Hughson-Westlake threshold search, air & bone conduction, masking, and audiogram interpretation — a clinical hearing test walkthrough
Before a single tone is presented, a valid pure tone audiogram depends on controlling every variable that could bias the threshold: ambient noise, transducer calibration, and the physical state of the ear canal. A rushed setup is the single most common source of erroneous audiograms in clinical practice.
A pure tone audiogram is meaningless without context. The audiologist first takes a case history: onset (sudden vs. gradual), laterality, tinnitus, vertigo, noise exposure history, ototoxic medication use, and family history of hearing loss. Sudden unilateral sensorineural loss is an audiologic emergency requiring same-day ENT referral (steroid treatment window closes within 2 weeks of onset).
Otoscopy follows immediately: the canal is inspected for cerumen impaction (a leading cause of falsely elevated air conduction thresholds — occlusion can add 15–40 dB of conductive-appearing loss), tympanic membrane perforation, middle ear effusion (visible as amber fluid or bubbles), and foreign bodies. Any occluding cerumen must be removed before valid air conduction testing, since the plug attenuates sound before it ever reaches the cochlea and would masquerade as a real conductive hearing loss.
Pure tone thresholds are measured relative to audiometric zero — the modal threshold of young, otologically normal adults, standardized as "0 dB HL" separately for every frequency and every transducer type under ANSI S3.6-2010 (internationally, ISO 389 series). A raw 0 dB HL tone at 125 Hz actually outputs roughly 45 dB SPL, while at 4000 Hz it outputs closer to 10.5 dB SPL — the audiometer automatically compensates for the ear's uneven physical sensitivity across frequency so that "0" always means "average normal threshold."
Testing must occur in a sound-treated booth meeting ANSI S3.1 maximum permissible ambient noise levels — otherwise low-level ambient noise itself masks the test tones near 0 dB HL and artificially elevates measured thresholds. Audiometers are calibrated at minimum annually with a sound level meter, artificial ear/mastoid coupler, and biological check against known listeners; calibration drift of just 5 dB can shift a patient from "normal" to "mild loss" classification.
Insert earphones (ER-3A/ER-5A) are increasingly preferred over supra-aural TDH-39 headphones: they reduce ear canal collapse artifacts in older/pediatric patients, provide better attenuation of ambient noise, and critically raise interaural attenuation from ~40 dB to ~55–70 dB, reducing how often masking is required.
Air conduction (AC) testing measures the softest tone a patient can reliably detect through the entire hearing pathway — outer ear, middle ear, cochlea, and auditory nerve. The Hughson-Westlake modified method of limits is the ASHA/BSA-standardized staircase procedure used worldwide to converge on this threshold efficiently and reproducibly.
1. Familiarization: present a clearly audible tone (e.g., 40 dB HL at 1000 Hz) so the patient understands the response task (button press or raised hand).
2. Descending phase: reduce level in 10 dB steps after each response, until the patient no longer responds.
3. Ascending phase ("up 5, down 10"): from the first no-response level, raise in 5 dB steps until a response occurs, then immediately drop 10 dB and ascend again in 5 dB steps.
4. Threshold criterion: the lowest hearing level at which the patient responds to at least 2 out of 3 (or 3 of 5) ascending trials is recorded as threshold. This bracketing from below, rather than descending onto the threshold, minimizes false positives from anticipation and reduces the effect of variable attention.
5. Test order: 1000 Hz is tested first (and re-tested for reliability, within 5 dB of the first measurement), then 2000, 4000, 8000 Hz, then back down to 500, 250, and 125 Hz. This order avoids fatigue effects from high-frequency testing biasing subsequent low-frequency thresholds.
A full diagnostic audiogram (both ears, AC + BC, all octaves) typically takes 20–30 minutes for a cooperative adult; considerably longer with masking or pediatric behavioral adaptations (visual reinforcement audiometry, play audiometry).
Results are plotted on a standardized chart: frequency (125 Hz–8 kHz) on a logarithmic x-axis, hearing level (−10 to 120 dB HL) on an inverted y-axis — better hearing (lower dB) plots near the top, worse hearing plots toward the bottom, mirroring intuitive "higher = louder needed" reading.
Unmasked air conduction: red circle (○) for right ear, blue X (✕) for left ear, connected by solid lines within each ear. Masked air conduction (when masking noise was required in the non-test ear): red triangle (△) right, blue square (▢) left. No response at the equipment's maximum output is indicated with a downward-pointing arrow from the last tested symbol, signifying "worse than this level, unmeasurable."
These ANSI S3.21 symbol conventions are near-universal across audiology clinics, which is what makes an audiogram instantly interpretable by any trained clinician without narrative description.
