Skip to content

LESSON 125

Introduction to Auditory Evoked Potentials

Auditory evoked potentials are small electrical responses to sound, recorded from the auditory nerve, the brainstem and the cortex. In this lesson we look at how the response is extracted from the EEG with signal averaging, filtering and artifact rejection, at latency and amplitude, at early, middle and late responses, and at their clinical uses.

Watch on YouTube

Follow along

Hello, and welcome to the AIHEAR Audiology Academy.
In this lesson, we will learn what auditory evoked potentials are and how they are recorded from the scalp.
We will talk about signal averaging and filtering, which pull the response out of brain activity, classification by latency, and uses.
Auditory evoked potentials, or AEPs, are electrical responses produced along the auditory pathway in response to sound.
These responses arise at different levels, starting from the cochlea and the auditory nerve and reaching up to the brainstem and the cortex.
Recording is done with surface electrodes attached to the scalp; no needles or invasive procedures are needed.
The responses are time-locked to the stimulus; that is, they always appear with the same delay after each stimulus.
This property lets us separate the response from random background brain activity.
Most AEPs do not require the patient to press a button or respond, so they are counted among the objective tests.
This is why they are so valuable in infants, in patients who cannot cooperate, and in assessing the nervous system.
In 1929, Hans Berger published his recording of the electrical activity of the human brain from the scalp.
This recording is called electroencephalography, or EEG.
In 1939, Pauline Davis described the response to sound that appears in the EEG of a waking person.
The real advance that made small responses visible was the spread of computer averaging methods in the nineteen sixties.
With the description of the brainstem response in humans in the early nineteen seventies, clinical electrophysiology grew quickly.
In an AEP system, the chain starts with the stimulus: sound is delivered through an earphone or an insert earphone.
Electrodes placed on the scalp pick up brain activity.
A differential amplifier greatly boosts the difference between two electrodes, while noise common to both is largely canceled.
The signal is digitized, filtered and averaged by the computer.
The result is a waveform plotted over time on the screen.
The basic problem is that the evoked response is very small.
Background EEG activity and muscle tension recorded from the scalp are usually in the range of tens of microvolts.
The auditory brainstem response, or ABR, is often below one microvolt.
On the screen, the gray line is a single recording and the blue line the response itself on the same scale; the response is lost in noise.
So seeing the response in a single recording is like looking for a needle in a haystack.
The solution rests on two basic tools: signal averaging and filtering.
The principle of signal averaging is simple: the same stimulus is presented hundreds or thousands of times.
The short recording after each stimulus, called a sweep, is aligned to the moment the stimulus started, and the sweeps are averaged.
Because the response occurs at the same time and in the same direction in every sweep, it is preserved in the average.
Random noise, however, is sometimes positive and sometimes negative, so it gradually cancels out in the average.
Residual noise decreases with the square root of the number of sweeps: four times as many sweeps halve the noise.
For example, with two thousand sweeps, noise becomes roughly forty five times smaller than in a single recording.
This is why ABR recordings often average one thousand to two thousand sweeps.
The second tool is filtering: each response carries energy in a certain frequency range.
A band-pass filter removes slow drifts and very fast noise outside this range from the recording.
For example, about one hundred to three thousand hertz is used for the ABR, and about one to thirty hertz for slow cortical responses.
Filter settings can change the shape and latency of the response, so norms are compared using the same filter settings.
The third safeguard is artifact rejection: sweeps that exceed a set voltage are left out of the average.
In this way, large disturbances such as patient movement or muscle contraction do not spoil the recording.
A comfortable, relaxed patient is the simplest way to improve recording quality.
We describe an evoked response with two basic measures: latency and amplitude.
Latency is the time from the stimulus onset to the peak of a wave, and it is measured in milliseconds.
Amplitude is the size of the wave, measured in microvolts; it is often taken as the difference between a peak and the trough that follows.
Latency reflects how fast the response travels along the pathway; amplitude reflects how many nerve fibers take part and how synchronously.
In the clinic, latency is more reliable; amplitude varies more from person to person.
We will read the waves and latencies of the ABR example on the screen in detail in later lessons.
AEPs are most often classified by when they appear after the stimulus, that is, by their latency.
Early responses appear within about the first ten milliseconds; electrocochleography, or ECochG, and the ABR belong to this group.
These responses come from the cochlea, the auditory nerve and the brainstem.
The middle latency response, or MLR, appears between about ten and fifty milliseconds and reflects the thalamus and the cortex.
Late, or cortical, responses appear after about fifty milliseconds and arise from the auditory cortex.
Responses such as the P300 and the mismatch negativity, or MMN, are related to the discrimination of sounds.
As a general rule, the longer the latency, the higher up the auditory pathway the source lies.
Electrocochleography records the responses of the cochlea and the auditory nerve in the first few milliseconds.
The ABR shows waves from the auditory nerve and the brainstem within about ten milliseconds.
The auditory steady-state response, or ASSR, is a rhythmic response to a continuously modulated sound.
The MLR contains waves from the thalamus and the cortex between ten and fifty milliseconds.
Cortical responses consist of the P1, N1 and P2 waves and show that sound has reached the cortex.
The P300 and the MMN reflect higher-level processes related to the discrimination of sounds.
Early responses are exogenous; their shape is set largely by physical properties of the stimulus, such as intensity, frequency and rate.
These responses are little affected by sleep and sedation, so they can be measured while infants sleep.
The ABR and electrocochleography are typical examples of this group.
Later responses are increasingly endogenous; the person's attention, alertness and task change the response.
This is why responses such as the P300 require an awake patient who can cooperate.
The MLR and the cortical responses can also be clearly affected by sleep and sedation.
The most common clinical use is estimating hearing thresholds in infants and children who cannot do behavioral testing.
Automated ABR devices are used in newborn hearing screening, especially for infants in intensive care.
Latency measurements help assess problems in the auditory nerve and the brainstem.
In some surgeries, recording continues throughout the operation to protect auditory nerve function.
They help in assessing cochlear implant candidates and, in some cases, in programming the device with electrically evoked responses.
Together with otoacoustic emissions, they allow auditory neuropathy spectrum disorder to be recognized.
Cortical responses are also used to show whether sound delivered by a device reaches the cortex.
A tip from practice: do not try to fix a noisy recording only by increasing the number of sweeps.
Tense neck muscles and movement greatly increase the noise; first make sure the patient is comfortable and relaxed.
In adults, a supported head and closed eyes work best; in infants, natural sleep is the most effective way.
Low and balanced electrode impedance is also essential; we will see this in detail in the next lesson.
Auditory evoked potentials are small electrical responses that occur time-locked to sound.
The response is pulled out of the EEG by averaging thousands of sweeps, filtering and artifact rejection.
By latency, early, middle and late responses are distinguished; the longer the latency, the higher up the pathway the source.
The main uses are threshold estimation, newborn screening, neurological assessment and surgical monitoring.
Test yourself: If the number of sweeps is raised from one thousand to four thousand, how does the residual noise change?
The correct answer is B: it roughly halves.
Noise decreases with the square root of the number of sweeps; the square root of four is two, so the noise is halved.
In the next lesson, we will learn where electrodes are placed on the scalp, and the 10-20 system.
Subscribe to the channel so you never miss a new lesson. See you soon!

Keywords

Click a keyword to search all lessons for that term.

Prerequisites

Comments

Questions and comments appear after approval.

No comments yet. Be the first to comment.