Color afterimages are caused by adaptive reversal in opponent channels

What it is

Staring at a highly saturated color block for ten-odd to several dozen seconds, then looking at a white or gray surface, produces a phantom patch in the complementary color — this is the negative color afterimage. Staring at red produces a green afterimage; staring at yellow produces a blue one. The direction always swings between the two poles of an opponent channel; it never comes out as an arbitrary color.

This isn't the intuitive "some residual glow of the original color left in the eye" — the afterimage color is the opposite of the original stimulus, and that opposite direction lines up exactly with the red-green and blue-yellow opponent axes. That is precisely why it counts as behavioral evidence for opponent-process coding.

Why it happens

Staring at one color for a long time drives the opponent channel that encodes it persistently toward one pole; the neural response handling that pole's signal gradually adapts (weakens, becomes harder to drive), while the response handling the opposite pole, not having been continuously driven, stays relatively fresh.

The moment gaze shifts to a neutral gray or white surface — a stimulus that should return the opponent channel to its zero point, favoring neither pole — the already-fatigued pole's response is now weaker, and the relatively fresh opposite pole momentarily has the upper hand. The channel's net output briefly shifts toward that "unfatigued" direction, and the brain interprets this shift as seeing the opposite color: the complementary afterimage. This shift fades as the adapted state gradually recovers, which is why the afterimage is only temporary, dissolving over several seconds to a dozen or so.

Studying it

The standard paradigm is an afterimage induction and report experiment: participants fixate a saturated color block for a standardized duration (duration itself is an important independent variable — longer typically produces a stronger, longer-lasting afterimage), then immediately switch to a neutral gray or white background and report or match the afterimage's color, intensity, and duration. Independent variables include the adapting color's hue and saturation and the fixation duration; dependent variables are the reported afterimage hue (whether it lands exactly on the opposite pole of the opponent axis), an intensity rating, and duration.

Beyond serving as behavioral evidence for opponent coding, this paradigm is also used to indirectly estimate an individual's opponent-channel adaptation and recovery rates, making it a common tool for studying the dynamic properties of color vision rather than just static color-classification ability.

Methodologically: afterimage strength and clarity depend heavily on steady fixation — small eye movements make the afterimage "drift" or fade prematurely, so lab studies typically control eye movements with a fixation point; afterimages reported under free-viewing conditions will be blurrier and less stable than those measured under strict fixation control.

Where it stops holding

• Afterimage strength and duration rise with fixation duration and stimulus saturation. A color glimpsed only briefly generally produces no noticeable afterimage; only sustained fixation on a highly saturated block for a substantial time reliably induces a clearly visible one.
• The afterimage "follows" gaze to a new fixation point, but is only easy to notice against a relatively uniform background; when the background itself is busy or has complex contrast, the afterimage can be masked and go unnoticed.
• This is a functional byproduct of normal color vision, not a visual impairment or abnormality. Nearly everyone with normal color vision can be made to see an afterimage; it's a direct manifestation of how opponent channels normally operate, not something that needs to be "fixed."
• How the afterimage presents changes for people with color vision deficiencies, since their opponent-channel computation differs; it should not be assumed that every user's induced afterimage direction and strength match exactly.

Applying it

• If an interface contains a large, highly saturated color block that users must fixate for a long time (a loading animation, a persistently displayed accent background, a warning color meant to hold attention), expect users to see the corresponding complementary afterimage on the next screen after looking away; avoid having that next screen happen to use that complementary color as the background for key information or status, which would let the afterimage interfere with the judgment that follows.
• For sequential tasks that require precise color judgment (comparing several color swatches for a match, one after another), inserting a brief neutral-gray transition between judgments reduces how much the previous color's residual shift on the opponent channels contaminates the next one.
• Verification: have testers operate the interface at a realistic pace, and immediately after switching to the next screen ask whether they notice a residual color phantom and whether it affected their judgment of the new content's color — rather than testing a single isolated color block's afterimage strength in the lab alone.

Related

• Same group: A1.05.5 Past the cones, signals are re-coded into red-green, blue-yellow, and black-white opponent channels · A1.05.6 Opponent coding explains why combinations like "reddish green" don't exist · A1.05.8 Interface color discriminability should be designed along opponent channels, not isolated hue
• Nearby: A1.05.1 The summed response of three cone types produces color perception
• Search terms: negative afterimage · color afterimage · opponent adaptation · complementary color