Click-evoked scalp potentials tracing the ascending auditory pathway — from cochlear nerve firing to newborn hearing screening and neurodiagnostic threshold estimation
The auditory brainstem response is the electrophysiological fingerprint of synchronized neural firing along the ascending auditory pathway. A single, brief acoustic transient recruits thousands of auditory nerve fibers to fire in near-perfect unison, and that synchronized volley can be tracked, relay by relay, all the way to the midbrain using nothing more than scalp electrodes.
A conventional ABR click is a 100 microsecond rectangular electrical pulse delivered to an insert earphone or supra-aural transducer, producing a broadband acoustic transient with energy concentrated between roughly 1–4 kHz. Because the click has an extremely fast rise time, it depolarizes a broad swath of the cochlear basilar membrane almost simultaneously — most effectively the basal, high-frequency region — driving thousands of type I spiral ganglion neurons to fire action potentials within a narrow, synchronized time window of about 1 millisecond.
This synchrony is the entire trick of the ABR. A tone or continuous sound smears neural firing out in time (each nerve fiber phase-locks to a different part of the ongoing waveform), and the individual, temporally dispersed action potentials cancel out when averaged. A click, by contrast, produces a compound action potential — the summed extracellular field of many synchronously firing axons — large enough, when averaged over many repetitions, to be recorded from the scalp far removed from the auditory nerve itself.
Jewett and Williston first described the characteristic five-to-seven vertex-positive peaks in 1971, and decades of intracranial recording, lesion studies, and animal models have refined the anatomical correlates of each wave:
• Wave I — the compound action potential of the distal auditory (cochlear) nerve, essentially identical to electrocochleography's N1 • Wave II — proximal auditory nerve as it enters the brainstem, near the cochlear nucleus • Wave III — cochlear nucleus and its crossing projections into the contralateral superior olivary complex • Wave IV — superior olivary complex, often fused with wave V into a "IV/V complex" • Wave V — lateral lemniscus input to the inferior colliculus; the most robust, highest-amplitude, and clinically dominant peak, present down to near-threshold intensities even when earlier waves have vanished
Because wave V survives at the lowest stimulus levels and shows the most reliable, reproducible morphology, it is the peak audiologists track for both screening pass/refer decisions and threshold estimation.
Wave V is generated primarily by the lateral lemniscus fiber tract as it terminates in the inferior colliculus of the midbrain — several synapses and multiple decussations away from the cochlea — which is precisely why a normal wave V implies intact function through the entire brainstem auditory pathway, not just the ear.
Extracting a 1 microvolt neural signal from a background of ongoing EEG, muscle artifact, and electrical noise 50–100 times larger requires disciplined electrode placement and an averaging strategy that exploits the one property noise does not share with the response: time-locked reproducibility.
The standard clinical montage places the non-inverting (active) electrode at the vertex (Cz) or high forehead, the inverting (reference) electrode on the ipsilateral earlobe or mastoid (A1/A2), and a ground electrode on the contralateral mastoid or low forehead. This vertex-to-mastoid configuration maximizes the recorded amplitude of the vertically-oriented dipole fields generated by the brainstem pathway.
Stimulus polarity is a key recording parameter: rarefaction clicks pull the eardrum outward first, condensation clicks push it inward first, and alternating polarity averages both directions together. Alternating polarity is often preferred clinically because it cancels the cochlear microphonic (a stimulus artifact that flips sign with polarity) and any stimulus-related electromagnetic artifact, though single-polarity recordings can better preserve wave I morphology for interpeak measurements.
Bandpass filtering is typically set around 100–3,000 Hz to emphasize the fast neural transients while rejecting slow EEG rhythms and 60 Hz line noise.
Ongoing EEG, muscle (EMG) artifact, and electrical interference are effectively random relative to the exact moment of stimulus onset, while the neural ABR response is time-locked and reproduces its waveform shape on every sweep. Averaging N time-locked sweeps together improves the signal-to-noise ratio by a factor of the square root of N, since random noise partially cancels through destructive summation while the identically-timed signal reinforces itself additively.
Averaging 2,000 sweeps therefore improves SNR by roughly √2000 ≈ 45-fold relative to a single sweep — enough to lift a 1 μV response cleanly above 50–100 μV background EEG. Because averaging time is clinically precious (especially in a sleeping or sedated infant), most systems interleave two independent running averages ("A" and "B" buffers) so a replicated, superimposable waveform can confirm that a peak is a genuine physiologic response rather than residual noise or artifact.
Any sweep in which the recorded voltage exceeds a fixed rejection threshold (commonly ±25–40 μV) — typically caused by a muscle twitch, eye movement, or sudden electrode noise — is discarded before it can contaminate the running average. In infants this is usually achieved passively by testing during natural or sedated sleep rather than active rejection algorithms alone.
