ABR waves I, III, and V: latencies and what shifts them

The auditory brainstem response is a series of tiny electrical waves generated as sound moves up the auditory pathway, recorded from electrodes on the scalp. Wave I comes from the distal part of the auditory nerve itself, close to the cochlea. Wave III is generated further up, around the cochlear nucleus and the superior olivary complex. Wave V, generally the largest and most robust wave in a normal recording, is generated near the lateral lemniscus and inferior colliculus, well into the brainstem. Because the waves are generated at successively higher points in the pathway, the time between them, not just their individual latencies, tells you something about where a problem sits.

Using a click stimulus at 80 to 90 dB nHL, the level at which clinical waveforms are usually easiest to read, published adult normative data puts wave I at roughly 1.6 to 1.65 ms, wave III at roughly 3.7 to 3.8 ms, and wave V at roughly 5.5 to 5.65 ms. The interpeak intervals, the differences between waves rather than their absolute latencies, run about 2.1 to 2.2 ms from I to III, about 1.7 to 1.8 ms from III to V, and about 4.0 ms from I to V overall.

Different published normative sets give slightly different numbers, and the simplified round numbers many students memorize, 1.5, 3.5, and 5.5 ms for waves I, III, and V with 2.0 ms interpeak intervals throughout, are a teaching approximation rather than a single precise standard. A lab should be judged against its own age-matched normative data, roughly plus or minus 2.5 to 3 standard deviations, rather than against a single memorized number from a textbook.

A conductive hearing loss simply attenuates the stimulus before it ever reaches the cochlea, so every wave in the response is delayed by roughly the same amount, on the order of 0.3 to 0.4 ms of added latency for every 10 dB of air-bone gap. Because every wave shifts together, the interpeak intervals, the differences between waves, stay normal even though every absolute latency is longer than expected. Recognizing that pattern, uniform delay with normal interpeak intervals, is what keeps a conductive component from being mistaken for a neural problem.

A cochlear, sensorineural loss produces a different pattern: latencies are prolonged at low and moderate stimulus levels but tend to normalize at high levels, because once the signal is loud enough to drive the surviving basal fibers, neural timing recovers. The interpeak intervals in a cochlear loss generally stay normal once a high enough level is used.

A retrocochlear lesion, most classically an eighth nerve tumor, behaves differently again: the delay does not shrink at high stimulus levels, because the problem is neural conduction itself rather than how much signal reaches the cochlea. Wave I can be entirely normal while waves III and V are both late, and the interpeak intervals from I to III, III to V, or I to V are themselves prolonged, often beyond about 4.4 ms for I to V. An interaural difference in wave V latency greater than about 0.4 ms between the two ears is another commonly used flag for a possible retrocochlear problem. The single most useful discriminator across all three patterns is whether the interpeak intervals, not just the absolute latencies, are normal or prolonged.