Neuroscience · Unit 3: Sensory systems

Principles of Sensory Coding

Every sense ends in the same currency: graded receptor potentials and trains of identical spikes. This lesson asks how that currency still tells you what, where and how much. You will follow a stimulus from transduction to a firing rate, read a tuning curve and a psychometric function, work out how lateral inhibition sharpens an edge in the Limulus eye, and weigh labelled-line against population codes in taste and smell.

Watch first Identical Spikes, Different Meanings · 3:06

In 1926 Edgar Adrian and Yngve Zotterman recorded the impulses from one stretch receptor in a frog muscle and found that a heavier load did not make the impulses bigger, only more frequent.

Every sense you have ends in that same currency. A fingertip ending, a rod in the retina, a taste cell and an olfactory neuron all produce graded voltages, and the afferent neurons they drive send trains of spikes that look alike. So how does a train of identical spikes say what touched you, where, and how hard? This lesson builds the answer from recordings: transduction, receptive fields, tuning, lateral inhibition and the old argument between labelled lines and populations.

The idea: a stimulus becomes a graded voltage, then a rate

Transduction is the conversion of a stimulus into an electrical signal at a sensory receptor. The first signal is the receptor potential, a change in membrane voltage whose size grows with stimulus strength. It is graded, like the synaptic potentials in integration and inhibition, and it spreads passively, as in the resting membrane.

The machinery differs by sense. In touch, force opens mechanically activated channels in the ending of the sensory neuron itself. In mice, Ranade and colleagues (Nature, 2014) deleted the channel Piezo2 from adult sensory neurons and Merkel cells, skin cells that contact some touch endings, and the animals lost much of their touch sensation. That result is about touch. Vertebrate photoreceptors use a receptor protein and a cGMP cascade, and light hyperpolarizes them; smell and taste also use receptor proteins. The cascades belong to senses and cGMP; here we follow what the voltage does next.

Where the ending is part of the afferent neuron, the receptor potential depolarizes the spike-initiating zone, and the larger it is, the faster the neuron fires action potentials. Where the receptor is a separate cell, as in photoreceptors and taste cells, it changes transmitter release onto the afferent instead. Either way, intensity reaches the brain mostly as a firing rate, which is what Adrian and Zotterman saw.

Which cells, which stimuli: receptive fields and adaptation

A receptive field is the region of the sensory surface, a patch of skin or of the visual field, where a stimulus can change a given cell’s activity. A fingertip afferent has a small one; an afferent from the back has a much larger one, which is one reason your fingertip resolves finer detail.

Receptors also differ in time. Rapidly adapting endings fire at the start of a maintained stimulus and then fall quiet; many fire again when it is removed. They report change: a Pacinian corpuscle in the skin signals vibration. Slowly adapting endings keep firing for as long as the stimulus lasts, as Merkel-cell afferents do during steady indentation, so they report the stimulus itself. Not every receptor stops firing under a constant stimulus. Adaptation as a general circuit motif is taken up in signaling circuits.

To summarize a cell’s preferences, plot its tuning curve: the mean firing rate against one stimulus parameter, such as the orientation of a bar or the frequency of a tone. Present the same stimulus many times and the spike count varies from trial to trial, so the curve is an average with spread around it.

Reading a tuning curve and a threshold

You might expect a neuron to be most useful at its peak. For fine discrimination, telling two nearby stimuli apart, it is often the reverse.

Worked problem 1, a tuning curve. In this made-up recording, a neuron fires 50, 40, 20 and 8 spikes/s at 0, 15, 30 and 45 degrees. Assume the trial-to-trial noise is low and the same at every angle.

Step 1: the change per step. From 0 to 15 degrees the rate falls by 10 spikes/s; from 15 to 30 degrees by 20 spikes/s; from 30 to 45 degrees by 12 spikes/s.

Step 2: the slope. The steepest segment is 15 to 30 degrees: 20 spikes/s per 15 degrees, about 1.3 spikes/s per degree, against about 0.7 near the peak.

