Sensory Systems · Unit 2: Touch

Meissner and Pacinian Corpuscles: Flutter and Vibration

What makes a Meissner or Pacinian afferent respond to change rather than to steady pressure, what sets the vibration frequencies it prefers, and what the brain does with the signal? This lesson starts from human flutter and vibration psychophysics, tests the classic idea that the Pacinian capsule is a mechanical filter against direct terminal recordings from duck corpuscles, follows the two afferents of the mouse Meissner corpuscle, and ends with vibration codes in mouse cortex and in freely moving mice. You will be able to compute an equal-pitch prediction from a fitted model and design the experiment that decides where rapid adaptation arises.

Watch first Is the Onion the Filter? · 3:10

Strip away the onion layers of a Pacinian corpuscle and you might expect the nerve ending inside to keep firing during a steady press. In isolated cat corpuscles in 1965 it mostly did not, and in duck corpuscles in 2025 the terminal, with its inner-core cells left in place after the outer layers were removed, still adapted in milliseconds.

So where does the brevity of a vibration receptor come from, what tunes it, and what does the brain make of the signal? The touch and pain lesson introduced Meissner and Pacinian afferents as rapidly adapting, with small and large receptive fields, and the LTMR lesson set them in the full class table. This lesson starts at mechanism. Two terms recur. A lamellar cell is a flattened, Schwann-cell-like cell wrapped around the nerve terminal inside a corpuscle. In the Pacinian corpuscle, lamellar Schwann cells hug the terminal in the inner core, and many concentric layers, the onion, make up the outer core.

AI-generated illustration: a cut-away Pacinian corpuscle drawn as an oval of nested concentric layers, the outer core, around a small inner core where flat lamellar Schwann cells hug a central nerve terminal. A thin axon leaves at the left. An inset zooms on the terminal and its lamellar cells.
A Pacinian corpuscle has an outer core of many concentric layers, the onion, around an inner core where lamellar Schwann cells hug the nerve terminal.

Flutter and hum: what human hands report

Put a vibrating probe on your fingertip and turn up the frequency. The percept changes kind. Talbot, Darian-Smith, Kornhuber and Mountcastle (1968, Journal of Neurophysiology) asked healthy adults to describe sinusoidal stimulation of the hand. At about 5 to 40 Hz they reported a light flutter. From around 60 to 80 Hz upward the sensation became a deeper, diffuse hum of vibration.

The same paper recorded single afferents from the median nerve of anaesthetized rhesus monkeys. That is a second preparation, not a measurement of the human subjects. The comparison tied flutter to the rapidly adapting afferents now linked to Meissner corpuscles and vibration to Pacinian afferents. Read the boundary between them as a description of two percepts, not a sharp receptor cut-off.

Frequency preferences also depend on species. Prsa and colleagues (2021, Nature Communications) found that behavioural vibration sensitivity peaked near 240 Hz at the human index fingertip and near 1000 Hz in their mouse forepaw preparation. “Pacinian tuning around 250 Hz” is therefore a human figure. Behavioural sensitivity is also not identical to the tuning of a single isolated afferent.

The onion that was not the filter: decapsulation then and now

Why does a Pacinian afferent fire at the start and end of a press but not during it? The classical answer was mechanical. Loewenstein and Mendelson (1965, Journal of Physiology) dissected Pacinian corpuscles from the cat mesentery and removed most of the lamellae. After decapsulation, the generator potential decayed much more slowly (later reviewers note this appeared mainly under intense compression). The impulses did not persist: the ending typically still fired a single spike. So adaptation had two stages. One lay in the mechanics around the terminal; another lay in the terminal’s spike-generating membrane, which stopped firing despite a lasting depolarization.

Loewenstein and Skalak (1966) then modelled the capsule. Think of a stack of water balloons: press slowly and the fluid flows sideways before the centre feels much; tap fast and the force reaches the core. In their model, fluid between the lamellae supplies viscosity and the lamellae supply elasticity, so the capsule suppresses low-frequency input and passes transients. For decades that high-pass filter was the textbook explanation of rapid adaptation.

The duck bill tested it directly. Ziolkowski and colleagues (2025, Science Advances) patch-clamped the afferent terminal inside Herbst corpuscles, the avian Pacinian, in bill skin from embryonic day 25 to 27 duck embryos. Removing the outer core left the rapid kinetics of the mechanotransduction current intact and did not significantly change the afferent’s frequency tuning. Slower inactivation appeared only as decapsulated terminals deteriorated. That last result offers one reading of 1965: a damaged, slowly dying terminal can look like a terminal that has lost its filter.

Keep the scope. This is a bird corpuscle, ex vivo, from late embryos. It shows the outer core is not needed for fast currents and tuning in that preparation; it does not show the capsule does nothing in a cat or a human hand. The same study found that depolarizing an inner-core lamellar Schwann cell by current injection lowered the afferent’s mechanical threshold, which moves attention inward, toward the cells that touch the terminal.

AI-generated illustration: two copies of a layered Herbst corpuscle with a patch pipette on the central nerve terminal. On the left the corpuscle is intact; on the right the outer layers are removed and only the inner core cells and terminal remain.
Ziolkowski and colleagues patch-clamped the terminal inside duck Herbst corpuscles, with and without the outer core. Removing it left the rapid kinetics of the transduction current intact in this ex vivo embryo preparation.

Two afferents in one Meissner corpuscle

The Meissner corpuscle sits in a dermal papilla of glabrous skin, a stack of lamellar cells with Aβ axons coiled between them. The usual cartoon has one axon per corpuscle. Neubarth, Emanuel, Ginty and colleagues (2020, Science) labelled mouse Meissner afferents by two genetic routes: Ret expression and a TrkB lineage label (cells that expressed TrkB during development, marked by an inducible Cre driver (TrkB-CreER, given tamoxifen), a gene that switches on a permanent reporter). In adult mice, most corpuscles in the paw pads held one afferent of each kind, intermingled. Both conducted in the Aβ range. Each kind tiled the skin with its own kind, its fields offset from the other’s.

AI-generated illustration: two skin cross-sections side by side, each with a dermal papilla holding a stack of flat lamellar cells. On the left a single axon coils between them; on the right two axons in different colours, one Ret-positive and one TrkB-lineage, are intermingled.
The textbook Meissner corpuscle has one axon. In mouse paw pads, Neubarth and colleagues found that most corpuscles held one Ret-positive and one TrkB-lineage afferent, intermingled between the lamellar cells.

The two differed in physiology. TrkB-lineage afferents were the more sensitive. Ret-positive afferents were less sensitive, and 4 of 10 recorded Ret-positive afferents also fired during maintained indentation. So “two types of rapidly adapting afferent” overstates the result: one subtype included units with a sustained response. And because TrkB protein was undetectable in adult terminals, the TrkB label marks a developmental lineage, not a current receptor.

Then the behaviour. Deleting TrkB from sensory neurons during development produced adult mice without the Meissner system. Mice were trained in an operant task to report a step indentation of the forepaw.

Worked problem, part (a): what does a doubled threshold mean? Adult control mice detected the push at a threshold of about 10 mN; mice lacking Meissner corpuscles, at about 20 mN (Neubarth et al. 2020).

  1. Ratio: 20 mN ÷ 10 mN = 2.0, a twofold rise.
  2. The mutants still had a measurable threshold, so they still detected touch.
  3. Conclusion: Meissner afferents contribute to gentle-touch detection, lowering the threshold by about half, but other afferents can support detection without them.

Two limits apply. The loss was developmental and TrkB was deleted in all sensory neurons (Advillin-Cre), not only Meissner afferents, so other afferents may have compensated or contributed, and an acute silencing could give a larger or smaller effect. And the task measured detection only, not discrimination of flutter or slip.

Accessory cells that feel force

If the terminal adapts without the outer layers, what are the lamellar cells for? Three papers from the Bagriantsev and Gracheva laboratories built the answer in steps, all in duck bill skin ex vivo.

  • 2020 (Nikolaev and colleagues, Science Advances; embryonic day 24 to 26). Lamellar cells of both Grandry and Herbst corpuscles produced mechanically activated currents. Grandry lamellar cells fired action potentials, predominantly through R-type calcium channels. Herbst outer-core cells were mechanosensitive but not excitable.
  • January 2023 (Ziolkowski, Gracheva and Bagriantsev, eLife). The Grandry terminal itself carried transduction currents at both the onset and the offset of indentation. The onset current inactivated in 8.95 ± 1.82 ms (mean ± SEM). Tetrodotoxin blocked the afferent’s action potentials. The terminal is a transducer in its own right.
  • Later 2023 (Nikolaev, Ziolkowski and colleagues, Science Advances). Selectively depolarizing one Grandry lamellar cell could trigger firing in its afferent, and calcium influx was required. Latencies ran from 20 ms to about 9 s. Chemical transmission was proposed but not shown.

So the corpuscle is bicellular: two mechanosensitive cells, each able to influence the spike. In duck the latencies are the problem. A 200 Hz vibration has a 5 ms cycle, so a signal that takes 20 ms or more cannot be what follows each cycle. Lamellar cells may set the terminal’s sensitivity rather than carry each cycle.

The mouse supplies the behavioural test. Chen and colleagues (2025, Science Advances; Huber laboratory) expressed optogenetic tools in mouse Pacinian lamellar Schwann cells. Activating them drove time-locked afferent spiking in anaesthetized adult mice, with first-spike latencies up to about 40 ms. The Etv1-CreER line marks inner-core lamellar cells but is also expressed in other cell types, so the manipulation is not perfectly cell-specific. Inhibiting them raised the hindpaw vibrotactile detection threshold of awake mice about fivefold. The authors note that optogenetic drive may not reproduce normal physiology. The channel in either cell is unproven (Piezo2 is the expected candidate in the terminal) and the signal from lamellar cell to terminal remains unidentified; the mechanotransduction channels lesson sets out what proof a transducer claim needs, and Merkel cells give a parallel case of a non-neuronal partner cell in touch.

Vibration in the brain and in the wild

What does cortex do with a Pacinian signal? Prsa, Morandell, Cuenu and Huber (2019, Nature) delivered vibrations to the mouse forelimb and recorded neurons in forelimb primary somatosensory cortex (S1). The responses depended on deep Pacinian input, were tuned to a feature combining frequency and amplitude, and did not follow the vibration cycle by cycle. That is a rate code for a composite feature, unlike the phase-locking of auditory nerve fibres described in hearing and balance.

Prsa and colleagues (2021) then measured the percept. Trained mice (C57BL/6, awake, head-fixed) and adult human volunteers compared vibrations, and in both species a change in amplitude shifted perceived pitch. For mice near a 450 Hz reference, the equal-pitch data were fitted with a power law (a fit to four mice):

P = f × (A / Aref)^0.32, with Aref = 5.6 µm

Worked problem, part (b): the frequency that keeps pitch constant. Reference: 450 Hz at 5.6 µm. Amplitude doubles to 11.2 µm. What frequency feels the same?

  1. Set pitches equal: 450 Hz × (5.6/5.6)^0.32 = f2 × (11.2/5.6)^0.32.
  2. Left side: 450 Hz × 1 = 450 Hz. Right side: f2 × 2^0.32.
  3. 2^0.32 = e^(0.32 × 0.693) = e^0.2218 = 1.248.
  4. f2 = 450 Hz ÷ 1.248 = 360.5 Hz.
  5. Change: (450 − 360.5) ÷ 450 = 0.199, a 19.9% decrease. Check: 360.5 × 1.248 = 449.9 Hz.
Equal-pitch matcher — amplitude shifts pitch⚙ original · interactive
  • equal-pitch curve f* (in this sketch)
  • reference, 450 Hz at 5.6 µm
  • your comparison vibration
Exponent used
0.32
(A/5.6)0.32
1.000
P = f × (A/5.6)0.32
450.0 Hz
Against the reference (450)
matched
Frequency that matches the reference at this A
f* = 450 ÷ 1.000 ≈ 450.0 Hz
Check
450.0 × 1.000 = 450.0 Hz

Reference point: the lesson's measured reference, 450 Hz at 5.6 µm.

In this sketch: the power-law fit P = f × (A/5.6 µm)0.32 near a 450 Hz reference in head-fixed C57BL/6 mice (a fit to four mice), assumed to hold from the reference amplitude up to double it (5.6 to 11.2 µm). Frequencies far from 450 Hz are extrapolation. The 1% band is a display tolerance in this sketch, not a perceptual threshold.

Rule to keep: the model is a local fit, so any amplitude other than 5.6 µm gives a prediction, not a measurement. Humans also showed amplitude-dependent pitch, but no human exponent is given, so none is applied here.

The amplitude ratio has no units, so the answer stays in hertz. This is a prediction of a local fit, assumed to hold at double amplitude; it is not a new measurement.

