Rod-Cone Interaction

Abstract

Rod-cone interaction, which MeSH classifies under mesopic vision, is the mutual influence between the two classes of photoreceptor that classical duplex theory treats as working in separate light ranges. At the twilight levels where both rods and cones are active, their signals combine and interfere: rod activity can suppress cone-mediated flicker, cancel it when the two are out of phase, and shift perceived hue in proportion to rod contrast. These interactions are built into the retina, where rods feed the cone pathway through gap junctions and shared bipolar and amacrine circuits. This article sets out the anatomical routes that couple the two systems, the psychophysical signatures of their interaction, the melanopsin contribution now folded into the account, and the silent-substitution methods used to isolate each receptor class.

Keywords: rod-cone interaction, mesopic vision, flicker cancellation, silent substitution

At dusk the world does not switch cleanly from the daytime vision of the cones to the night vision of the rods; for a wide band of light levels both receptor systems run at once, and what the eye reports is neither a pure cone image nor a pure rod image but a blend in which each system reaches into the other. Classical textbook vision divides the retina into two tidy regimes—rods for the dark, cones for the light—but the division is a simplification that breaks down precisely where most real seeing happens, in the long twilight of the mesopic range. Rod-cone interaction is the study of that overlap: how signals from the two receptor classes meet in the retina, where they reinforce, suppress, and distort one another before a single message is sent to the brain (Buck, 2003; Zele & Cao, 2015).

What Rod-Cone Interaction Is

Rod-cone interaction is the set of ways in which signals originating in the rod photoreceptors and signals originating in the cone photoreceptors influence one another on their path through the retina and into perception. The starting point is the duplex theory of vision, which holds that the retina contains two largely independent systems: a rod system of high sensitivity and poor acuity that serves vision in dim light, and a cone system of lower sensitivity, high acuity, and color that serves vision in bright light. The theory is a good first approximation, and for the extremes of lighting—starlight on one end, daylight on the other—it is nearly exact. What it misses is the broad middle, the mesopic range, where both systems are driven at once and no longer behave as if the other were absent (Buck, 2014).

In that middle range the two systems are not merely active in parallel; they interact. Rod signals alter the gain, timing, and apparent color of cone signals, and the reverse influences occur as well. These effects are not incidental noise but systematic, measurable consequences of the fact that rods and cones share much of the retinal circuitry downstream of the receptors. The sections that follow build the subject from the wiring upward: first the retinal circuits that physically couple the two systems, then the psychophysical signatures of their interaction, the twilight range in which it matters most, its effect on color, the newer melanopsin contribution, and finally how each receptor class is isolated and measured.

The Retinal Circuits That Couple Rods and Cones

Rod and cone signals have somewhere to meet because the retina does not keep them apart. Rods greatly outnumber cones and lack a private line to the brain; instead their signals are routed into pathways that cones also use, and three such routes are now recognized (Daw, Jensen, & Brunken, 1990). The primary rod pathway sends the rod signal to a dedicated rod bipolar cell, which does not contact a ganglion cell directly but passes its signal to the AII amacrine cell; the AII cell then injects the rod signal into the cone bipolar pathways, so that the rod message reaches the ganglion cells over cone wiring. The secondary pathway is more direct still: rods and cones are joined by electrical gap junctions, so that a rod signal can flow straight into a neighboring cone and travel outward along the cone pathway. A tertiary route connects rods to certain cone bipolar cells directly.

The electrical coupling at the heart of these routes is made of gap junctions—channels built largely from the connexin36 protein—and Bloomfield and Völgyi showed how pervasively such junctions knit the retina together, coupling rods to cones and linking the AII amacrine cells into a network whose conductance is itself under adaptive control (Bloomfield & Völgyi, 2009). Because that coupling can be turned up or down with the ambient light level, the degree of rod-cone interaction is not fixed but changes as the eye adapts, strongest in the mesopic range and suppressed toward the photopic extreme. Sharpe and Stockman traced how the rod signal, routed through these shared pathways, shapes vision at low light levels, arguing that understanding the rod contribution means understanding the pathways it borrows rather than the receptor alone (Sharpe & Stockman, 1999).