A bone oscillator (typically the Radioear B-71) is placed on the mastoid process (or forehead) and vibrates the skull directly, bypassing the outer ear canal and middle ear ossicular chain to stimulate the cochlea mechanically. Comparing bone conduction (BC) thresholds — a direct readout of cochlear/neural sensitivity — against air conduction thresholds isolates where in the auditory pathway a hearing loss originates.
Because bone conduction stimulates the cochlea directly, BC thresholds represent the best possible "sensorineural reserve" of the ear — the outer/middle ear are not part of the pathway. Air conduction thresholds represent the whole system. The gap between them is diagnostic:
• AC elevated, BC normal (gap > 10 dB): conductive hearing loss. The cochlea works fine, but something is blocking or damping sound in the outer/middle ear — cerumen, perforated eardrum, otosclerosis (fixed stapes footplate), middle ear effusion, ossicular discontinuity.
• AC and BC both elevated by a similar amount (gap ≤ 10 dB): sensorineural hearing loss. The cochlea or auditory nerve itself is the site of the deficit — noise damage to outer hair cells, presbycusis, ototoxicity, Ménière disease, vestibular schwannoma.
• AC elevated more than BC, and BC itself also elevated: mixed hearing loss — both a conductive component and an underlying sensorineural component are present simultaneously (e.g., otosclerosis in an already presbycusic ear).
Bone conduction testing has a much lower maximum output than air conduction (the oscillator can only vibrate the skull so hard before distortion), so profound losses may leave BC thresholds untestable — the audiogram shows arrows at the equipment ceiling rather than a true floor.
When a supra-aural earphone or foam plug covers the ear canal during bone conduction testing of that ear, low-frequency (250–1000 Hz) bone-conducted sound pressure builds up in the now-sealed canal instead of escaping, artificially improving (lowering) the measured BC threshold by up to 20 dB — the "occlusion effect." This mainly affects patients with normal or near-normal middle ears; it is negligible in conductive pathology where the middle ear itself is already compromised.
Clinically, audiologists either leave the non-test ear canal open during BC testing, apply an occlusion-effect correction factor, or use insert earphones with venting to minimize the artifact. Ignoring the occlusion effect can make a purely sensorineural loss appear to have a spurious low-frequency conductive component.
Sound presented to one ear does not stay confined to that ear — it can cross the skull and stimulate the opposite cochlea. When the level needed in the test ear is loud enough to cross over and be detected by the non-test ear, the audiogram would (falsely) record the better ear's threshold as belonging to the worse ear. Masking noise is presented to the non-test ear to prevent this "cross-hearing" contamination.
Masking for air conduction is indicated whenever the difference between the test ear's air conduction threshold and the non-test ear's bone conduction threshold reaches or exceeds the minimum interaural attenuation (conservatively 40 dB for supra-aural phones): AC(test ear) − BC(non-test ear) ≥ 40 dB.
Masking for bone conduction is required far more often, because interaural attenuation for bone-conducted sound is approximately 0 dB — the skull conducts vibration to both cochleae almost equally regardless of which mastoid is stimulated. The clinical rule of thumb: mask bone conduction whenever there is ANY air-bone gap of 10 dB or more in the test ear, since without masking you cannot be certain whether you are measuring the test ear's cochlea or the better cochlea on the other side via cross-conduction.
Failing to mask when indicated is one of the most common serious errors in audiometry — it can make a unilateral profound hearing loss look like a bilateral moderate loss, sending the patient down the wrong management pathway entirely (e.g., missing a vestibular schwannoma on the "silent" side).
Once masking is triggered, narrowband noise (centered on the test frequency, since white/broadband noise would waste energy outside the critical band and cause unnecessary loudness) is introduced to the non-test ear at an initial effective masking level, then the test tone is re-presented to find threshold.
The "plateau method" increases masking level in 10 dB steps while re-measuring the test-ear threshold at each step. Initially, added masking may cause the measured threshold to rise (because the previous "threshold" was really the crossed-over signal in the non-test ear being masked away). Once genuine masking is sufficient, further increases in masking level produce a stable plateau — the same test-ear threshold across a 15–20 dB range of masking levels — confirming the true, unmasked-by-crossover threshold of the test ear has been isolated. If the plateau never appears and threshold keeps rising with masking level, this indicates "masking dilemma" — the masking noise itself is crossing back over to the test ear, common in severe bilateral conductive losses, and requires insert earphones or narrower masking bands to resolve.
A textbook masking dilemma: a patient with bilateral severe conductive hearing loss (e.g., bilateral atresia) can have an air-bone gap so large in both ears that no amount of masking in one ear stays confined there — it always leaks back to the test ear. Audiologists resolve this using insert earphones (higher interaural attenuation) or accept a "best estimate" threshold with documented uncertainty.
