The auditory brainstem response (ABR) is a series of five to seven neurogenic potentials, or waves, that occur within the first 10ms following acoustic stimulation (Sohmer and Feinmesser, 1967; Jewett, Romano, and Williston, 1970). The potentials are the scalp-recorded synchronous electrical activity from groups of neurons in response to a rapid-onset (<1 ms) stimulus. An example of these potentials, with their most common labeling scheme using Roman numerals, is shown in Figure 1. Waves I and II are generated in the auditory nerve, wave III is predominantly from the cochlear nucleus, and waves IV and V are predominantly from the superior olivary complex and lateral lemniscus (Moller, 1993).

Figure 1..  

The normal ABR elicited by a high-level click stimulus, with waves I–VI labeled.

The ABR is valuable in audiology and neurology because of its reliability and highly predictable changes in many pathological conditions affecting the auditory system. The ABR may be used in a number of ways in adults. Perhaps the most common use is in the diagnostic assessment of a hearing loss, either to determine the site of the lesion or to determine the function of the neural system. A second use is to estimate auditory sensitivity in patients who are unable or unwilling to provide accurate behavioral thresholds. A third use is for monitoring the auditory nerve and brainstem pathways during surgery for auditory nerve tumors or vestibular nerve section. For diagnostic and neural monitoring purposes, the latencies and amplitudes of the most reliable waves, waves I, III, and V, are analyzed. For estimation of auditory sensitivity, the lowest detection level for wave V is used to approximate auditory threshold. Responses are replicated to ensure reliability of the waveforms.

Because the potentials are small (<1 µV) and embedded in high levels of background electrical noise, several techniques are used to enhance the visibility of the potentials (ASHA, 1988). Surface electrodes are attached to the scalp of the patient, with the placement in line with the orientation of the dipoles of the neural generators. Pairs of electrodes are used such that one electrode picks up positive activity and one electrode picks up negative activity from the neural generators. Differential amplification/common mode rejection enhances the electrical activity that is different between two electrodes (the potentials) and reduces the activity that is the same between electrodes (the random background electrical noise). The physiological response is filtered to eliminate the extraneous electroencephalographic activity and external line noise. Signal averaging increases the magnitude of the time-locked response and minimizes the random background activity. Artifact rejection eliminates unusually high levels of electrical activity that are impossible to eliminate through averaging.

The diagnostic interpretation of the ABR is based on the latencies and amplitudes of the component waves. The latencies, or times of occurrence, of the waves are the more reliable measures because latencies for a given person remain stable across recording sessions unless intervening pathology has occurred. The latency is evaluated in absolute and relative terms. Absolute latency is measured from the arrival of the stimulus at the ear to the positive peak of waves I, III, and V. Relative latency is measured between relevant peaks within the same ear or between ears. The absolute latency reflects the state of the auditory system to the generation site of each wave but may be affected by conductive or cochlear pathology, making it difficult to isolate any delay from neural pathology. Interpeak intervals allow an estimation of neural conduction time and are less dependent on peripheral pathology than absolute latencies are. Interaural values limit variability by using the nonsuspect ear as a control in a patient with unilateral pathology, but interaural values also may be affected by asymmetrical conductive or cochlear pathology.

The amplitudes of the waves are more variable than the latencies and therefore are less useful for determining normality of the waveform. Amplitudes, measured from the positive peak to the averaged baseline or from positive peak to subsequent negative trough, are affected by the quality of the electrode contact, physiological noise levels of the patient, and amplitudes of adjacent waves. Consequently, absolute amplitude values are little used except in the case of absent waves. Relative amplitudes, particularly the amplitude ratio of waves I and V, may be useful measures to control for measurement variables. A decrement in wave V amplitude relative to wave I amplitude may suggest auditory nerve or low brainstem pathology.

For the estimation of auditory threshold, wave V may be traced to its detection threshold, which is typically defined as the lowest stimulus level at which wave V can be seen. The ABR does not measure hearing per se, but correlates with auditory sensitivity in most cases. Clicks, which are broadband stimuli, provide estimates of average sensitivity in the range of 2000–4000 Hz because of cochlear physiology that biases the response to the high-frequency region of the cochlea (Fowler and Durrant, 1993). For better definition of thresholds across the frequency range, frequency-specific stimuli, such as tone pips or clicks with ipsilateral masking, may be used (Stapells, Picton, and Durieux-Smith, 1993). Because of limitations in the signals and the cochlea, thresholds for frequencies below 1000 Hz are difficult to obtain, and other methods, such as middle or late auditory-evoked potentials or steady-state evoked potentials, may yield better results or additional information (see Hall, 1992; Jacobson, 1993).