The Psychophysics of Interaction

The wiring predicts that rod and cone signals should influence one another perceptually, and they do, in ways that can be measured without opening the eye. The clearest early demonstration concerned flicker. Goldberg, Frumkes, and Nygaard found that steady or slowly changing rod stimulation suppresses the visibility of cone-mediated flicker: a background that excites the rods makes a flickering cone stimulus harder to see, a phenomenon named suppressive rod-cone interaction (Goldberg, Frumkes, & Nygaard, 1983). The rods, in other words, can turn down the gain of the cone flicker channel, an inhibitory influence that reveals the two systems are anything but independent.

A second signature concerns timing. Rod signals are slower than cone signals—they take longer to pass through the retina—so when a light flickers fast enough to drive both systems, the rod and cone contributions arrive at different phases. MacLeod showed that at certain flicker frequencies the delayed rod signal arrives almost exactly out of phase with the cone signal, so that the two cancel and the perceived flicker is nulled: rods, as the title of the report put it, cancel cones in flicker (MacLeod, 1972). The first demonstration lets the reader set a rod-to-cone delay and a flicker frequency and watch the combined signal grow and vanish as the two components drift in and out of phase.

Table 1 sets out the principal forms the interaction takes, the direction of influence, and the perceptual signature by which each is recognized.

The Mesopic Range

Rod-cone interaction matters because it fills the range of light levels in which most natural dim-light vision occurs. Between the scotopic range, dim enough that only rods respond, and the photopic range, bright enough that cones dominate and rods are saturated, lies the mesopic vision range, roughly the light of dusk, moonlight, and dim interiors, where both systems contribute (Stockman & Sharpe, 2006). Across this range the balance shifts continuously: as light falls, the cone contribution wanes and the rod contribution grows, and the eye's spectral sensitivity slides with it. The peak of sensitivity moves from about 555 nanometers in cone-dominated vision toward about 507 nanometers in rod-dominated vision—the Purkinje shift—which is why reds darken and blues brighten as evening falls.

Figure 1 shows the two reference curves whose peaks define that slide. The scotopic luminous efficiency function, carried by the rods, peaks near 507 nanometers; the photopic function, carried by the cones, peaks near 555 nanometers. As light falls and vision passes from cone-dominated to rod-dominated, the eye's effective sensitivity migrates from the right-hand curve toward the left-hand one, and the roughly 48-nanometer gap between their peaks is the Purkinje shift.

Figure 1

The Purkinje Shift: Scotopic and Photopic Luminous Efficiency

This continuous reweighting is exactly why a single luminous efficiency function cannot describe the mesopic eye: neither the photopic curve nor the scotopic curve is correct there, and the effective sensitivity is a level-dependent mixture of the two (Zele & Cao, 2015). The second demonstration lets the reader move a luminance level across the scotopic, mesopic, and photopic ranges and watch the rod and cone weights trade places while the spectral sensitivity peak slides between the two canonical values.

Rods and Color

Because rod signals enter the cone pathways, they can reach the neural channels that compute color, and so the rods—though they carry no color information of their own and there is only one kind of rod—can nonetheless alter the hue a cone stimulus appears to have. Cao, Pokorny, Smith, and Zele measured this rod intrusion directly and found a clean regularity: the shift that rod activity produces in perceived color grows linearly with rod contrast, so that a given increment of rod signal produces a fixed increment of hue shift (Cao, Pokorny, Smith, & Zele, 2008). The linearity matters because it means the rod contribution to color can be treated as an additional input to the color channels with a constant weight, rather than as an unpredictable distortion.

That a system with a single photopigment can influence color at all is one of the more counterintuitive consequences of rod-cone interaction: color requires a comparison across receptor types, and the rods gain entry to that comparison only by sharing circuitry with the cones (Buck, 2003). The third demonstration lets the reader raise and lower rod contrast and read off the resulting hue shift, tracing the straight line that Cao and colleagues measured.

A Third Photoreceptor

The classical account has two receptor classes, but the retina has a third photoreceptive element: a small population of retinal ganglion cells that contain the pigment melanopsin and respond to light on their own, independently of the rods and cones that also feed them. These intrinsically photosensitive ganglion cells were first known for driving the pupil and the circadian clock, but they also contribute to conscious vision, and their signal interacts with the rod and cone signals much as those two interact with each other. Zele, Adhikari, Feigl, and Cao showed that melanopsin activation adds to the perceived brightness of a stimulus beyond what the cones alone would predict (Zele, Adhikari, Feigl, & Cao, 2018), and Cao, Chang, and Gai found that melanopsin activation even shifts the stimulus that observers judge to be pure white (Cao, Chang, & Gai, 2018).

