Auditory Acuity

Abstract

Auditory acuity is a form of auditory perception: the sharpness with which the ear detects faint sounds and distinguishes one sound from another. It is operationalized chiefly as the absolute threshold, the quietest level a listener can reliably detect at each frequency, charted as the audiogram and summarized by the pure-tone average. This article treats auditory acuity as a worked case in psychophysics, covering how thresholds are measured, why adaptive staircase procedures converge efficiently on them, and how the audible field is bounded by equal-loudness contours relating physical intensity to perceived loudness. It also surveys the modern recognition that a clinically normal audiogram can mask cochlear synaptopathy, a hidden loss of the neural channel carrying sound to the brain, which has reshaped how acuity is understood and what its measures are presumed to capture.

Keywords: auditory acuity, absolute threshold, audiogram, psychophysics

Of all the senses, hearing is the one most often assessed by a single number: the faintest sound a person can detect. That number, the auditory threshold, is both the oldest measure in experimental psychology and the one on which a clinical diagnosis of hearing loss still turns. Yet the threshold is not a fixed property of the ear read off like a ruler. It is a statistical boundary estimated from a listener's responses, shaped by the method used to find it, and interpreted against a model of what the quiet end of hearing should look like. The study of auditory acuity is therefore a study of psychophysics applied to a practical problem, and the century of work it spans runs from the first equal-loudness measurements to the discovery that the standard measure can miss a real injury entirely.

What Auditory Acuity Is

Auditory acuity, in the sense the term carries in perception and audiology, is the keenness of hearing: the capacity to detect sounds that are faint and to tell apart sounds that differ. MeSH defines it as the ability to hear and distinguish one sound from another, assessed with whisper tests, tuning fork tests, or bone conduction tests. It is a specific capacity within the broader domain of auditory perception, the part concerned with the sheer sensitivity of the system rather than with the recognition of speech or music. Where auditory perception asks what a listener hears and understands, auditory acuity asks how little sound suffices for detection and how finely two sounds can be discriminated.

The dominant operationalization reduces this keenness to the absolute threshold: the lowest sound-pressure level, at a given frequency, that a listener detects on some criterion proportion of presentations. Because sensitivity varies across the audible range, acuity is not one number but a function of frequency, and measuring it means tracing that function. The resulting plot, the audiogram, charts threshold in decibels of hearing level against frequency, with the reference zero at each frequency set to the median threshold of healthy young adults, so that a listener's audiogram reads directly as a departure from normal acuity (Carhart & Jerger, 1959). A single summary of the audiogram, the pure-tone average of thresholds at 500, 1000, and 2000 Hz, has long served as the one-number index of acuity and the threshold on which the grading of hearing loss is built (Moore, 2012). Figure 1 shows a schematic audiogram with the standard severity bands against which an individual's thresholds are read.

Figure 1

The Audiogram and Standard Severity Bands

Measuring Auditory Acuity

The clinical measurement of acuity inherited its logic from the psychophysics of the nineteenth century but had to be made fast and reproducible enough for routine use. The procedure that became standard, the Hughson-Westlake method, presents a pure tone at a clearly audible level and then lowers it in 10 dB steps until the listener fails to respond, after which it is raised in 5 dB steps; the threshold is taken as the lowest level at which the listener responds on at least half of the ascending trials (Carhart & Jerger, 1959). The asymmetry of the steps, down in tens and up in fives, is deliberate: it brackets the threshold quickly from above and then approaches it finely from below, trading a little efficiency for robustness against momentary lapses of attention. The first demonstration lets thresholds be set across the audiogram and reads off the resulting pure-tone average and its severity grade.

Detection alone does not exhaust acuity. A listener with a normal audiogram may still struggle to resolve a talker from a noisy background, and this frequency selectivity, the ability of the ear to separate nearby frequencies into distinct channels, is a second component of keen hearing. It is captured by the concept of the auditory filter, a bandpass channel centered on each frequency whose width sets the finest frequency difference the ear can resolve there. Measurements of these filters give the equivalent rectangular bandwidth, which grows with center frequency and quantifies how the cochlea trades frequency resolution for coverage across the spectrum (Glasberg & Moore, 1990). The width of the auditory filter is not independent of detection: broader filters admit more masking noise, raising thresholds in background sound even when thresholds in quiet are normal, which is one reason a single pure-tone audiogram underdescribes a listener's real-world acuity. Table 1 sets out the principal measures of auditory acuity and what each captures.

The Psychophysics of the Threshold

The threshold the audiogram reports is a statistical construct, not a sharp cutoff. If a tone is presented repeatedly at a fixed low level, a listener detects it on some trials and not others, so the probability of detection rises gradually with level along a psychometric function rather than jumping from zero to one. The threshold is defined as the level corresponding to some chosen point on that function, conventionally a probability partway between chance and certainty. This means any threshold estimate inherits the variability of the underlying responses, and the efficiency of a measurement is a question of how to place trials so as to estimate that point with the fewest presentations. Signal detection theory pressed the point further, questioning whether a fixed sensory threshold exists at all: it showed that a detection response reflects both the listener's sensitivity and a decision criterion that shifts with expectation and payoff, so that what older methods recorded as a threshold partly measured a willingness to respond rather than the sensitivity of the ear alone (Swets, 1961). Modern threshold methods are designed to neutralize that criterion, which is why a forced choice between intervals is often preferred to a simple yes-no judgment.

The classical methods of limits and constant stimuli spend many trials at levels far from the threshold, where responses are nearly certain and so carry little information. Adaptive procedures fix this by letting the listener's responses steer the level. In the transformed up-down method, the level descends after a run of correct detections and ascends after a miss, so that the track hovers around the point it is designed to find. A two-down one-up rule, dropping the level after two correct responses and raising it after one failure, converges on the level detected about 71 percent of the time, and the threshold is estimated by averaging the levels at the reversals, the points where the track changes direction (Levitt, 1971). Because the procedure concentrates trials near the threshold and the reversal average is insensitive to the starting level, it is both efficient and largely free of the expectation bias that plagues fixed-sequence methods. The second demonstration runs such a staircase and shows it closing in on a hidden threshold. The entire enterprise rests on the older insight that sensation grows with the logarithm of stimulus intensity, which is why acuity is measured in the logarithmic decibel in the first place, and why equal ratios of sound pressure, not equal differences, correspond to equal steps of sensation (Stevens, 1957).

Loudness and the Audible Field

Detection marks the lower edge of a two-dimensional region, the audible field, within which hearing operates. The field is bounded below at each frequency by the threshold of hearing and above by the threshold of discomfort, and its shape across frequency is the fullest picture of acuity. The foundational mapping of this field came from the measurement of equal-loudness contours: curves tracing the sound-pressure level that, at each frequency, matches the loudness of a reference tone at 1000 Hz. Fletcher and Munson's measurements established that these contours are not flat, so that the ear is far more sensitive in the 2 to 4 kHz region than at low or very high frequencies, and that the contours flatten at high levels, meaning the frequency dependence of loudness is itself level-dependent (Fletcher & Munson, 1933). Loudness is indexed in phons, the level in decibels of the matching 1000 Hz tone, so that by definition a contour passing through 40 dB at 1000 Hz is the 40-phon contour. Perceived loudness, as distinct from the phon scale that merely ranks it, grows as a power function of intensity, the basis of the sone scale on which a doubling of the number corresponds to a doubling of subjective loudness (Stevens, 1957).

The ear's uneven sensitivity across frequency is not an incidental fact but a structural feature with consequences for acuity. The heightened sensitivity around 3 to 4 kHz, which arises partly from the resonance of the ear canal, is also the region where noise-induced damage first appears, producing the characteristic notch in the audiogram of a listener with a history of loud exposure. The spacing of the equal-loudness contours also explains why frequency selectivity and loudness interact: within a critical band, the roughly constant-ratio frequency region over which the ear integrates energy, tones add their loudness differently than they do across bands, and the width of these bands, measured by their masking of one another, tracks the auditory filters that set frequency resolution (Zwicker et al., 1957). The third demonstration plots the equal-loudness contours and shows how the audible field narrows at the frequencies where acuity is poorest.

Worked Example

Consider a right ear with pure-tone thresholds of 30 dB HL at 500 Hz, 40 dB HL at 1000 Hz, and 50 dB HL at 2000 Hz, of the kind the first demonstration computes. The standard pure-tone average takes the arithmetic mean of these three values, which is the sum, 120, divided by three, giving 40 dB HL. On the common clinical scale, where normal acuity extends to 25 dB HL, mild loss spans 26 to 40, and moderate loss 41 to 55, a pure-tone average of 40 dB HL sits exactly at the upper boundary of mild loss. The listener detects conversational speech but misses its quieter consonants, the pattern a threshold in this range predicts.

Now add the 4000 Hz threshold, 60 dB HL, and compute the four-frequency average that many clinicians prefer because it is more sensitive to high-frequency loss. The sum is now 180 across four frequencies, giving 45 dB HL, which crosses into the moderate band. The single fact this makes vivid is that the grade assigned to a listener's acuity depends on which frequencies the summary includes: the same ear is classified as mild by the three-frequency average and moderate by the four-frequency average, because the loss is concentrated at the high frequencies that the three-frequency average omits. The choice of summary is therefore not a neutral bookkeeping decision but one that can change the clinical category, which is why the audiogram itself, not any single average, remains the primary record of acuity (Moore, 2012).

Discussion

The measurement of auditory acuity has moved from treating the threshold as a property of the ear to treating it as one window, and an incomplete one, onto a system that fails in more ways than sensitivity alone. For most of the twentieth century the audiogram was effectively the definition of hearing: a normal audiogram meant normal hearing, and a diagnosis of loss was a diagnosis of raised thresholds. The psychophysical tradition had always known that detection was only part of the story, since frequency selectivity, temporal resolution, and loudness growth are measurable and partly independent, but the audiogram's practical convenience let it stand for acuity as a whole (Moore, 2012). The clinical reliance on a single number was reinforced by the real correlation between elevated thresholds and the everyday difficulties that bring people to audiologists, chief among them difficulty understanding speech in noise, which specialized tests such as the QuickSIN and the Words-in-Noise test were developed to quantify directly rather than infer from the audiogram (Killion et al., 2004; Wilson, 2011).

The limits of the threshold model became impossible to ignore with the discovery of cochlear synaptopathy. Animal work showed that noise exposure causing only a temporary threshold shift, from which the audiogram fully recovers, can permanently destroy a large fraction of the synapses between the inner hair cells and the auditory nerve, silently removing much of the neural channel that carries sound to the brain (Kujawa & Liberman, 2009). Because the surviving fibers suffice to detect quiet tones, the audiogram stays normal, yet the degraded neural representation is expected to impair hearing precisely where listeners most often complain, in noisy settings where the full population of fibers is needed (Liberman & Kujawa, 2017). This mechanism gave a physiological name to the long-suspected category of hidden hearing loss, a deficit in supra-threshold processing invisible to the standard measure of acuity (Plack et al., 2014). Whether and how synaptopathy manifests in human listeners with normal audiograms remains actively contested, but its existence has already redefined what the threshold does and does not certify (Lobarinas et al., 2017).

Current Directions

Contemporary work on auditory acuity develops along three fronts. The first is the search for a human marker of cochlear synaptopathy, testing whether electrophysiological and behavioral measures can detect, in listeners with clinically normal audiograms, the supra-threshold deficit that animal models predict; results so far are mixed, and the field is working to establish which measures are sensitive and specific enough to serve as a diagnostic (Lobarinas et al., 2017). The second reframes acuity in terms of the cognitive cost of listening rather than detection alone: the Framework for Understanding Effortful Listening treats degraded acuity as something a listener compensates for by recruiting attention and working memory, so that two people with identical audiograms can differ sharply in the effort hearing demands of them (Pichora-Fuller et al., 2016). The third situates acuity within public health, following epidemiological findings that hearing loss is among the largest modifiable risk factors for later-life cognitive decline and that its global burden, measured in years lived with disability, is rising as populations age (Lin et al., 2011; Livingston et al., 2020; GBD 2019 Hearing Loss Collaborators, 2021). Running through all three is a single revision of the founding assumption: auditory acuity is no longer equated with the threshold, but understood as a layered capacity of which detection is only the most easily measured layer.

Common Misconceptions

A normal audiogram means hearing is undamaged.

Cochlear synaptopathy can destroy a large share of auditory-nerve synapses while leaving pure-tone thresholds normal, because the surviving fibers still detect quiet tones. The audiogram certifies detection acuity, not the integrity of the whole auditory system (Kujawa & Liberman, 2009; Plack et al., 2014).

The auditory threshold is a fixed level the ear either passes or fails.

Detection probability rises gradually with level along a psychometric function, so the threshold is a chosen point on that curve estimated from variable responses, not a sharp cutoff (Levitt, 1971).

The ear is equally sensitive to all frequencies.

Equal-loudness contours show the ear is far more sensitive around 2 to 4 kHz than at low or very high frequencies, and the shape of that sensitivity changes with level, which is why loudness and acuity depend jointly on frequency and intensity (Fletcher & Munson, 1933).

Glossary

Absolute threshold.

The lowest level of a stimulus, at a given frequency, that a listener detects on some criterion proportion of presentations; the primary measure of detection acuity.

Adaptive procedure.

A psychophysical method in which the stimulus level on each trial depends on the listener's previous responses, concentrating trials near the threshold to estimate it efficiently.

Audiogram.

A plot of hearing threshold in decibels of hearing level against frequency, the standard clinical record of auditory acuity.

Auditory acuity.

The keenness of hearing: the ability to detect faint sounds and to distinguish one sound from another, quantified chiefly by the absolute threshold at each frequency.

Auditory filter.

A bandpass channel of the cochlea centered on a given frequency whose width sets the finest frequency difference the ear can resolve there; its breadth is reported as the equivalent rectangular bandwidth.

Cochlear synaptopathy.

The loss of synapses between the inner hair cells and the auditory nerve, caused by noise or aging, which degrades the neural representation of sound while often leaving the audiogram normal.

Critical band.

The frequency region, of roughly constant ratio, over which the ear integrates acoustic energy; tones within one band mask and combine differently than tones in separate bands.

Decibel of hearing level (dB HL).

A scale of sound level referenced at each frequency to the median threshold of healthy young adults, so that a listener's threshold reads directly as a departure from normal acuity.

Equal-loudness contour.

A curve tracing the sound-pressure level that, at each frequency, matches the loudness of a reference tone at 1000 Hz; the set of contours maps how sensitivity varies across the audible range.

Frequency selectivity.

The ear's ability to separate nearby frequencies into distinct channels; a component of acuity independent of detection sensitivity and central to hearing in noise.

Hidden hearing loss.

A deficit in supra-threshold auditory processing, such as that from cochlear synaptopathy, that impairs hearing in noise while leaving the pure-tone audiogram within normal limits.

Phon.

The unit of loudness level, equal to the sound-pressure level in decibels of an equally loud 1000 Hz tone; it labels the equal-loudness contours.

Psychometric function.

The curve relating the probability of detecting or discriminating a stimulus to its level; the threshold is defined as a chosen point along it.

Pure-tone average.

The arithmetic mean of a listener's thresholds at 500, 1000, and 2000 Hz; the standard one-number summary of the audiogram and the basis for grading hearing loss.

Reversal.

A point in an adaptive staircase where the direction of level change switches; the mean of the reversal levels estimates the threshold.

Signal detection theory.

The framework holding that a detection response reflects both sensory sensitivity and a decision criterion, so that a measured threshold mixes the keenness of the ear with the listener's willingness to respond.

Sone.

The unit of perceived loudness on a ratio scale, defined so that a doubling of the number corresponds to a doubling of subjective loudness; loudness grows as a power function of intensity.

Key Researchers

Harvey Fletcher

(1884-1981). Physicist at Bell Telephone Laboratories whose measurements of equal-loudness contours and the audible field, with Wilden Munson, established the frequency dependence of loudness and laid the quantitative foundation for audiometry. Wikipedia - Wikidata

M. Charles Liberman

(contemporary). Professor of Otolaryngology at Harvard Medical School and former director of the Eaton-Peabody Laboratories whose discovery of cochlear synaptopathy, with Sharon Kujawa, revealed the hidden hearing loss that a normal audiogram can conceal. ORCID - Wikidata

Frank R. Lin

(contemporary). Otolaryngologist and epidemiologist at Johns Hopkins University and director of the Cochlear Center for Hearing and Public Health whose studies linked hearing loss to cognitive decline and reframed acuity as a public-health concern. Faculty Page

Brian C. J. Moore

(b. 1946). Emeritus Professor of Auditory Perception at the University of Cambridge whose work on auditory filters, frequency selectivity, and loudness models formalized the components of acuity beyond the pure-tone threshold. ORCID - Wikipedia - Wikidata

S. S. Stevens

(1906-1973). Director of the Psycho-Acoustic Laboratory at Harvard University who formulated the power law relating subjective loudness to intensity and defined the sone scale, grounding the psychophysics on which acuity measurement rests. Wikipedia - Wikidata

Frequently Asked Questions

What is auditory acuity?

Auditory acuity is the keenness of hearing, the ability to detect faint sounds and to distinguish one sound from another. It is measured chiefly as the absolute threshold at each frequency, plotted as the audiogram, and summarized by the pure-tone average (Carhart & Jerger, 1959).

How is auditory acuity measured?

The standard clinical method presents a pure tone and lowers it in 10 dB steps until the listener stops responding, then raises it in 5 dB steps, taking the threshold as the lowest level detected on at least half of the ascending trials (Carhart & Jerger, 1959).

What is the pure-tone average?

The pure-tone average is the arithmetic mean of a listener's thresholds at 500, 1000, and 2000 Hz. It is the standard one-number summary of the audiogram and the basis for grading hearing loss from mild to profound (Moore, 2012).

Why are hearing thresholds measured in decibels?

Sensation grows roughly with the logarithm of stimulus intensity, so equal ratios of sound pressure correspond to equal steps of sensation. The logarithmic decibel captures the enormous range of audible intensities in a manageable scale (Stevens, 1957).

What is an adaptive staircase procedure?

It is a method in which the stimulus level follows the listener's responses, descending after correct detections and ascending after misses. A two-down one-up rule converges on the level detected about 71 percent of the time, estimated by averaging the reversal levels (Levitt, 1971).

What are equal-loudness contours?

They are curves tracing the sound-pressure level that, at each frequency, matches the loudness of a 1000 Hz reference tone. They show the ear is most sensitive around 2 to 4 kHz and far less sensitive at low and very high frequencies (Fletcher & Munson, 1933).

Can hearing be impaired with a normal audiogram?

Yes. Cochlear synaptopathy can destroy many auditory-nerve synapses while leaving pure-tone thresholds normal, producing hidden hearing loss that degrades hearing in noise without changing the audiogram (Kujawa & Liberman, 2009; Plack et al., 2014).

Why does hearing loss matter beyond the ear?

Epidemiological work identifies hearing loss as a major modifiable risk factor for later-life cognitive decline and a large and growing contributor to the global burden of disability as populations age (Livingston et al., 2020; GBD 2019 Hearing Loss Collaborators, 2021).

References