Step 3: the reading. A small change near 22 degrees moves the rate about twice as much as the same change near the peak, so it is easier to detect. Butts and Goldman (PLoS Biology, 2006) showed this in low-noise models of fine discrimination, and also that as variability grows the best-encoded stimulus can move back toward the peak. The answer depends on noise and task.

Behaviour has its own curve. A psychometric function plots the proportion of correct responses against stimulus level. In a two-alternative forced-choice (2AFC) task the observer must pick which of two options held the stimulus, so guessing gives 50% correct. That makes 50% chance, not threshold, and the usual threshold sits halfway between chance and perfect, at 75%. (The 50% convention belongs to simple detection, where threshold is the level detected on half the trials.)

Worked problem 2, a psychometric function. Take a model 2AFC observer judging an angular difference x, with chosen, not measured, parameters (midpoint 2 degrees, spread 1 degree, no lapses):

p(x) = 0.5 + 0.5 × Φ((x − 2°) / 1°)

Φ is the standard normal cumulative distribution.

Step 1: threshold. Set p = 0.75. Then Φ(z) = (0.75 − 0.5)/0.5 = 0.5, so z = 0 and x = 2 degrees.

Step 2: 90% correct. Φ(z) = (0.90 − 0.5)/0.5 = 0.8, so z = 0.84 and x = 2.84 degrees.

Step 3: performance at 3 degrees. z = (3 − 2)/1 = 1 and Φ(1) = 0.841, so p = 0.5 + 0.5 × 0.841 = 0.921, or 92.1% correct.

Now the two curves can be compared. Britten, Shadlen, Newsome and Movshon (Journal of Neuroscience, 1992) trained rhesus monkeys to report the direction of moving dots and recorded single neurons in visual area MT on the same trials, with the display matched to each neuron’s receptive field and preferred direction. Under these conditions most neurons were about as sensitive as the monkey. That is a statement about signal. It does not show that one neuron made the choice: in a later model by the same group, pools of at least 100 weakly correlated neurons reproduced the observations best, and that number is a model estimate, not a count. Causal tests came from manipulation; Salzman, Britten and Newsome (Nature, 1990) biased monkeys’ choices by microstimulating MT.

Neighbours that suppress: lateral inhibition

The lateral eye of the horseshoe crab Limulus is a compound eye of ommatidia, each sending its own fibre into the optic nerve. Hartline, Wagner and Ratliff (Journal of General Physiology, 1956) recorded with a microelectrode inside one ommatidium. Lighting its neighbours suppressed its steady discharge, cutting the lateral connections to it abolished the inhibition, and the interaction exaggerated the difference between brightly and dimly lit regions. Think of each unit reporting its own light minus a share of its neighbours’ light.

Worked problem 3, an edge. This is a feedforward simplification. The Hartline-Ratliff model is recurrent (each unit is inhibited by its neighbours’ output, not their input) and has thresholds. Six receptors in a row receive excitation 10, 10, 10, 20, 20, 20 (made-up units). Each output equals its own excitation minus 0.2 times the excitation of each adjacent neighbour, and the row continues beyond both ends.

output = own excitation − 0.2 × (left neighbour + right neighbour)

Interior dim: 10 − 0.2 × (10 + 10) = 6.

Dim at the edge: 10 − 0.2 × (10 + 20) = 4.

Bright at the edge: 20 − 0.2 × (10 + 20) = 14.

Interior bright: 20 − 0.2 × (20 + 20) = 12.

The plateaus differ by 6 units (12 against 6), but across the edge the difference is 10 (14 against 4). The dim side dips below its plateau and the bright side overshoots, the same kind of edge exaggeration as the light and dark Mach bands people see at the ends of a luminance ramp (an abrupt step does not ordinarily show them, so this is an analogy). Limulus is not the vertebrate retina, whose surrounds are built by other cells; see the retina.

