A Stimarium Dossier
The right amount of noise can sharpen a sense
A touch too faint to feel becomes easier to detect when a random buzz is added to the skin. The buzz now and then lifts the faint signal over the point where a nerve fires. Too little buzz does nothing, and too much drowns the signal. The effect is called stochastic resonance.
What’s known, what’s not: Known, Our read, Open
Noise helps people find a touch too faint to feel
In 1996, at Boston University, volunteers had a small probe pressed against their skin, too faintly for them to detect reliably. Then James Collins and colleagues added a random vibration to the stimulus. At the right strength, people detected the faint touch more often. [S02]
The same noise made a touch they could already feel harder to detect. The noise helped only the touch that was too weak on its own. [S02]
Physicists call this stochastic resonance: a benefit from randomness. It can appear wherever a weak signal meets a threshold, a line the signal has to cross to count. [S05]
The noise helped only the touch that was too weak on its own.
How a random jiggle carries a steady signal
In a simple threshold model, a nerve fires when its input crosses a line. A weak, regular signal that stays below that line produces no firing on its own. [S05]
Add random noise and the input jitters. The jitter crosses the line most often when it rides on the signal's peaks, so the firing starts to follow the signal's rhythm. [S01, S05]
The benefit has a peak. With too little noise, the signal rarely gets over the line. With too much, the nerve fires at random and the signal is lost again. Detection rises with noise, then falls. [S01, S05]
Detection rises with noise, then falls.
Animals that detect more with noise in the water
The first demonstration in a living sensor came in 1993. Frank Moss and colleagues moved a crayfish tail fan back and forth through water about ten times a second, near the limit its sensory hairs could detect. A middling dose of added random water motion made the nerve signal follow that movement better. Stronger noise made it worse. [S01]
In 1999 the group turned to paddlefish, which find plankton by sensing the plankton's faint electric fields with a long snout. With a noisy electric field of the right strength in the tank, the area in which the fish detected plankton grew wider. [S04]
Swarms of the plankton the fish eat give off electrical noise of their own. The study showed the effect in a tank; whether wild paddlefish depend on it is untested. [S04]
Clear for balance in the lab, mixed beyond it
In 2003, Collins's team set vibrating gel insoles at 90 percent of each foot's threshold, so the buzz was too faint to feel. Twelve older adults, average age 73, and fifteen young adults stood on them. With the noise on, the older group swayed less on every measure the team took. [S03]
A month of seated insole use at home did not improve walking speed in the study tests. In a 2022 randomized trial, which assigned participants to groups by chance, older adults with type 2 diabetes used the insoles while sitting for 22 minutes a day, with vibration at 90 percent of each foot's threshold. After a month, walking tests found no improvement, and the trial's balance measures showed no difference either. [S07]
Hearing results are mixed. A 2009 study traced an apparent noise benefit in a listening task to how people chose their answers. A 2020 study found no consistent evidence for stochastic resonance in its auditory tasks, with responses differing across people. [S06, S09]
A month of seated insole use at home did not improve walking speed in the study tests.
Noise as an ingredient
The idea began far from biology. In 1981, three Italian physicists, Roberto Benzi, Alfonso Sutera and Angelo Vulpiani, described it in a simplified climate model, as a way small, regular changes in Earth's orbit might tip the planet in and out of ice ages. [S08]
Whether nervous systems evolved to use their own internal noise this way is an open question. Mark McDonnell and Derek Abbott argued in 2009 that it would be more surprising if the brain did not exploit noise than if it did. [S05]
A sense with a threshold has a use for noise, within a narrow band. A little randomness can lift a faint, regular signal over the threshold. Each sensor has its own best amount, and beyond it the noise gets in the way again.
A sense with a threshold has a use for noise, within a narrow band.
Think with the idea.
Take me sideways
A stuck drawer often opens with a jiggle when a steady pull fails. Jiggle too hard and you lose your grip. How much shaking helps open it depends on how stuck the drawer is.
Analogy.Challenge it
The cleanest results come from animal sensors and short lab tests. In people, hearing results are mixed, one apparent gain came from how people answered, and a month of insole use did not speed walking.
Scope warning.Trace it
Set it beside the lime lumps in Roman concrete, long blamed on sloppy mixing, that can seal cracks. In both, the part that looks like a defect does the work, and only within limits.
Cross-domain synthesis, not a finding of the cited studies.Take it with you
When a method keeps missing a faint pattern, ask whether some variation would help it catch the pattern before assuming the only fix is more precision.
Reflection prompt, not medical advice.What’s known, what’s not
↑ Back to where you were- Added random vibration helped people detect a touch too faint to feel, and made an already-felt touch harder to detect. [S02]
- In crayfish sensory hairs, moderate noise made nerve signals follow a weak water movement better; stronger noise made them worse. [S01]
- Insoles buzzing below what people could feel reduced older adults’ sway in a lab test. In older adults with diabetes, a month at home changed neither walking nor sway. [S03, S07]
We think noise, dosed to suit each sensor, can help a signal cross a threshold, the level needed for detection. In people, we read the benefit as demonstrated for touch and balance in the lab, and unproven beyond it.
- Do nervous systems put their own internal noise to this use, or only noise added from outside?
- Could a buzz tuned to each person's threshold reduce falls, not just sway?
What the three labels meanClose explanations
Known marks what the cited evidence directly supports, including what named researchers argue.
Our read is Stimarium’s own interpretation of that evidence.
Open marks a question the cited evidence can’t answer yet, or a possibility we are raising ourselves.