The melanopsin contribution is clearest in the pupil, where Barrionuevo and Cao showed that luminance and chromatic signals interact with melanopsin activation in distinct ways to control the pupillary light response (Barrionuevo & Cao, 2016). The upshot is that rod-cone interaction is now understood as one part of a larger problem of photoreceptor interaction, in which three input systems with different spectral sensitivities and time courses combine to produce a single percept of brightness and color.

Measuring the Interaction

Studying how receptor classes interact requires a way to drive one class while holding the others fixed, and the technique that makes this possible is silent substitution. The method exploits the fact that each receptor class has its own spectral sensitivity: by exchanging one light for another chosen so that the two produce identical excitation in every receptor class except the target, a stimulus can modulate, say, the rods alone while the cones and melanopsin see no change at all. The target receptor is driven; the others are silenced. Maguire and colleagues used silent substitution to isolate a rod electroretinogram from the light-adapted human eye—recording a rod-driven electrical response under background conditions that would normally be assumed to saturate the rods—demonstrating that rod signals persist and can be measured well into the range where cones are active (Maguire et al., 2016).

Silent substitution is what turned rod-cone interaction from a set of qualitative demonstrations into a quantitative science, because it allows the contribution of each receptor class to be measured in isolation and then recombined to test how the classes sum, suppress, or cancel. It is the methodological thread that runs through the modern work on mesopic vision, rod color intrusion, and the melanopsin contribution alike.

Worked Example

The flicker cancellation that MacLeod described is the cleanest place to work the numbers, and the first demonstration reproduces this arithmetic. Treat the rod and cone contributions to a flickering light as two sinusoids of the same frequency and, for simplicity, the same amplitude, differing only in phase because the rod signal is delayed. If the rod signal lags the cone signal by a time delay τ, then at flicker frequency f the phase lag in degrees is φ equals 360 times f times τ. Two equal sinusoids differing in phase by φ sum to a single sinusoid whose amplitude is 2A times the absolute value of the cosine of φ over 2. When φ is zero the two reinforce and the amplitude is 2A; when φ is 180 degrees they are exactly opposed and the amplitude is zero—complete cancellation, a nulled flicker.

Put numbers to it. A rod delay of about 33 milliseconds relative to the cones is typical in the mesopic range. The phase lag reaches 180 degrees—the cancellation point—when 360 times f times 0.033 equals 180, that is when f equals 180 divided by (360 times 0.033), which is about 15 hertz. So a light flickering near 15 hertz should drive the rod and cone signals into opposition and null the perceived flicker, while at half that frequency, about 7.5 hertz, the phase lag is only 90 degrees and the combined amplitude is 2A times the cosine of 45 degrees, or about 1.41 times A—appreciable flicker. The single delay thus predicts both a frequency of maximum cancellation and the gradual return of visible flicker on either side of it.

The same logic of weighted combination runs through the mesopic range more generally. There the perceived brightness is neither the pure rod nor the pure cone estimate but a weighted sum of the two, with the rod weight rising and the cone weight falling as the light dims; the spectral sensitivity peak slides from 555 toward 507 nanometers—a Purkinje shift of 48 nanometers—as that weighting changes. Rod-cone interaction, in both the flicker case and the brightness case, is at bottom the arithmetic of combining two signals whose relative weight and relative phase depend on the light level.

Discussion

The study of rod-cone interaction is, in a sense, the correction of a useful oversimplification. Duplex theory cleanly separated the retina into a scotopic rod system and a photopic cone system, and that separation remains the right way to teach vision and the right description of the two extremes of lighting. But the separation was never anatomical: rods were always wired into cone pathways, and the moment both systems are driven at once their signals combine (Sharpe & Stockman, 1999; Buck, 2014). The psychophysical signatures—suppression, cancellation, hue shift—are the behavioral fingerprints of that shared wiring, and the anatomy of gap junctions and the AII amacrine network is its physical basis (Bloomfield & Völgyi, 2009).

What makes the subject more than a footnote to duplex theory is that the mesopic range it governs is where a great deal of consequential human vision actually happens—driving at dusk, moving through dim interiors, reading instruments at night. In those conditions the eye is neither the daytime nor the nighttime instrument of the textbook but a blend whose spectral sensitivity, temporal response, and even color appearance depend on the exact light level and on how the two receptor systems are interacting at that moment. The practical problem of specifying a mesopic luminous efficiency function—a single curve that describes the twilight eye—is unsolved precisely because the eye's sensitivity there is not a fixed function but a moving weighted average of two (Stockman & Sharpe, 2006).

Current Directions

The most active recent front has been the folding of melanopsin into what was a two-receptor problem. The intrinsically photosensitive ganglion cells add a third input with its own slow time course and distinct spectral sensitivity, and recent work has measured its contribution to conscious brightness, to the color judged to be white, and to the pupil, using the same silent-substitution logic developed for the rods and cones (Zele, Adhikari, Feigl, & Cao, 2018; Cao, Chang, & Gai, 2018; Barrionuevo & Cao, 2016). The classical two-receptor interaction is increasingly treated as a special case of a three-way photoreceptor interaction, and disentangling the three contributions under natural, changing illumination is an open problem.

A second front is methodological and applied. Silent-substitution recordings such as the light-adapted rod electroretinogram are maturing into clinical tools, because a signal that isolates one receptor class can reveal receptor-specific disease that a mixed response would hide (Maguire et al., 2016). On the applied side, the fact that mesopic sensitivity is a level-dependent mixture of rod and cone contributions bears directly on lighting design and on standards for night driving, where specifying how bright a dim scene appears requires a model of exactly the interaction this field studies (Zele & Cao, 2015). The open questions run from the molecular control of gap-junction coupling to the engineering of light sources matched to the twilight eye.

Common Misconceptions

Rods and cones work in completely separate light ranges.

Only at the extremes. Across the whole mesopic range—dusk, moonlight, dim interiors—both systems are active at once and their signals combine and interfere (Zele & Cao, 2015).

Rods cannot affect color because they are colorblind.

A single rod type carries no color of its own, yet because rod signals enter the shared cone pathways they reach the color channels and shift perceived hue, by an amount that grows linearly with rod contrast (Cao, Pokorny, Smith, & Zele, 2008).

Adding rod light can only make a cone stimulus easier to see.

Rod stimulation can suppress cone signals: a rod-exciting background lowers the visibility of cone-mediated flicker, an inhibitory interaction, not a simple addition (Goldberg, Frumkes, & Nygaard, 1983).

There are only two kinds of photoreceptor.

A third photoreceptive element, the melanopsin-containing retinal ganglion cell, responds to light in its own right and contributes to brightness, color, and the pupil alongside the rods and cones (Zele, Adhikari, Feigl, & Cao, 2018).

Glossary

AII amacrine cell.

A retinal interneuron that receives the rod bipolar signal in the primary rod pathway and injects it into the cone bipolar circuits, routing rod signals onto cone wiring.

Cone.

A photoreceptor of lower sensitivity and higher acuity that mediates color vision and operates in bright, photopic light; one of three cone types underlies trichromatic color.

Connexin36.

The protein that forms the neural gap junctions coupling rods to cones and linking AII amacrine cells, the molecular basis of electrical rod-cone coupling.

Duplex theory.

The classical account dividing the retina into an independent rod system for dim light and cone system for bright light; a good approximation at the extremes but not in the mesopic middle.

Flicker cancellation.

The nulling of perceived flicker when the delayed rod signal arrives out of phase with the cone signal, so that the two sinusoidal contributions sum to zero.

Gap junction.

A channel that directly connects the interiors of two adjacent cells, allowing electrical signals to pass between them; in the retina it couples rods to cones.

Luminous efficiency function.

A curve describing the eye's relative sensitivity to each wavelength; it differs for photopic and scotopic vision and has no single fixed form in the mesopic range.

Melanopsin.

The photopigment of intrinsically photosensitive retinal ganglion cells, a third photoreceptive system that contributes to brightness, color, the pupil, and the circadian clock.

Mesopic vision.

Vision at twilight light levels, between the scotopic and photopic ranges, where both rods and cones are active and rod-cone interaction is strongest.

Photopic vision.

Vision in bright light, mediated by the cones, with high acuity and color; rods are saturated and contribute little.

Primary rod pathway.

The route by which a rod signal passes to a rod bipolar cell, then to the AII amacrine cell, and from there into the cone bipolar pathways toward the ganglion cells.

Purkinje shift.