Lateral inhibition at a light step⚙ original · interactive
Interior dim
6
Dim at the edge
4
Bright at the edge
14
Interior bright
12
Plateau difference (interior bright − interior dim)
6
Across-edge difference (bright edge − dim edge)
10

In this sketch: k and the bright level are made-up values (the lesson uses k = 0.2, bright 20); each output = own excitation − k × (left + right neighbour), computed from the inputs (feedforward), no thresholds, row continuing past both ends, made-up units. Not the vertebrate retina.

This sketch is a feedforward simplification of the recurrent Hartline–Ratliff model of the Limulus eye, without thresholds; made-up units. Rule to keep: output = own excitation − k × (left neighbour + right neighbour).

Labelled lines or populations

In a labelled line, a fibre’s activity means one thing, say “sweet”, however it was activated. In a population code, quality is read from the pattern across many fibres. Taste researchers still argue between the two (Ohla and colleagues, Chemical Senses, 2019).

Smell leans toward populations at the receptor level. In olfactory sensory neurons, Malnic, Hirono, Sato and Buck (Cell, 1999) found that one receptor recognizes several odorants and one odorant is recognized by several receptors, so an odorant is coded by a combination. How neurons expressing the same receptor are wired to the same targets in the olfactory bulb is a developmental question, covered in partners, layers and self-avoidance.

Taste offers a test. Wu, Dvoryanchikov, Pereira, Chaudhari and Roper (Nature Communications, 2015) imaged taste neurons in the geniculate ganglion of anaesthetized mice expressing GCaMP3 and found that some neurons narrowly tuned at low concentrations became broadly tuned at higher ones.

Design problem

In that study, 28 of 101 neurons responded to more than one taste below the EC50 (the concentration giving half the maximal response), and 79 of 155 did so at about 1.5 to 2 times the EC50. A classmate says this refutes labelled lines. How would you evaluate the claim, and what experiment would test it?

One way to do it

Start with the counts. 28 of 101 is 28%, and 79 of 155 is 51%, but they come from different samples of neurons, so the rise could partly reflect which cells were imaged. The stronger evidence is the same neuron recorded at a low and a higher concentration (about 3.5 to 7 times the EC50 in that paired series), which the paper also shows for a subset of neurons.

Next, scope. Broad peripheral tuning says what the ganglion sends, not how the brain decodes it. A decoder could still read a narrowly tuned subset, or the combination.

Then test it causally. Silence or remove one receptor class, or one molecularly defined group of afferents, and measure behaviour at low and high concentrations. If labelled lines carry quality, losing the “sweet” class should remove sweet responses at both concentrations; if quality is read from patterns, the loss should depend on concentration and blur discrimination among tastes. The result is a scoped counterexample to strict, concentration-invariant labelled lines in mouse geniculate neurons, not a refutation of every labelled-line model.

How we measure it

Intracellular and single-fibre recording from receptors

A fine electrode inside a receptor, or on a nerve dissected down to one active fibre, records the receptor potential or the spike train of one sensory unit. Changing the stimulus, or the cell's neighbours, while recording shows how that one unit encodes it.

Single-unit recording paired with psychophysics

An animal trained on a perceptual task performs while one neuron is recorded. From the spike counts you build a neurometric function and compare it with the animal's psychometric function on the same trials, which tests whether the neuron carries enough signal to account for the behaviour.

Calcium imaging of sensory ganglia

Neurons expressing a genetically encoded calcium indicator such as GCaMP3 brighten when they fire. Imaging a ganglion while you apply stimuli records the tuning of many neurons at once, though calcium signals are slower and less direct than recorded spikes.

Conditional knockout with behavioural and afferent readouts

Deleting a gene only in chosen cells of an adult animal, such as Piezo2 in sensory neurons and Merkel cells, then testing behaviour and recording single afferents, links one molecule to transduction in the cell and to perception in the animal.

Finish line

Test yourself

8 questions

Practice and MCAT-style, with explained answers.

Next · Lesson 10

The Retina

A rod in the dark is depolarized and releasing glutamate. A flash of light makes it stop. How does turning a cell off let you see?

