Absolute threshold
Also known as: Detection threshold
The faintest stimulus a person can detect at all.
What it means
The absolute threshold is the minimum intensity of a stimulus needed for it to be detected, classically defined as the level detected on half of the trials because there is no single sharp cutoff. Detection probability rises gradually with intensity rather than switching on abruptly, reflecting noise in the nervous system and the observer. Signal detection theory later showed that this 'threshold' is not purely sensory: an observer's response criterion — willingness to say 'yes' — shifts the apparent threshold independently of true sensitivity. The concept anchors the measurement of sensory limits across vision, hearing, taste, smell, and touch. It pairs with the difference threshold to define the basic boundaries of perception.
How it is measured
You cannot ask for a threshold; you infer it from a pattern of yes and no. Classical psychophysics offered three routes: the method of limits, where intensity ascends and descends until the report flips; the method of constant stimuli, where fixed levels are presented in random order; and the method of adjustment, where the observer sets it. Each produces a psychometric function — detection probability against intensity — and the threshold is read off at an agreed criterion, conventionally the halfway point, though in two-interval forced choice the convention is 75%, midway between chance and certainty. Modern work uses adaptive staircases that place each trial near the current best estimate, converging in far fewer trials. Forced-choice designs, where the observer says which of two intervals held the signal, are preferred because the observer's willingness to say yes no longer moves the number.
What the evidence shows
Hecht, Shlaer and Pirenne (1942) measured the least light a dark-adapted eye needs, producing the field's most durable result: 54 to 148 quanta arriving at the cornea, of which only about 5 to 14 are absorbed by rods at their 60% seeing criterion, spread across roughly 500 rods so that no rod absorbs more than one. A rod must therefore respond to a single photon, and the threshold reflects how many such responses the system demands before committing to seeing. Later electrophysiology confirmed that rods do signal single photons. Whether the whole observer can is still contested: Tinsley and colleagues (2016) used a quantum light source and reported above-chance detection of single photons, but the statistics were challenged, the challenge was itself withdrawn, and no independent replication has settled it.
The threshold is not a constant
Treating the absolute threshold as a fixed property of a person is the commonest error. It moves with the state of the system and of the observer. Dark adaptation drops the eye's threshold by orders of magnitude over roughly forty minutes. Where attention is pointed, what the observer expects, what a miss costs relative to a false alarm, fatigue, and the preceding stimulus all shift it. It varies widely between people and drifts with age, most visibly in hearing at high frequencies. Signal detection theory's deeper claim is not just that criterion contaminates the measure. The classical high-threshold model — a gate below which nothing passes — fits the data badly. Vary what the observer has to lose and trace hit rate against false-alarm rate: the resulting ROC curve bends, where a gate plus guessing predicts a straight line.
Where it matters in practice
The threshold is the working unit of sensory measurement wherever a limit must be certified rather than described. Audiometry reports hearing thresholds in decibels against a young-normal reference, and hearing-aid fitting begins from that curve. Perimetry maps light thresholds across the visual field to catch glaucoma long before the patient notices a gap. Water utilities and food producers set limits on taint compounds against the detection thresholds of trained panels. Alarm and haptic design runs the logic backwards: a signal placed at threshold is missed half the time by definition, so warnings are engineered with a deliberate margin above it. The rule common to all of them: state the criterion and the method, because a threshold quoted without both is not comparable to anything.
Examples
On a clear, dark night the human eye can detect a candle flame from many kilometers away — near its absolute threshold for light.
Add sugar to water one grain at a time and there is no moment it flips to sweet; near the threshold you taste it on about half the sips and miss it on the rest.
A radiologist who missed a tumour last month starts calling more faint shadows real. Her eyes are no better; her willingness to say yes has moved, and with it the apparent threshold.
A glaucoma patient reports seeing normally, but perimetry finds a wedge of the visual field where the light must be far brighter than elsewhere before it is reported at all. The threshold map shows the loss the patient's own experience does not: subjective vision registers no gap, because nothing arrives to be noticed as missing.
Natural gas is odourless, so utilities add a sulphur compound at a concentration well above the average nose's detection threshold — the margin exists because thresholds vary widely between people.
First described in Foundational psychophysics; Fechner (1860).
Key references
- Tinsley, J. N., Molodtsov, M. I., Prevedel, R., Wartmann, D., Espigulé-Pons, J., Lauwers, M., & Vaziri, A. (2016). Direct detection of a single photon by humans. Nature Communications, 7, 12172. doi.org/10.1038/ncomms12172
- Watson, A. B., & Pelli, D. G. (1983). QUEST: A Bayesian adaptive psychometric method. Perception & Psychophysics, 33(2), 113-120. doi.org/10.3758/BF03202828
- Swets, J. A. (1961). Is there a sensory threshold? Science, 134(3473), 168-177. doi.org/10.1126/science.134.3473.168
- Tanner, W. P., Jr., & Swets, J. A. (1954). A decision-making theory of visual detection. Psychological Review, 61(6), 401-409. doi.org/10.1037/h0058700
- Hecht, S., Shlaer, S., & Pirenne, M. H. (1942). Energy, quanta, and vision. Journal of General Physiology, 25(6), 819-840. doi.org/10.1085/jgp.25.6.819