The movement of peak spectral sensitivity from about 555 nanometers toward about 507 nanometers as vision shifts from cone-dominated to rod-dominated, darkening reds and brightening blues at dusk.

Rod.

A photoreceptor of high sensitivity and low acuity that mediates vision in dim, scotopic light; there is a single rod type, so rods alone carry no color.

Scotopic vision.

Vision in very dim light, mediated by the rods alone, with high sensitivity but no color and poor spatial detail.

Silent substitution.

A stimulus method that exchanges one light for another to modulate a single receptor class while holding the excitation of all others constant, isolating that class's contribution.

Suppressive rod-cone interaction.

The reduction of cone-mediated flicker visibility by rod stimulation, an inhibitory influence of the rod system on the cone system.

Key Researchers

Stewart A. Bloomfield

. SUNY College of Optometry; a retinal neuroscientist of the gap-junctional circuitry that couples rod and cone pathways, establishing how electrical coupling through the AII amacrine cell and the photoreceptor network routes rod signals into cone channels. Faculty Page

Steven L. Buck

. University of Washington (Professor Emeritus); mapped how rod signals intrude on and alter cone-mediated hue and luminance perception, and authored the standard reference reviews of rod-cone interaction in human vision. Faculty Page - Google Scholar

Dingcai Cao

. University of Illinois at Chicago; leads quantitative studies of rod contributions to color and of rod-cone-melanopsin interactions under mesopic and scotopic illumination, establishing that rod intrusion on color is linear with rod contrast. Faculty Page - Google Scholar

Andrew Stockman

. UCL Institute of Ophthalmology, University College London; co-authored the Stockman-Sharpe cone fundamentals adopted as a CIE standard and the foundational analyses of rod pathways, mesopic vision, and luminous efficiency that frame how rod and cone signals combine. ORCID - Faculty Page - Google Scholar

Andrew J. Zele

. Queensland University of Technology; maps vision across the mesopic and scotopic ranges and the interactions of rods, cones, and melanopsin that govern brightness, color, and the pupil, bringing silent-substitution and photoreceptor-isolating methods to the problem. Faculty Page - Wikidata

Frequently Asked Questions

What is rod-cone interaction?

It is the set of ways in which signals from the rod and cone photoreceptors influence one another in the retina and in perception, rather than acting as the two independent systems that classical duplex theory describes. The interactions are strongest in the mesopic range, where both systems are active at once (Buck, 2003).

Why do rods and cones interact at all if they serve different light levels?

Because they share retinal wiring. Rods lack a private line to the brain, so their signals are routed into the cone pathways through gap junctions and the AII amacrine cell, and the moment both systems are driven their signals meet and combine (Daw, Jensen, & Brunken, 1990).

What is the mesopic range?

It is the band of twilight light levels—dusk, moonlight, dim interiors—between the rod-only scotopic range and the cone-dominated photopic range, where both receptor systems contribute and the eye's sensitivity is a level-dependent mixture of the two (Stockman & Sharpe, 2006).

How can rods affect perceived color if they are colorblind?

A single rod type carries no color of its own, but because rod signals enter the shared cone pathways they reach the neural channels that compute color and shift perceived hue, by an amount that grows linearly with rod contrast (Cao, Pokorny, Smith, & Zele, 2008).

What is flicker cancellation?

The rod signal is slower than the cone signal, so at certain flicker frequencies the two arrive out of phase and cancel, nulling the perceived flicker—the effect MacLeod summarized as rods cancelling cones in flicker (MacLeod, 1972).

What is suppressive rod-cone interaction?

It is an inhibitory effect in which rod stimulation reduces the visibility of cone-mediated flicker: a background that excites the rods turns down the gain of the cone flicker channel (Goldberg, Frumkes, & Nygaard, 1983).

How is the contribution of each receptor class measured?

By silent substitution, which exchanges one light for another so as to modulate a single receptor class while holding the excitation of all others constant; it has been used to record even a rod response from the light-adapted eye (Maguire et al., 2016).

Is there a third photoreceptor involved?

Yes. Melanopsin-containing retinal ganglion cells respond to light on their own and contribute to brightness, to the stimulus judged to be white, and to the pupil, so rod-cone interaction is increasingly treated as part of a three-way photoreceptor interaction (Zele, Adhikari, Feigl, & Cao, 2018; Cao, Chang, & Gai, 2018).

References