Once all air and bone conduction thresholds are plotted, the audiogram is read as a single integrated picture: how much hearing is lost (degree), which frequencies are affected (configuration), and where in the auditory pathway the problem lies (type — conductive, sensorineural, or mixed). This classification directly drives the treatment pathway.
The Pure Tone Average (PTA) — the arithmetic mean of thresholds at 500, 1000, and 2000 Hz, the frequencies most important for speech understanding — is the single number most often quoted to summarize overall hearing sensitivity per ear:
• −10 to 25 dB HL: Normal hearing • 26–40 dB HL: Mild hearing loss (soft speech and distant conversation difficult) • 41–55 dB HL: Moderate hearing loss (normal conversational speech difficult without amplification) • 56–70 dB HL: Moderately severe hearing loss • 71–90 dB HL: Severe hearing loss (only loud speech understood at close range) • > 90 dB HL: Profound hearing loss (little or no auditory access to speech even amplified; candidacy for cochlear implant evaluation)
PTA can be a poor single summary when the audiogram configuration is steeply sloping (e.g., normal low frequencies, profound high frequencies from noise exposure) — in these cases the configuration itself (flat, sloping, rising, notched, cookie-bite) is reported alongside PTA, since it dramatically affects speech intelligibility even at matched PTA values.
Conductive hearing loss (air-bone gap > 10 dB, bone conduction ≤ 25 dB HL): the cochlea is healthy but sound is blocked or damped en route. Common causes: cerumen impaction, otitis media with effusion, otosclerosis, ossicular chain discontinuity, tympanic membrane perforation. Highly treatable — often medically or surgically reversible (myringotomy tubes, stapedectomy, tympanoplasty), or manageable with a conventional or bone-anchored hearing aid.
Sensorineural hearing loss (air and bone conduction elevated together, gap ≤ 10 dB): damage lies in the cochlea (most common — outer hair cell loss from noise or aging) or the auditory nerve/central pathway (retrocochlear — e.g., vestibular schwannoma, requiring MRI referral especially if asymmetric). Generally permanent; managed with hearing aids, cochlear implants for severe-profound loss, and assistive listening technology — not surgically reversible in the vast majority of cases.
Mixed hearing loss (air-bone gap present superimposed on an elevated bone conduction baseline): both mechanisms co-exist, e.g., otosclerosis developing in an ear that already has age-related sensorineural loss. Management addresses the conductive component first (often surgically) since that may unmask better residual hearing before committing to amplification strategy for the sensorineural component.
A completed audiogram is only clinically useful once it is translated into a report and a decision: does this patient need hearing aids, medical/surgical evaluation, further diagnostic testing, or simple reassurance and monitoring? Standardized reporting ensures referring physicians and other audiologists can act on the results without re-testing.
A complete audiology report includes: the audiogram itself with all symbols and masking notations, PTA per ear, word recognition/speech discrimination scores (percentage correct on phonetically balanced word lists at a comfortable loud level — cross-checks that the pure tone thresholds predict real-world speech understanding), tympanometry (middle ear pressure/compliance, screening for effusion or perforation independent of behavioral thresholds), and acoustic reflex thresholds where relevant.
Speech Recognition Threshold (SRT) — the softest level at which spondee words are correctly repeated 50% of the time — should agree with the PTA within about 6–8 dB; a larger discrepancy flags either non-organic (functional) hearing loss, patient inconsistency, or a retrocochlear/central processing issue requiring further workup rather than face-value acceptance of the audiogram.
Certain audiometric patterns trigger mandatory referral regardless of the patient's subjective complaint: asymmetric sensorineural loss (≥15 dB difference between ears at two or more frequencies) warrants MRI to rule out vestibular schwannoma; sudden onset (within 72 hours) sensorineural loss is a medical emergency needing ENT evaluation and often systemic or intratympanic corticosteroids within the first 1–2 weeks for the best chance of recovery; a significant air-bone gap with normal otoscopy suggests otosclerosis, referred for surgical stapedectomy candidacy; and any conductive loss with abnormal tympanometry or otoscopic findings (effusion, perforation, cholesteatoma) goes to ENT before any hearing aid fitting is considered.
For confirmed permanent hearing loss without a treatable conductive component, the pathway moves to amplification: PTA-based candidacy for hearing aids typically begins around 25–40 dB HL depending on patient need and lifestyle, while bilateral severe-to-profound sensorineural loss with poor word recognition scores (<50–60% aided) triggers cochlear implant candidacy evaluation.
Universal newborn hearing screening, mandated in the US since the early 2000s (otoacoustic emissions or automated auditory brainstem response before hospital discharge), reduced the average age of confirmed hearing loss diagnosis from roughly 2.5–3 years to under 3 months — critically, this falls within the first-year "critical period" for spoken language development, and children diagnosed and fitted with amplification/intervention before 6 months of age show significantly better language outcomes than those diagnosed later.