BPTA (Bayesian Pure Tone Audiometry)

How BPTA Differs from Traditional Audiometry

The traditional Hughson-Westlake procedure (used in most clinical audiology offices) uses a fixed algorithm: lower the volume until the person stops responding, then raise it back up. This staircase pattern converges on the threshold but requires many tone presentations per frequency.

BPTA replaces the staircase with a probabilistic model. After each response, the algorithm updates its estimate of where the threshold most likely is, then presents the next tone at the level that will yield the most useful information. This adaptive strategy typically reaches a reliable threshold estimate in fewer trials, reducing test time and listener fatigue.

Accuracy of BPTA for Screening

BPTA is designed for screening, not clinical diagnosis. In a quiet environment with calibrated headphones, BPTA results correlate well with clinical audiogram thresholds at the screened frequencies, typically within ±10 dB for most users.

Factors that reduce accuracy include background noise (masking the test tones), uncalibrated or low-quality headphones, and test-taker variability in response behavior. For these reasons, BPTA results should be understood as a screening estimate: sufficient to indicate whether further evaluation is warranted, but not a substitute for a clinical evaluation in a sound-treated booth.

What BPTA Measures

LSTN's Hearing Test uses a BPTA protocol to measure air conduction thresholds (the quietest tones detectable through headphones) at key frequencies covering the speech range. Results are classified per ear as normal, mild loss, moderate loss, or significant loss, consistent with AAA hearing loss classification standards.