Continue →

← Lesson 8 · Synaptic Plasticity

PIEZO2 — the mechanically activated channel behind much of touch
PIEZO2 — the mechanically activated channel behind much of touch ● LIVE · RCSB PDB 6KG7
open on RCSB PDB ↗
Loading live structure 6KG7 from RCSB PDB…

Three curved blades surround a central pore. The blades are thought to bend the membrane, which is one proposed way that force on the membrane could open the channel.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Rhodopsin — the photoreceptor's light-absorbing receptor
Rhodopsin — the photoreceptor's light-absorbing receptor ● LIVE · RCSB PDB 1F88
open on RCSB PDB ↗
Loading live structure 1F88 from RCSB PDB…

Seven transmembrane helices hold retinal in a pocket. Absorbing a photon changes retinal's shape and starts the cascade that turns light into a change in membrane voltage.

Drag to rotate · scroll or pinch to zoom · right-drag to pan
Edgar Adrian

Edgar Adrian

With Yngve Zotterman, recorded impulses from a single sensory ending and showed that a stronger stimulus raised the frequency of identical impulses, and that the frequency fell while a steady stimulus was held: rate coding and adaptation.

By the 1920s physiologists knew that nerves carried electrical impulses, but a whole nerve mixes the activity of many fibres, so no one could say what a single sensory ending sent to the brain. Edgar Adrian, at Cambridge, combined a capillary electrometer with valve amplification sensitive enough to see the small impulses of single fibres. In 1926, working with the Swedish physiologist Yngve Zotterman, he cut successive strips from a frog muscle until its activity came from one stretch receptor. Each impulse was the same size whatever the load. What changed with a heavier load was how often the impulses came, and during a maintained stretch the frequency fell. Adrian shared the 1932 Nobel Prize in Physiology or Medicine with Charles Sherrington for their work on the functions of neurons. The single- fibre record made it possible to ask, for any sense, what the spike train of one receptor says about the stimulus.

✓ verbatim Adrian & Zotterman, J. Physiol. (1926), 'The impulses produced by sensory nerve-endings', Part 2; Adrian, Nobel Lecture, 12 December 1932 (nobelprize.org).
H. Keffer Hartline

H. Keffer Hartline

With Henry Wagner and Floyd Ratliff, showed in the lateral eye of the horseshoe crab Limulus that each receptor unit is inhibited by its illuminated neighbours, and that this lateral inhibition exaggerates differences between light and dark regions.

Hartline trained in medicine and turned to the physiology of vision, and he chose an animal whose eye could be taken apart one unit at a time. The lateral eye of Limulus, the horseshoe crab, is a compound eye built of ommatidia, each with a receptor unit whose axon runs in the optic nerve, so a single fibre reports a single unit. In the 1956 Journal of General Physiology paper with Henry Wagner and Floyd Ratliff, a microelectrode inside one ommatidium recorded a steady discharge to steady light. Lighting neighbouring ommatidia suppressed that discharge. Cutting the lateral connections to the ommatidium abolished the inhibition, which placed it in a network of connections among the units rather than in the optics. The interaction exaggerated the difference in response between brightly and dimly lit regions of the eye. Hartline and Ratliff went on to describe the inhibition with equations in which each unit is inhibited by its neighbours' output. Hartline shared the 1967 Nobel Prize in Physiology or Medicine with Ragnar Granit and George Wald. The work made contrast enhancement a measurable computation done by a circuit.

✓ verbatim Hartline, Wagner & Ratliff, J. Gen. Physiol. 39:651–673 (1956); Hartline, Nobel Lecture, 12 December 1967 (nobelprize.org).
William Newsome

William Newsome

With Kenneth Britten, Michael Shadlen and J. Anthony Movshon, recorded single neurons in visual area MT of rhesus monkeys while the monkeys judged motion, and found that most neurons were about as sensitive as the animal under matched conditions.

The question was old: how does the activity of neurons in a sensory area relate to what an animal can perceive? Britten, Shadlen, Newsome and Movshon trained rhesus monkeys to report the direction of motion in a display of moving dots, in which the experimenters could set what fraction of dots moved together. While a monkey worked, they recorded a single neuron in the middle temporal area, MT, and set the display to fit that neuron's receptive field and preferred direction. Because neuron and monkey saw the same display on the same trials, the two sensitivities could be compared directly. Under these conditions most neurons were about as sensitive as the monkey. A 1996 paper from the same group then modelled how pooled MT signals could produce the choices, and the simulations fitted the observations best with pools of at least 100 weakly correlated neurons. Microstimulation experiments from the Newsome laboratory tested the link by biasing MT activity. The work made it possible to put neuronal and behavioural sensitivity on one scale.

✓ verbatim Britten, Shadlen, Newsome & Movshon, J. Neurosci. 12:4745–4765 (1992); Shadlen, Britten, Newsome & Movshon, J. Neurosci. 16:1486–1510 (1996).
Ardem Patapoutian

Ardem Patapoutian

His laboratory identified the Piezo proteins as mechanically activated ion channels, and with Sanjeev Ranade and colleagues showed that removing Piezo2 from adult mouse sensory neurons and Merkel cells caused a profound loss of touch.

Touch begins when force opens an ion channel in the membrane of a sensory ending, but for decades no one could say which protein did it in mammals. The Patapoutian laboratory at Scripps Research looked for a protein whose loss removed mechanically activated currents from a cultured cell line, and in 2010 reported two, Piezo1 and Piezo2, as components of mechanically activated channels. The next question was whether Piezo2 mattered in an animal. In the 2014 Nature paper led by Sanjeev Ranade, the group deleted Piezo2 from the sensory neurons and Merkel cells of adult mice. The mice showed a profound loss of touch sensation, and single primary afferents were recorded in an ex vivo preparation of skin with its nerve attached. Patapoutian shared the 2021 Nobel Prize in Physiology or Medicine with David Julius. The work gave touch a molecular entry point: a channel that can be removed, mutated and recorded, linking one protein to the firing of an afferent and to behaviour.

✓ verbatim Coste et al., Science 330:55–60 (2010); Ranade et al., Nature 516:121–125 (2014); Patapoutian, Nobel Lecture, December 2021 (nobelprize.org).
Test yourself
Question 1 of 8 Score 0 / 0

A receptor cell is stimulated at three increasing strengths, and its membrane is recorded before any spike is generated. How does its receptor potential change?
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.

Molecular structures live data

  • PIEZO2 — the mechanically activated channel behind much of touch — RCSB PDB 6KG7 · coordinates CC0 · Mouse PIEZO2, cryo-EM structure (Wang et al., Nature, 2019)
  • Rhodopsin — the photoreceptor's light-absorbing receptor — RCSB PDB 1F88 · coordinates CC0 · Bovine rhodopsin, X-ray crystal structure (Palczewski et al., Science, 2000)

Simulations

  • Lateral inhibition at an edge · Original interactive — Biochemistrypedia (Deppmann Lab).

Scientist profiles & quotes

  • Edgar Adrian — Adrian & Zotterman, J. Physiol. (1926), 'The impulses produced by sensory nerve-endings', Part 2; Adrian, Nobel Lecture, 12 December 1932 (nobelprize.org).
  • H. Keffer Hartline — Hartline, Wagner & Ratliff, J. Gen. Physiol. 39:651–673 (1956); Hartline, Nobel Lecture, 12 December 1967 (nobelprize.org).
  • William Newsome — Britten, Shadlen, Newsome & Movshon, J. Neurosci. 12:4745–4765 (1992); Shadlen, Britten, Newsome & Movshon, J. Neurosci. 16:1486–1510 (1996).
  • Ardem Patapoutian — Coste et al., Science 330:55–60 (2010); Ranade et al., Nature 516:121–125 (2014); Patapoutian, Nobel Lecture, December 2021 (nobelprize.org).

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

Animated explainer

Identical Spikes, Different Meanings · 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

This lesson carries no raster figures. Every structure is rendered live from RCSB PDB or PubChem, 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

Your skin, eyes, nose and tongue have special cells that notice things like a poke, light or a smell. Each one turns what it notices into a tiny electrical change, and nerves carry it to your brain as clicks. A bigger poke makes the clicks come faster. Your brain tells what you touched by which cells are clicking and how fast.
High school

A sensory receptor converts a stimulus, such as pressure or light, into a change in membrane voltage called a receptor potential, and its size grows with the stimulus. That voltage sets how often action potentials fire along a sensory nerve, so stronger stimuli give higher firing rates. Each receptor responds only to stimuli in its own patch of skin or of the visual world, its receptive field. Some receptors keep firing while a stimulus lasts, others only when it changes. Neighbouring receptors can suppress one another, which makes edges stand out.
College

Transduction channels or receptor cascades turn a stimulus into a graded receptor potential, which is encoded as a firing rate in the afferent. A unit's receptive field and its tuning curve (mean rate against a stimulus parameter) say what it responds to, and rapidly or slowly adapting endings say whether it reports change or steady state. Because fine discrimination depends on how fast the rate changes, a tuning curve's flank can carry more information about small changes than its peak. Lateral inhibition, shown directly in the Limulus eye, subtracts neighbours' activity and exaggerates edges. Quality can be coded by labelled lines, in which a fibre's activity means one thing, or by patterns across a population; smell is combinatorial at the receptor level, and taste afferents in mice can broaden their tuning as concentration rises.
Graduate

Sensory coding links three measurable things: the stimulus-response function of single units (tuning curves with their trial-to-trial variability), the decoder that reads a population, and the observer's psychometric function. Butts and Goldman (2006) showed that in low-noise fine-discrimination models the best-encoded stimuli lie on steep flanks, but that as variability grows the most informative stimulus can move toward the peak, so "where a neuron is most useful" depends on noise and task. Britten and colleagues (1992) matched displays to single MT neurons in rhesus monkeys and found neurometric sensitivity similar to psychometric sensitivity; Shadlen and colleagues (1996) then modelled pools of at least 100 weakly correlated neurons, because correlated noise does not average away. In awake mice, part of the ongoing activity in visual cortex tracks the animal's own behaviour (Stringer and colleagues, 2019), so "noise" may be signal about something else. An open question: for taste, where single geniculate neurons in mice broaden their tuning with concentration, does the brain read labelled lines, population patterns or a mixture that changes with stimulus strength?
One receptor, one fibre, many impulses

To hear one sensory ending, Adrian and Zotterman had to make sure only one was active. They cut successive strips from a frog muscle until the record showed impulses of a single size, the sign of a single active fibre. Then they loaded the muscle with weights. The impulses never grew; their frequency did, and during a maintained stretch it declined. Those two observations, intensity carried by frequency and a response that fades under a constant stimulus, remain the first two facts taught about sensory coding.

— Adrian & Zotterman, J. Physiol. (1926), 'The impulses produced by sensory nerve-endings', Part 2; Adrian, Nobel Lecture (1932).
Cutting the connection removes the inhibition

A correlation between neighbours' light and a unit's suppressed firing could have come from stray light or optics. Hartline, Wagner and Ratliff tested the circuit directly. They recorded the steady discharge of one Limulus ommatidium, lit its neighbours, and saw the discharge fall. When they cut the lateral connections to the recorded ommatidium, the neighbours' light no longer inhibited it. The cut turned an association into a test: the inhibition depended on connections between units, which is what made lateral inhibition a property of the network.

— Hartline, Wagner & Ratliff, J. Gen. Physiol. 39:651–673 (1956).
Smell is spelled with several letters

Malnic, Hirono, Sato and Buck recorded which olfactory sensory neurons responded to a set of odorants and identified the receptor each responsive neuron expressed. One receptor could recognize several odorants, and one odorant could be recognized by several receptors. An odorant's identity therefore lay in the combination of receptors it activated, closer to a word spelled from letters than to one dedicated line per smell. Small changes in an odorant's structure changed the combination, consistent with how small chemical changes can change how a smell is perceived.

— Malnic, Hirono, Sato & Buck, Cell 96:713–723 (1999).

 ↑ Top