Clicks at high presentation levels are the most commonly used stimuli for diagnostic ABRs because their abrupt rise times elicit the necessary degree of synchrony to obtain the full complement of waves. Although the normal ABR to a click is dominated by neurons associated with 2000–4000 Hz, a peripheral hearing loss at those frequencies may effectively filter the stimulus, causing different frequency regions to dominate the response in different hearing losses, which will affect the ensuing ABR latencies (see Fowler and Durrant, 1993). ABR latencies from suspect ears are compared with the norms at equivalent sound pressure levels at the cochlea. The typical effects of auditory pathology on the latencies and thresholds of the ABR are discussed below.

Normal Hearing

Normative latencies may vary somewhat among clinics, but are typically derived from the mean and ±2 standard deviations of the latencies of waves from a jury of listeners with normal hearing. Examples of normative values for waves I, III, and V are shown as the bounded areas in the lower panel in Figure 2. Typical normal interpeak intervals are <2.51 ms for interpeak interval I–III, <2.31 ms for III–V, <4.54 ms for I–V, and <0.4 ms for interaural V latency (Bauch and Olsen, 1990). Figure 2 includes the ABRs for high-level signals to threshold for a person with normal hearing, along with the latencies of those waves plotted against the normal range. All latencies are within the normal limits, and the threshold is 40 dB peak sound pressure level (pSPL), which equals 10 dB nHL (normalized hearing level).

Figure 2..  

The top panel illustrates the normal ABR elicited by clicks from 110 dB pSPL to 30 dB pSPL, with wave V labeled down to the visual detection threshold of 40 dB pSPL. The lower panel shows the latencies of waves I, III, and V plotted against the normative values for those waves, indicated by the bounded areas.

Conductive Hearing Loss

The absolute latencies are prolonged and the amplitudes are reduced relative to the degree of the conductive component of a hearing loss, but the interwave intervals are normal because the neural system is intact. Consequently, a person with a 30-dB conductive hearing loss will produce an ABR to a 90-dB nHL click that is approximately equivalent to the ABR to a 60-dB nHL click for a person with normal hearing. An example of a waveform and the resulting latencies from a 30-dB flat, conductive hearing loss are shown in Figure 3. All absolute latencies are prolonged, but the I–V latency difference is within normal limits. The threshold is 70 dB pSPL.

Figure 3..  

The top panel illustrates the ABR for a person with a mild conductive hearing loss, with wave V labeled to the visual detection threshold of 70 dB pSPL. The lower panel shows the latencies for waves I, III, and V plotted above the normative values, indicated by the bounded areas.

Cochlear Hearing Loss

The degree and configuration of the hearing loss affect the latencies of the waves in a cochlear hearing loss, although typically the I–V interval is normal because the neural system is intact. Most mild to moderate cochlear losses do not affect the latencies of the ABR for high-level click stimuli, although the amplitudes of the waves may be reduced. Severe high-frequency hearing losses may reduce the amplitudes and prolong the absolute latencies of the waves with little effect on the I–V latency interval, although wave I may be absent. An example of the waveform and resulting latencies from a moderate, flat hearing loss are shown in Figure 4. Absolute and interwave latencies are within normal limits, and the threshold is 100 dB pSPL.

Figure 4..  

The top panel illustrates the ABR for a person with a moderate, flat cochlear hearing loss, with wave V labeled to its visual detection threshold of 100 dB pSPL. The lower panel shows the latencies for waves I, III, and V plotted within the normative values, indicated by the bounded areas.

Retrocochlear Hearing Loss

Retrocochlear pathology refers to any neural pathology of the auditory system that is beyond the cochlea and may include such disorders as acoustic neuromas, multiple sclerosis, brainstem strokes, and head trauma. Retrocochlear pathology may produce a variety of effects on the latencies and morphology of the ABR depending on the type, location, and size of the pathology. Effects may include absence of waves, prolonged absolute latencies or interwave intervals, or prolonged interaural wave V latencies. In Figure 5, the waveform and resulting latencies are shown from a patient with an acoustic neuroma and a mild, high-frequency hearing loss. Wave I is within normal limits at the two highest levels, but the absolute latency for wave V, and consequently the I–V interval, is prolonged beyond the normal limits. The wave V threshold may be variable in people with retrocochlear pathology, and may not provide a useful estimation of behavioral threshold.

Figure 5..  

The top panel illustrates the ABR for a person with a vestibular schwannoma, with the reliable waves labeled. The lower panel shows the latencies for waves I and V plotted against the normative values for those waves. Wave I is within normal limits, whereas wave V is significantly prolonged, yielding a prolonged interwave I–V interval. The threshold for wave V does not correspond to the behavioral auditory threshold.