyone. To get an approximation of the patient's behavioral threshold at 1000 Hz, just apply your CF: 40 dB nHL - 25 = 15 dB eHL. When you're fitting hearing aids, you would fit 1000 Hz as though you got behavioral data and the threshold was 15 dB HL.
|Non-inverting (active)||Cz or Fz|
|Inverting (reference)||Earlobe (A) or mastoid (M) of the test ear|
|Ground (common)||Contralateral earlobe (A) or mastoid (M)|
|Non-inverting (active)||Cz (center of head) or Fz (high forehead)|
|Inverting (reference)||Each earlobe (A) or each mastoid (M)|
|Ground (common)||Fpz (low forehead)|
|High-pass filter||5 - 30 Hz|
|Low-pass filter||2000 - 3000 Hz|
|Intensity||80 dB nHL|
|Rate||17/s or 27/s|
*Using alternating polarity cancels out most of the cochlear microphonic, and combines rarefaction and condensation clicks. Alternating polarity is a good rule of thumb, but not your only (or even the best) option. Consider collecting rarefaction and condensation separately, so you can see both, and use the average of the two runs to label the waves.
The ABR has the distinct advantage of not requiring a behavioral response
from the patient. That means we can test newborns within just a few days
after birth. Getting a newborn hearing screening (NBHS) with portable ABR
equipment is routine medical care in the United States. NBHS is
especially important in the neonatal intensive care unit (NICU), as this
population is more likely to have established risk factors for hearing
loss. While otoacoustic emissions (OAEs) can also be used to screen
newborns, the ABR is the gold standard
Another common use of the ABR is to estimate hearing thresholds in populations that cannot respond, such as infants, small children, individuals with developmental delays, etc. Because the click ABR cannot be used to estimate frequency-specific information, this type of ABR uses frequency-specific stimuli. All the click ABR can tell you is that some nerve fibers, somewhere, are responding, but you have no idea which frequency. This topic is discussed in detail in the next lesson: ABR threshold estimation.
This lesson has focused mostly on the neurodiagnostic ABR: using a click to elicit a broad response from as many auditory nerve fibers as possible. Using the parameters in the tables above (or similar), a clinician can see all seven waves. More frequency-specific stimuli typically do not elicit all ABR waves, in part because fewer neurons are responding, and in part frequency-specific stimuli are usually presented at or near hearing threshold levels. So it's up to the neurodiagnostic ABR to provide detailed information about the integrity of the auditory brainstem.
Going through this lesson has already introduced you to the procedures: