Touch

Abstract

Touch is a type of sensation: the sense that detects mechanical contact, vibration, and skin stretch through specialized receptors embedded in the skin. Four kinds of low-threshold mechanoreceptor encode different features—steady pressure, flutter, vibration, and stretch—and send signals along fast myelinated nerves to the somatosensory cortex, where the body surface is laid out as a distorted map that devotes the most territory to the most sensitive regions. A separate, slower system of unmyelinated afferents carries the affective, pleasant quality of gentle stroking. This article sets out the mechanoreceptors and the molecular channel that makes them sensitive to force, the pathway from skin to cortex and its map, the distinct affective-touch system, and how tactile acuity is measured, distinguishing touch as the sensory ability from its interpretation.

Keywords: touch, mechanoreceptor, somatosensory cortex, tactile acuity

A fingertip drawn across a tabletop reads almost everything about the surface—its grain, its edges, whether it is dry or damp, warm or cold—before the eyes are turned to it. Touch is the sense that reads the world through direct contact, the only sense distributed across the entire body surface rather than gathered into a single organ, and the first to develop in the womb. For most of its history it was studied as a map of spots and thresholds on the skin; only in recent decades have the receptors that give it its acuity, the molecular channel that makes them fire, and the separate pathway that makes a caress feel pleasant been worked out, turning a sense once charted with calipers into a clear account of how mechanical force becomes feeling (Abraira & Ginty, 2013; McGlone, Wessberg, & Olausson, 2014).

What Touch Is

Touch, or the tactile sense, is the detection of mechanical stimulation of the skin: pressure, vibration, stretch, and the movement of objects across the body surface. It is one branch of the broader somatosensory system, which also carries temperature, pain, and the body's sense of its own position, and it is classified in MeSH as a form of sensation, the registration of physical stimuli by the sensory receptors. Unlike vision or hearing, touch has no single organ; its receptors are spread through the skin of the entire body, which makes it the sense with the largest and most variable sensory surface.

A useful distinction runs through the whole subject: touch is really two senses sharing a skin. Discriminative touch resolves the fine spatial and temporal detail that lets the hand read a texture or a Braille character, and it is fast, precise, and carried by thick myelinated nerves. Affective touch registers the emotional quality of contact—the comfort of a caress—and it is slow, diffuse, and carried by thin unmyelinated fibers (McGlone, Wessberg, & Olausson, 2014). The sections that follow build the sense from the skin upward: first the mechanoreceptors and the molecular channel that make them sensitive to force, then the route from skin to cortex and the map it builds, then the separate affective system, and finally how touch is measured.

The Mechanoreceptors of the Skin

The acuity of touch rests on four types of low-threshold mechanoreceptor in the skin, and their division of labor follows two simple dimensions: how quickly each stops firing when a steady stimulus is held, and how large a patch of skin each one listens to (Johnson, 2001). Slowly adapting receptors keep firing throughout a sustained indentation and so report steady pressure and form; rapidly adapting receptors fire only when the stimulus changes and so report movement and vibration. Crossed with receptive-field size—small and sharply bounded, or large and diffuse—these two properties define the four afferent classes. Merkel cells are slowly adapting with small fields and encode fine spatial detail and edges; Meissner corpuscles are rapidly adapting with small fields and detect low-frequency flutter and micro-slip; Pacinian corpuscles are rapidly adapting with large fields and are exquisitely tuned to high-frequency vibration; Ruffini endings are slowly adapting with large fields and signal skin stretch (Abraira & Ginty, 2013).

What makes any of these receptors respond to force at all is a single class of molecule. Coste and colleagues identified Piezo1 and Piezo2 as the mechanically activated cation channels that open when a membrane is deformed, converting physical pressure directly into an electrical current (Coste et al., 2010). Woo and colleagues then showed that Piezo2 is the channel Merkel cells use to transduce touch (Woo et al., 2014), and Chesler and colleagues found that people lacking functional Piezo2 are profoundly impaired at discriminative touch while their sense of pain is spared, proving the channel's necessity in humans (Chesler et al., 2016). The first demonstration lets the reader apply a stimulus to each receptor type and watch how its firing pattern and adaptation differ.

Table 1 sets out the four mechanoreceptor types, their adaptation and receptive-field properties, and the feature each one encodes.

From Skin to Cortex

The four mechanoreceptors feed their signals into large myelinated nerve fibers—the Aβ afferents—that conduct quickly, on the order of tens of meters per second, which is what lets touch guide a movement in close to real time (Johansson & Flanagan, 2009). These fibers enter the spinal cord and ascend the dorsal column–medial lemniscus pathway to the thalamus and on to the primary somatosensory cortex in the parietal lobe. The defining feature of that cortex is that it is laid out as a map of the body: adjacent patches of skin project to adjacent patches of cortex, so the body surface is reproduced as an orderly, if distorted, image.

Figure 1 shows the central fact of that map. It is not drawn to the body's real proportions but to the density of its innervation: the fingertips, lips, and tongue, packed with mechanoreceptors, command vastly more cortical territory than the trunk or legs, which is why those regions feel so much more finely. This disproportion—the sensory homunculus—was first charted by Penfield and Boldrey from electrical stimulation of the cortex in awake neurosurgical patients (Penfield & Boldrey, 1937).

Figure 1

The Sensory Homunculus: Cortical Area Scales with Innervation Density

The map is not fixed. When the pattern of input changes—after the loss of a finger, or with extensive practice at a tactile skill—the cortical territories reorganize, expanding for heavily used regions and shrinking for silent ones, which makes the somatosensory map one of the clearest demonstrations of cortical plasticity in the adult brain. How much cortex a region commands tracks how densely its skin is innervated, a quantity Corniani and Saal mapped across the whole human body, from the richly supplied fingertip to the sparsely supplied back (Corniani & Saal, 2020). The second demonstration lets the reader select a body region and see how its innervation density, its tactile acuity, and its share of cortex rise and fall together.

Affective Touch

For most of its history the physiology of touch was the physiology of discriminative touch—the fast, myelinated system that resolves spatial detail. But the skin carries a second, quite different population of fibers. Olausson and colleagues used microneurography to record from unmyelinated, slowly conducting afferents in human hairy skin and found that they respond best not to pressure or vibration but to slow, gentle stroking at the speed and temperature of a caress, and that their activity projects to the insular cortex rather than the usual somatosensory areas (Olausson et al., 2002). Löken and colleagues then showed that the firing rate of these C-tactile afferents tracks how pleasant people rate a stroking stimulus, peaking at intermediate stroking velocities (Löken et al., 2009).

This is the neural basis of affective touch: a dedicated channel whose job is not to tell what is touching the skin but how it feels to be touched, carrying the emotional and social value of gentle contact. It explains why a caress and a poke of identical force feel utterly different, and why affectionate touch is so central to bonding and well-being. The existence of two parallel touch systems—a fast discriminative one and a slow affective one, with separate receptors, separate fibers, and separate cortical targets—is one of the more striking findings of modern sensory neuroscience (McGlone, Wessberg, & Olausson, 2014).

Measuring the Sense

Because tactile experience is private, measuring it means tasks that externalize it, and the classic measure is spatial acuity: how finely the skin can resolve two nearby points as two rather than one. In the two-point discrimination task, a pair of contacts is applied at a known separation and the observer reports whether they feel one point or two; the separation at which they are reliably told apart is the two-point threshold. That threshold varies enormously across the body, from roughly two millimeters on the fingertip to several centimeters on the back, tracking the innervation density and cortical magnification of each region (Mancini et al., 2014). Mancini and colleagues mapped this acuity for both touch and pain across the whole body surface and found the two senses share the same gradient, finest at the extremities and coarsest on the trunk.

As with any threshold, the transition from one point to two is not abrupt but graded, described by a psychometric function—the curve relating the probability of reporting two points to the separation between them. The third demonstration runs such a task, letting the reader choose a body region and a separation and read off the predicted probability that two points are felt. The two-point threshold is the historical standard, but it is not a clean measure of spatial resolution: Craig and Johnson argued that observers exploit non-spatial cues—a difference in felt intensity between one contact and two—so the threshold overestimates true acuity, and they recommended the grating-orientation task, in which a grooved surface is rotated until its ridges can no longer be told from a smooth one, as the more valid test (Craig & Johnson, 2000). Modern work goes beyond thresholds to model the full peripheral signal: Saal and colleagues built a simulation that reproduces the millisecond-precise spike trains of all the tactile afferents in the hand for any touched object, turning tactile acuity from a single number into a complete population code (Saal, Delhaye, Rayhaun, & Bensmaia, 2017).

Worked Example

The central facts of touch become concrete when the numbers are worked, and the demonstrations reproduce this arithmetic. Take the mechanoreceptors first. Their division of labor is a clean two-by-two: each type is either slowly or rapidly adapting, and has either a small or a large receptive field. Two adaptation rates times two field sizes gives exactly four combinations, and those four are the four afferent classes—Merkel (slow, small), Meissner (fast, small), Pacinian (fast, large), and Ruffini (slow, large). The scheme is complete: there is no fifth low-threshold class because there is no fifth cell in the table.

Now spatial acuity. The two-point threshold is about 2 millimeters on the fingertip and about 40 millimeters on the back, a twenty-fold difference across the body surface, and this ratio is why a coin can be identified by the fingertips but not pressed to the back. Model the probability of feeling two points as a logistic function of separation, p equals one divided by one plus e to the minus (x minus m) over s, where x is the separation, m is the two-point threshold for that region, and s sets the steepness. On the fingertip, with m equal to 2 millimeters and s equal to half a millimeter, a separation of 1 millimeter gives p about 0.12, a separation of 2 millimeters gives p exactly 0.5, and a separation of 3 millimeters gives p about 0.88. The same curve on the back, with m equal to 40 millimeters, is shifted twenty-fold to the right: a 1-millimeter separation is hopeless there, felt as a single point.

Finally the cortical map. Because cortical area scales with innervation density rather than with physical size, a region's share of somatosensory cortex is roughly proportional to how densely its skin is supplied with receptors. The fingertip, with the highest innervation density on the body, commands a share of cortex out of all proportion to its small surface area, while the far larger trunk, sparsely innervated, commands very little. Acuity, innervation density, and cortical magnification rise and fall together, which is the single thread tying the receptor count on the skin to the two-point threshold to the size of the hand in the homunculus.

Discussion

Touch entered the modern era as a sense charted by its thresholds—maps of pressure spots, two-point distances, vibration limits—and leaves it as a sense understood from the molecule up. The identification of the Piezo channels gave the mechanoreceptors a physical basis for responding to force, and the demonstration that people without Piezo2 lose discriminative touch while keeping pain established that this single channel is the gateway for tactile sensation in humans (Coste et al., 2010; Woo et al., 2014; Chesler et al., 2016). The four-receptor scheme, the fast Aβ pathway, and the magnified cortical map together explain how mechanical contact becomes the fine spatial sense that guides the hand (Johnson, 2001; Johansson & Flanagan, 2009).

Two features keep touch from being a solved problem. The first is the discovery that it is not one sense but two: alongside the discriminative system runs the affective C-tactile system, with its own receptors, fibers, and cortical target, whose job is to register the emotional value of contact rather than its spatial detail (Olausson et al., 2002; Löken et al., 2009). The second is the active hand. Touch is rarely passive; we explore objects with deliberate movements, and Lederman and Klatzky showed that people use stereotyped exploratory procedures—lateral motion for texture, pressure for hardness, enclosure for shape—each tuned to extract one property efficiently (Lederman & Klatzky, 1987; Lederman & Klatzky, 2009). The sense that reads the world by contact is molecularly simple at its receptors and cognitively rich at the moving hand, and the distance between the two is where the psychology of touch now lives.

Current Directions

The most active recent front is the molecular and anatomical dissection of the touch periphery. Neubarth and colleagues resolved how Meissner corpuscles are built and innervated, showing that two distinct afferent types intermingle within a single corpuscle to underlie gentle-touch perception, a finer-grained picture than the one-receptor-one-fiber scheme (Neubarth et al., 2020). Handler and Ginty reviewed how the whole mechanosensory system is assembled and activated, consolidating the molecular account of touch that the Piezo discoveries opened (Handler & Ginty, 2021). These advances turn the classic four-receptor scheme from a functional classification into a developmental and molecular one.

A second front is engineering touch back into people who have lost it. Saal and colleagues' whole-hand simulation of tactile afferent activity is a step toward biomimetic feedback: if the exact spike trains the hand would produce can be computed for any contact, they can in principle be delivered to a prosthetic limb's sensors or directly to the nerves, restoring a sense of touch that feels natural rather than buzzing or tingling (Saal, Delhaye, Rayhaun, & Bensmaia, 2017). The open questions now run from the molecule to the machine: how the receptors are specified in development, how the affective and discriminative streams are combined centrally, and how faithfully an artificial touch can reproduce the natural population code.

Common Misconceptions

Touch is a single, uniform sense.

Touch is at least two parallel senses: a fast, myelinated discriminative system that resolves spatial detail, and a slow, unmyelinated affective system that signals the pleasant quality of gentle contact, with separate receptors, fibers, and cortical targets (McGlone, Wessberg, & Olausson, 2014).

The whole skin is equally sensitive.

Tactile acuity varies about twenty-fold across the body, from roughly two millimeters on the fingertip to several centimeters on the back, tracking innervation density and cortical magnification rather than being uniform (Mancini et al., 2014).

The homunculus reflects the body's real proportions.

The cortical map is scaled to innervation density, not physical size. The fingertips and lips command far more cortex than the much larger trunk because their skin is far more densely supplied with receptors (Penfield & Boldrey, 1937).

There is one kind of touch receptor.

Glabrous skin carries four distinct low-threshold mechanoreceptors—Merkel, Meissner, Pacinian, and Ruffini—that differ in how fast they adapt and how large their receptive fields are, each encoding a different feature of contact (Johnson, 2001).

Glossary

Aβ afferent.

A large, fast-conducting myelinated nerve fiber that carries signals from the skin's mechanoreceptors to the spinal cord, enabling touch to guide movement in near real time.

Affective touch.

The emotional, pleasant quality of gentle contact, carried by a slow unmyelinated system distinct from the discriminative touch that resolves spatial detail.

C-tactile afferent.

A thin, slowly conducting unmyelinated fiber in hairy skin, tuned to slow gentle stroking and projecting to the insular cortex; the basis of affective touch.

Cortical magnification.

The disproportionate share of somatosensory cortex devoted to densely innervated body regions such as the fingertips and lips, the basis of the distorted homunculus.

Discriminative touch.

The fast, precise, myelinated touch system that resolves fine spatial and temporal detail, as in reading a texture or a Braille character.

Exploratory procedure.

A stereotyped hand movement tuned to extract one object property—lateral motion for texture, pressure for hardness, enclosure for shape—in active haptic exploration.

Mechanoreceptor.

A sensory receptor that responds to mechanical deformation of the skin—pressure, vibration, or stretch—and converts it into a neural signal.

Meissner corpuscle.

A rapidly adapting mechanoreceptor with a small receptive field, detecting low-frequency flutter and the micro-slip that guides grip (RA1).

Merkel cell.

A slowly adapting mechanoreceptor with a small receptive field, encoding steady pressure, edges, and fine spatial form; transduces touch through the Piezo2 channel (SA1).

Pacinian corpuscle.

A rapidly adapting mechanoreceptor with a large receptive field, exquisitely tuned to high-frequency vibration and to distant events felt through a tool (RA2/PC).

Piezo2.

The mechanically gated ion channel that opens when a receptor membrane is deformed, converting skin pressure into an electrical current; necessary for discriminative touch in humans.

Rapidly adapting.

A receptor property of firing only when a stimulus changes, so reporting movement and vibration rather than steady contact.

Receptive field.

The patch of skin within which a stimulus will change a given receptor's firing; small fields give fine spatial resolution, large fields give coarse, diffuse sensitivity.

Ruffini ending.

A slowly adapting mechanoreceptor with a large receptive field, signaling skin stretch and thereby hand shape and finger position (SA2).

Sensation.

The registration of physical stimuli by the sensory receptors, the broad category under which touch is classified as a mechanical sense.

Slowly adapting.

A receptor property of firing throughout a sustained stimulus, so reporting steady pressure and form rather than only its onset.

Somatosensory cortex.

The region of the parietal lobe that receives tactile signals and represents the body surface as an orderly but distorted map, the sensory homunculus.

Two-point threshold.

The smallest separation at which two simultaneous contacts on the skin are reliably felt as two points rather than one; the classic measure of tactile spatial acuity.

Key Researchers

David D. Ginty

(b. 1962). Harvard Medical School (Howard Hughes Medical Institute); dissected the molecular and developmental biology of the mechanosensory neurons of touch, from the Piezo2-dependent Merkel-cell afferents to the intermingled afferents of the Meissner corpuscle. ORCID - Faculty Page - Google Scholar - Wikipedia - Wikidata

Roberta L. Klatzky

. Carnegie Mellon University; with Susan Lederman, established the exploratory-procedures framework of active haptic perception, showing how deliberate hand movements are tuned to extract specific object properties. ORCID - Faculty Page - Google Scholar - Wikipedia - Wikidata

Susan J. Lederman

. Queen's University (Kingston, Ontario); pioneered the study of haptic object recognition, demonstrating that touch is an active, exploratory sense and mapping the hand movements people use to perceive texture, shape, and material. Faculty Page - Google Scholar - Wikipedia - Wikidata

Ardem Patapoutian

(b. 1967). Scripps Research (Howard Hughes Medical Institute); discovered the Piezo1 and Piezo2 mechanically activated ion channels that let receptors convert force into electrical signals, work recognized with the 2021 Nobel Prize in Physiology or Medicine. ORCID - Faculty Page - Google Scholar - Wikipedia - Wikidata

Åke B. Vallbo

(b. 1933). University of Gothenburg; pioneered microneurography, the recording of single tactile afferents in awake humans, characterizing the mechanoreceptive units of the hand and, later, the C-tactile afferents of affective touch. Wikipedia - Wikidata

Frequently Asked Questions

What is touch?

Touch, or the tactile sense, is the detection of mechanical stimulation of the skin—pressure, vibration, stretch, and the movement of objects across the body. It is a form of sensation, distributed across the whole body surface rather than concentrated in one organ (Abraira & Ginty, 2013).

What are the receptors for touch?

Glabrous skin carries four low-threshold mechanoreceptors: Merkel cells for steady pressure and form, Meissner corpuscles for flutter, Pacinian corpuscles for high-frequency vibration, and Ruffini endings for skin stretch. They differ in how fast they adapt and how large their receptive fields are (Johnson, 2001).

How does the skin turn pressure into a nerve signal?

Through mechanically gated ion channels. Piezo2 opens when a receptor membrane is deformed, letting current flow and converting physical force into an electrical signal; people without functional Piezo2 lose discriminative touch (Coste et al., 2010; Chesler et al., 2016).

What is the sensory homunculus?

It is the map of the body in the somatosensory cortex, drawn not to physical proportions but to innervation density, so the fingertips and lips command far more cortex than the trunk. It was first charted by Penfield and Boldrey from cortical stimulation (Penfield & Boldrey, 1937).

Why are the fingertips so much more sensitive than the back?

Because tactile acuity tracks innervation density and cortical magnification. The fingertip has the densest supply of mechanoreceptors and the largest cortical representation, giving a two-point threshold around two millimeters, against several centimeters on the back (Mancini et al., 2014).

What is affective touch?

It is the emotional, pleasant quality of gentle contact, carried by a separate system of slow unmyelinated C-tactile afferents that respond best to a caress and project to the insular cortex, distinct from the fast discriminative system (Olausson et al., 2002; Löken et al., 2009).

How is tactile acuity measured?

Chiefly by the two-point threshold—the smallest separation at which two contacts are felt as two rather than one—which varies about twenty-fold across the body. Modern work models the full population of afferent spike trains the hand produces (Mancini et al., 2014; Saal, Delhaye, Rayhaun, & Bensmaia, 2017).

Is touch an active or a passive sense?

Largely active. We explore objects with deliberate movements, and Lederman and Klatzky showed that people use stereotyped exploratory procedures—lateral motion for texture, pressure for hardness, enclosure for shape—each tuned to extract one property efficiently (Lederman & Klatzky, 1987).

References