At a typical click intensity of 70–80 dB nHL in a normally-hearing adult, well-established normative latencies are approximately: wave I ≈ 1.5 ms, wave III ≈ 3.5 ms, and wave V ≈ 5.5–6.0 ms, yielding a wave I–V interpeak interval of about 4.0 ms. These values shift with age (neonates show prolonged, still-myelinating latencies that mature toward adult values by 12–18 months), stimulus rate, and click polarity, so every clinical lab maintains its own normative reference curves.
Universal Newborn Hearing Screening (UNHS) uses automated ABR devices to deliver an objective pass/refer decision within minutes, without requiring a trained audiologist to visually interpret a waveform at the bedside — making population-wide screening of every newborn logistically feasible before hospital discharge.
Devices such as the Natus ALGO 3/5 and GSI Corti deliver click stimuli at a fixed screening intensity — typically 35 dB nHL — and average the recorded response using disposable surface electrodes. Rather than displaying a waveform for a clinician to eyeball, the device runs an automated statistical algorithm (commonly a normalized cross-correlation or a template-matching detection statistic) that compares the recorded, averaged response against a stored normative wave V template derived from thousands of known-normal infant ears.
If the recorded waveform correlates with the template above a preset confidence threshold within a bounded number of sweeps, the device outputs "PASS." If the correlation fails to reach threshold within the maximum allotted sweeps (commonly capped to bound test time), it outputs "REFER." This turns a subjective interpretive task into an objective, reproducible, technician-administerable screening result.
Many UNHS programs use a two-stage protocol: otoacoustic emissions (OAE) testing first, as a fast, inexpensive cochlear-function screen, followed by AABR for infants who refer on OAE or who are in the NICU (where AABR is preferred because it also probes retrocochlear and neural pathway integrity, which OAE cannot assess — critical for detecting auditory neuropathy spectrum disorder).
False positive refers are common and expected: residual vernix caseosa or amniotic debris in the ear canal, middle ear fluid from the birth process, or a fussy/active infant generating myogenic artifact can all cause a true-normal ear to refer. This is precisely why the JCIH 1-3-6 benchmark builds in a rescreen and full diagnostic step rather than treating a single refer as a diagnosis — the initial ~2–4% refer rate falls substantially after outpatient rescreening.
The Joint Committee on Infant Hearing (JCIH) 1-3-6 rule is the backbone of Universal Newborn Hearing Screening policy worldwide: every infant should be screened by 1 month of age, receive a full diagnostic audiologic evaluation by 3 months if they refer, and — if hearing loss is confirmed — be enrolled in early intervention services by 6 months of age.
As click intensity is reduced, wave V does not simply shrink and vanish uniformly — it follows a predictable, well-characterized latency-intensity function: latency prolongs and amplitude diminishes in a systematic curve that clinicians use to estimate hearing sensitivity in patients who cannot provide reliable behavioral responses.
At high stimulus intensities, the traveling wave on the basilar membrane recruits a broad region and the most sensitive, fastest-responding high-frequency basal fibers dominate the synchronized response, producing a short-latency wave V. As intensity decreases, fewer fibers reach firing threshold, recruitment shifts toward more apical (lower-frequency, slower-responding) regions of the cochlea, and neural conduction time through each synaptic relay lengthens slightly — the net effect is a smoothly prolonging wave V latency and a shrinking wave V amplitude as intensity drops toward threshold.
This relationship is remarkably consistent across normally-hearing ears, which is why a "latency-intensity curve" — plotting wave V latency against stimulus dB — can be compared against normative curves to classify the type and even estimate the degree of hearing loss: pure conductive loss shifts the entire curve upward in a roughly parallel fashion, while cochlear (sensorineural) loss steepens the curve near threshold.
Clinical threshold search uses a bracketing strategy: start at a comfortably audible intensity (e.g., 70 dB nHL) where wave V is unambiguous, then descend in 10–20 dB steps, replicating each waveform to confirm reproducibility, until wave V is no longer reliably identifiable. The lowest intensity at which a reproducible, replicated wave V can still be identified is taken as the electrophysiologic threshold — typically within 10–20 dB of the true behavioral pure-tone threshold in adults, and used similarly (with wider margins) in infants who cannot be behaviorally tested.
Because a broadband click stimulates mostly the 1–4 kHz basal cochlear region, click-evoked ABR threshold is not a full audiogram — it is weighted toward that frequency range. For frequency-specific information, brief tone-burst stimuli (typically 500 Hz, 1 kHz, 2 kHz, and 4 kHz, gated with a rise/fall envelope to preserve some frequency specificity while retaining enough spectral splatter to synchronize firing) are used to build an estimated frequency-specific audiogram in infants — essential for fitting hearing aids appropriately across frequencies.