Finally, natural behaviour. Turecek and Ginty (2024, Neuron) recorded single Pacinian neurons in the L4 dorsal root ganglion of awake, freely moving mice. The neurons fired during the mouse’s own movements as well as during external vibration, and sensitivity varied widely between neurons. The most sensitive responded to substrate vibrations generated 2.5 m away, the furthest point tested on that surface, not a known limit. Whether the brain separates self-generated from external vibration was not tested. The dorsal horn lesson follows these signals into the spinal cord and brainstem, and prediction returns to how the brain discounts its own movements.

Design problem

Decapsulation of cat Pacinian corpuscles (1965) and duck Herbst corpuscles (2025) seem to give different answers about the outer core. Design an experiment that decides whether the lamellae filter the stimulus or the terminal itself adapts.

One way to do it
  1. Measure the right signal. Voltage-clamp the terminal inside one corpuscle and record the transduction current before and after removing the outer core, in the same corpuscle. Record spikes from the same fibre separately, so generator current and spiking are not confused.
  2. Control for damage. Track time after dissection and the terminal’s leak current. If adaptation slows only as leak rises, deterioration, not loss of a filter, explains the change. Include sham-handled corpuscles kept for the same time.
  3. Probe the filter directly. Deliver step indentations and sinusoids from 10 to 1000 Hz at matched displacement at the terminal (measured, for example, by imaging the inner core), so that any change in tuning can be assigned to mechanics or to the membrane.
  4. Test in a mammal. Duck results do not transfer automatically. In mouse Pacinian corpuscles, add optogenetic silencing of lamellar Schwann cells (as Chen and colleagues did in 2025) with an afferent recording and a behavioural detection threshold.
  5. Interpret. If the terminal adapts at the same speed and keeps its tuning once the outer core is gone, adaptation does not need the outer core (the inner-core cells are still present, so a separate test is needed to rule them out); if it slows only in healthy decapsulated terminals, the capsule filters. Human corpuscles cannot be patch-clamped this way, so human answers would need other methods, such as microneurography and psychophysics.

How we measure it

Terminal patch clamp with single-fibre recording

A pipette is sealed onto the afferent terminal inside an intact corpuscle while a second electrode records spikes from the same fibre. You see the transduction current itself and can ask whether its timing, rather than the capsule's, sets adaptation. So far this has been done in ex vivo duck bill skin.

Cre and lineage genetics with operant psychophysics

Genetic labels (Ret, a TrkB lineage tracer) mark afferent subtypes, a conditional deletion removes an end organ, and trained mice report whether they felt a calibrated push. The result shows what that end organ contributes to detection, within the limits of a developmental rather than acute loss.

Optogenetics in non-neuronal cells

A light-gated channel or pump expressed in lamellar Schwann cells lets you activate or inhibit them while recording the afferent or testing behaviour. Activation shows what the cells can do; inhibition shows what the behaviour needs, both under artificial drive.

Chronic DRG recording in freely moving mice

A stabilized electrode in the lumbar dorsal root ganglion records single sensory neurons while the mouse moves about freely. It asks what an afferent reports in natural life instead of during imposed sine waves.

Finish line

Test yourself

8 questions

Practice and MCAT-style, with explained answers.

Next · Lesson 7

Hairy Skin, Stroking and Affective Touch

Brush a forearm at a few centimetres per second and one slow, unmyelinated fibre fires hardest. Is that why it feels nice?

Continue →

← Lesson 5 · Merkel Cells: Touch Receptors That Are Not Neurons

Werner Loewenstein

Werner Loewenstein

With M. Mendelson, removed the outer lamellae of isolated cat Pacinian corpuscles and separated the prolonged generator potential from the brief impulse discharge (1965); with R. Skalak, modelled the capsule as a viscoelastic mechanical filter (1966).

By the 1960s the Pacinian corpuscle was a favourite preparation for asking how a sense organ turns pressure into nerve impulses, because a single large corpuscle with one axon could be dissected out of the cat mesentery and studied on its own. The puzzle was why its response was so brief. In a 1965 paper in the Journal of Physiology, Loewenstein and Mendelson stripped away most of the onion-like lamellae and pressed on what remained. The generator potential, the graded depolarization of the nerve ending, could now last much longer during a held compression. The impulses did not: the decapsulated ending still typically fired a single spike. So adaptation had at least two stages, one in the mechanics around the ending and one in the ending's own spike-generating membrane. The following year, with R. Skalak, Loewenstein built a mechanical model of the capsule in which fluid between the lamellae supplies viscosity and the lamellae supply elasticity, so that slow pressure is dissipated before it reaches the core. That model gave the field a testable claim about the capsule, which direct terminal recordings could later put to the test.

✓ verbatim Loewenstein & Mendelson, J. Physiol. 177:377–397 (1965); Loewenstein & Skalak, J. Physiol. (1966), doi 10.1113/jphysiol.1966.sp007827.
David Ginty

David Ginty

His laboratory showed that most mouse Meissner corpuscles are innervated by two genetically distinct Aβ afferents (Neubarth et al. 2020) and, with Josef Turecek, recorded Pacinian neurons in freely moving mice (2024).

The Ginty laboratory works on the low-threshold mechanoreceptors of the mouse, using genetic tools to label, record and remove one afferent type at a time. In the 2020 Science paper led by Nicole Neubarth and Alan Emanuel, the group labelled Meissner afferents by two genetic routes, Ret expression and developmental TrkB expression. Most corpuscles in the paw pads held one afferent of each kind, the two endings intertwined among the lamellar cells. The TrkB-lineage afferents were the more sensitive, and the fields of each kind tiled the skin while staying offset from the other kind. Deleting TrkB from sensory neurons during development left adult mice without the Meissner system, and their threshold for detecting a gentle push on the forepaw roughly doubled. In 2024 Josef Turecek and Ginty recorded single Pacinian neurons in the dorsal root ganglion of awake, freely moving mice and found them active during the animal's own movements and during vibration of a surface tapped as far as 2.5 m away. Together the papers made it possible to ask what each end organ contributes to behaviour rather than only to a nerve recording.

✓ verbatim Neubarth et al., Science 368:eabb2751 (2020); Turecek & Ginty, Neuron 112:3267–3277 (2024).
Daniel Huber

Daniel Huber

His laboratory, with Mario Prsa, recorded vibration responses in mouse forelimb somatosensory cortex (2019), measured amplitude-dependent vibrotactile pitch in mice and humans (2021), and with Y.-T. Chen and colleagues tested Pacinian lamellar Schwann cells in mouse behaviour (2025).

The Huber laboratory studies how the mouse cortex represents touch, and much of that work concerns vibration, a signal that the 2019 work traced to deep tissue and carried by Pacinian afferents. In a 2019 Nature paper, Mario Prsa and colleagues delivered vibrations to the forelimb of mice and recorded neurons in the forelimb region of primary somatosensory cortex. The neurons responded to a combination of frequency and amplitude rather than to frequency alone, did not follow each cycle of the vibration, and depended on Pacinian input. A 2021 paper trained mice and asked adult humans to compare vibrations, and found in both species that changing amplitude shifted perceived pitch. In 2025, Chen and colleagues turned to the corpuscle itself: optogenetic activation of lamellar Schwann cells drove the Pacinian afferent, and inhibiting them raised the vibration detection threshold of mice about fivefold. The work tied a cortical code, a percept and a non-neuronal cell to one sense.

✓ verbatim Prsa, Morandell, Cuenu & Huber, Nature 567:384–388 (2019); Prsa et al., Nat. Commun. (2021), doi 10.1038/s41467-021-25476-9; Chen et al., Sci. Adv. (2025), doi 10.1126/sciadv.adt5110.
Slav Bagriantsev

Slav Bagriantsev

With Elena Gracheva, used the touch-specialist duck bill to record mechanically activated currents from lamellar cells (2020) and directly from afferent terminals inside Grandry and Herbst corpuscles (2023, 2025).

Mammalian corpuscles are small and scattered, which makes their nerve terminals hard to reach with a patch pipette. The Bagriantsev and Gracheva laboratories turned to the bill of the domestic duck, which is packed with Grandry corpuscles (the avian counterpart of Meissner corpuscles) and Herbst corpuscles (the counterpart of Pacinian ones). Working on isolated bill skin from late-stage embryos, Nikolaev and colleagues reported in 2020 that the lamellar cells themselves produce mechanically activated currents, and that the lamellar cells of Grandry corpuscles fire calcium-dependent action potentials. Ziolkowski then patch-clamped the afferent terminal inside the corpuscle and recorded transduction currents at the onset and the offset of indentation (2023). A paired study showed that depolarizing one lamellar cell could make its afferent fire (2023), and a 2025 study removed the outer core of Herbst corpuscles and found that the terminal's fast currents and its tuning survived. The duck made it possible to measure, rather than infer, where in a corpuscle force first becomes current.

✓ verbatim Nikolaev et al., Sci. Adv. 6:eabe6393 (2020); Ziolkowski, Gracheva & Bagriantsev, eLife 12:e84179 (2023); Nikolaev et al., Sci. Adv. 9:eadi4147 (2023); Ziolkowski et al., Sci. Adv. (2025), doi 10.1126/sciadv.adt4837.
Test yourself
Question 1 of 8 Score 0 / 0

In Loewenstein and Mendelson's 1965 experiments on isolated cat Pacinian corpuscles, what changed after most of the lamellae were removed and the ending was held compressed?
Sources & Integrity

Every structure, slide, simulation, and clip on this page is listed below with the source it came from. None of the 3D structures is drawn by hand or by AI: each one loads by accession ID from RCSB PDB or PubChem. This panel is generated from the lesson's own data file, so what it lists is what the page loads.

Simulations

  • Equal-pitch matcher · Original interactive — Biochemistrypedia (Deppmann Lab).

Scientist profiles & quotes

  • Werner Loewenstein — Loewenstein & Mendelson, J. Physiol. 177:377–397 (1965); Loewenstein & Skalak, J. Physiol. (1966), doi 10.1113/jphysiol.1966.sp007827.
  • David Ginty — Neubarth et al., Science 368:eabb2751 (2020); Turecek & Ginty, Neuron 112:3267–3277 (2024).
  • Daniel Huber — Prsa, Morandell, Cuenu & Huber, Nature 567:384–388 (2019); Prsa et al., Nat. Commun. (2021), doi 10.1038/s41467-021-25476-9; Chen et al., Sci. Adv. (2025), doi 10.1126/sciadv.adt5110.
  • Slav Bagriantsev — Nikolaev et al., Sci. Adv. 6:eabe6393 (2020); Ziolkowski, Gracheva & Bagriantsev, eLife 12:e84179 (2023); Nikolaev et al., Sci. Adv. 9:eadi4147 (2023); Ziolkowski et al., Sci. Adv. (2025), doi 10.1126/sciadv.adt4837.

Quotes are verbatim from the cited source. Profile text is original to this site and was checked against the sources cited.

Animated explainer

Is the Onion the Filter? · narration: an AI narrator voice (Gemini, Charon) reading narration written from this lesson's text · animation: AI-generated in code (Manim) by Claude, a schematic with no drawn molecular structures, plots computed from their equations · captions from the narration script.

Images & figures

3 illustrations in this lesson are AI-generated: drawn by an image model from a written spec, then checked by a separate AI judge against the spec and the lesson text. None shows a chemical structure or data.

  • a cut-away Pacinian corpuscle drawn as an oval of nested concentric layers, the outer core, around a small inner core where flat lamellar Schwann cells hug a central nerve terminal. A thin axon leaves at the left. An inset zooms on the terminal and its lamellar cells.
  • two copies of a layered Herbst corpuscle with a patch pipette on the central nerve terminal. On the left the corpuscle is intact; on the right the outer layers are removed and only the inner core cells and terminal remain.
  • two skin cross-sections side by side, each with a dermal papilla holding a stack of flat lamellar cells. On the left a single axon coils between them; on the right two axons in different colours, one Ret-positive and one TrkB-lineage, are intermingled.

Every 3D structure is rendered live from an accession ID, and every plot is drawn from its equation in the browser.

integrity Machines check the mechanics. The build confirms that every local file exists, that every question's answer key points at a real choice, and that any AI-generated scientific figure names the person who checked it; a weekly job re-resolves every accession ID against its database. Questions are original; MCAT-style items are written for this site and are not real AAMC questions. Banned sources (the Tymoczko textbook, publisher decks, copyrighted figures) are not used.

Like I'm five

Under your fingertips are tiny touch sensors shaped like little onions and coils. Some notice a gentle flutter and some notice a fast buzz. Scientists used to think the onion layers did the work of ignoring slow pushes. New experiments in ducks and mice show the nerve tip inside, with its inner helper cells, still does much of that once the outer layers are gone, and those helper cells can feel touch too.
High school

Two kinds of touch organ in your skin report change rather than steady pressure. Meissner corpuscles respond best to slow vibrations that feel like flutter, and Pacinian corpuscles to fast ones that feel like a hum. For decades the Pacinian's layered capsule was thought to act like a shock absorber that blocks slow pushes. Recordings from duck corpuscles show the nerve ending inside keeps its fast response even with the outer layers removed, and the flattened cells wrapped around the ending can sense force and help trigger the nerve. Mouse studies show that each Meissner corpuscle usually holds two different nerve fibres and that losing these corpuscles makes gentle touches harder to detect.
College

Meissner and Pacinian afferents are rapidly adapting Aβ low-threshold mechanoreceptors. In human psychophysics (Talbot et al. 1968), 5 to 40 Hz vibration feels like flutter and higher frequencies feel like a deeper hum, matching the frequency ranges of the two afferent classes recorded in monkeys. The classical account located Pacinian adaptation in the capsule, a viscoelastic high-pass filter, although decapsulated cat endings still fired only one spike. Direct terminal patch clamp in duck Herbst corpuscles found fast transduction currents and tuning intact after outer-core removal, while inner-core lamellar Schwann cells are mechanosensitive in duck and can drive the afferent there, and optogenetic activation of mouse lamellar Schwann cells drives the afferent too. Mouse Meissner corpuscles usually contain a Ret-positive and a TrkB-lineage afferent, and their developmental loss doubles the forepaw detection threshold. In mouse S1, vibration responses combine frequency and amplitude, and in both mice and humans amplitude changes perceived pitch.
Graduate

Rapid adaptation in these end organs is now a multicellular problem. In duck Grandry terminals, onset transduction currents inactivate in about 9 ms (8.95 ± 1.82 ms, mean ± SEM) and a second current appears at offset; in duck Herbst corpuscles, removing the outer core leaves fast current kinetics and frequency tuning statistically unchanged, and slowing appears only as decapsulated terminals deteriorate, which may explain why older decapsulation work saw prolonged generator potentials. Lamellar cells carry their own mechanically activated currents, Grandry lamellar cells spike through mainly R-type calcium channels, and depolarizing one can trigger afferent firing after 20 ms to about 9 s, with calcium entry required. In mice, inhibiting Pacinian lamellar Schwann cells raises vibrotactile thresholds about fivefold. Centrally, mouse forelimb S1 responses reportedly encode a frequency-amplitude feature without cycle-by-cycle entrainment, and a fitted equal-pitch rule, P = f × (A/Aref)^0.32 near 450 Hz, captures amplitude-dependent pitch. Open question: how does a lamellar cell signal to the terminal fast enough to matter for each cycle of a 200 Hz vibration, and which channel transduces in the terminal and in the lamellar cell? The signal is unidentified and the channel unproven.
A duck bill as a touch laboratory

The duck bill is dense with Grandry and Herbst corpuscles, the bird versions of Meissner and Pacinian end organs. That density is why the Bagriantsev and Gracheva groups worked on bill skin taken from duck embryos close to hatching. In that tissue they could seal pipettes onto single lamellar cells, and later onto the afferent terminal inside an intact corpuscle while recording spikes from the same fibre. That made the bird a place to ask, by direct measurement, where in a corpuscle force first becomes current.

— Nikolaev et al., Sci. Adv. 6:eabe6393 (2020); Ziolkowski, Gracheva & Bagriantsev, eLife 12:e84179 (2023).
Feeling a tap 2.5 metres away

Turecek and Ginty recorded single Pacinian neurons in the lumbar dorsal root ganglion of awake mice that were free to move. The neurons fired during the mouse's own movements, and the most sensitive ones also responded to vibrations of a wooden surface (a branch) generated as far as 2.5 m away, the furthest distance tested on that surface. Sensitivity varied widely between neurons. The result reframes the Pacinian afferent as a reporter of the body's own movement and of distant events in the substrate, not only of a probe pressed on the skin.

— Turecek & Ginty, Neuron 112:3267–3277 (2024).
Flutter versus hum

The 1968 paper is often cited as Mountcastle's, but its first author is W. H. Talbot. It had two arms. Healthy adult volunteers described sinusoidal stimulation of the hand: at roughly 5 to 40 Hz it felt like a light flutter, and from around 60 to 80 Hz upward like a deeper, diffuse hum. Separately, the group recorded single afferents from anaesthetized rhesus monkeys. The comparison linked the two percepts to two afferent populations, a link drawn across species that later work in humans and mice has refined.

— Talbot, Darian-Smith, Kornhuber & Mountcastle, J. Neurophysiol. 31:301–334 (1968).

 ↑ Top