Beyond hearing screening, the ABR is a sensitive neurodiagnostic probe of the auditory nerve and brainstem pathway itself. A retrocochlear lesion — classically a vestibular schwannoma (acoustic neuroma) compressing the auditory nerve — delays neural conduction on the affected side, producing measurable, comparable asymmetries between the two ears.
A vestibular schwannoma arising from the vestibular portion of cranial nerve VIII physically compresses adjacent cochlear nerve fibers as it grows within the internal auditory canal and cerebellopontine angle, slowing neural conduction velocity on the affected side without necessarily abolishing cochlear (peripheral) function. Because wave V timing depends on that conduction, comparing wave V latency between the two ears at matched stimulus intensities isolates this neural delay from any peripheral hearing loss.
An interaural latency difference (ILD) — the difference in wave V latency between right and left ears — exceeding roughly 0.2 to 0.4 ms (thresholds vary slightly by laboratory and normative dataset) is considered abnormal and raises suspicion for a retrocochlear lesion on the side with the longer, delayed latency, prompting referral for contrast-enhanced MRI.
Historically, before high-resolution MRI became widely available, ABR with ILD and interpeak-interval analysis was the primary screening tool for acoustic neuroma, achieving roughly 90–95% sensitivity for tumors a centimeter or larger — though sensitivity drops substantially for the small, intracanalicular tumors under about 1 cm that modern gadolinium-enhanced MRI now detects almost universally, which is why MRI has become the diagnostic gold standard while ABR remains a useful, low-cost adjunct and monitoring tool.
In addition to the interaural comparison, the absolute wave I–V interpeak interval (normally roughly 4.0 ms, with an upper normal limit near 4.4 ms) measures conduction time from the auditory nerve to the midbrain within a single ear. A prolonged I–V interpeak interval, particularly when wave I is normal in latency but wave III and V are delayed, localizes the abnormality to the retrocochlear pathway (nerve or brainstem) rather than to the cochlea itself — since a purely cochlear (sensorineural) hearing loss delays all waves roughly proportionally to the degree of hearing loss without disproportionately stretching the interpeak intervals.
This wave-by-wave localization is what makes the ABR uniquely valuable among audiologic tests: it is not just a hearing test but a functional map of neural conduction timing along a specific anatomical pathway, capable of distinguishing cochlear, auditory nerve, and brainstem sources of a hearing complaint.
The entire architecture of Universal Newborn Hearing Screening exists because timing matters enormously for the developing brain. Auditory input during the first months of life shapes the central auditory system during a critical period, and decades of outcome research now demonstrate a clear, quantifiable benefit to identifying and treating hearing loss early rather than late.
Christine Yoshinaga-Itano's landmark Colorado studies, beginning in the late 1990s, provided the pivotal evidence base for UNHS policy: children with congenital hearing loss identified and enrolled in early intervention before 6 months of age showed language development scores approaching the normal range for hearing children by school age, while children identified later — even by 18–24 months, an age once considered reasonably early — showed persistent, substantial language delays that often did not fully close.
The underlying neurodevelopmental logic is that the central auditory pathway, and the language-processing cortical networks that depend on patterned auditory input, undergo experience-dependent maturation during a sensitive early window. Depriving that system of structured auditory input during the first months to years of life — even if hearing is eventually restored with amplification or cochlear implantation — produces effects that are progressively harder to fully remediate the later intervention begins.
A "Refer" result on newborn AABR screening is not a diagnosis — it triggers the JCIH 1-3-6 pathway: outpatient diagnostic audiologic evaluation (including full-frequency ABR threshold estimation, OAE, and tympanometry) by 3 months to confirm and characterize the type and degree of hearing loss, followed by enrollment in early intervention services — hearing aid fitting, cochlear implant candidacy evaluation for severe-to-profound loss, and family-centered early intervention/education services — by 6 months of age.
Universal Newborn Hearing Screening programs, now covering over 98% of newborns in the United States and widely adopted internationally, have compressed the average age of identification of congenital hearing loss from roughly 2–3 years in the pre-screening era down to a few months — directly enabling the early-intervention window that outcome studies show is so consequential for lifelong language, literacy, and academic achievement.
Because moderate-or-greater congenital hearing loss affects roughly 1 to 3 of every 1,000 live births, universal screening of every newborn — rather than screening only infants with known risk factors — was necessary policy: risk-factor-based screening alone was shown to miss roughly half of all infants with congenital hearing loss.