Ghostly afterimages are the result of our brain stabilizing our vision, according to German researchers, following their detailed investigation into how our mental hardware smooths our darting glances into a steady image.
Multiple times each second, our eyes dart around in rapid movements known as saccades, which our brains must smooth out to provide us with a stable visual image. A part of this stabilization involves afterimages as the brain attempts to interpret and predict its surroundings.
In a recent paper published in Science Advances, a team of researchers investigated just how accurate those predictions are.
The Remarkable Brain
According to the team’s findings, our brains are incredibly accurate in their predictions, but some systemic errors persist.
Since Aristotle, these ghostly afterimages produced by the brain have been documented and pondered by philosophers and scientists. Their tendency to drift across our field of view as our eyes move sometimes betrays a discrepancy between the image’s accuracy and its placement.
The team wanted to isolate the afterimages uniquely, placing their experimental subjects in complete darkness and removing any markers in the surroundings that the brain could use to estimate eye movement.
In each round, participants first viewed a bright flash, followed by a second brief illumination elsewhere in the room, drawing their attention to a different spot. Once participants clearly saw an afterimage, the second light was illuminated in one of several positions, drawing their attention. The researchers then asked the participants to report whether they saw the afterimage to the left, right, or directly in line with the second light source.
This data allowed the team to determine where participants saw the afterimages, while eye-tracking measurements allowed them to determine where the retinas were truly pointed. By comparing this data, the team determined how closely the brain’s perception of where the eye was pointed actually matched the tracking measurement.
Brain Image Accuracy
Generally, the brain proved quite adept at placing the afterimage, yet as the distance of the eye movement increased, so did the placement discrepancy.
“On average, the perceived shift of the afterimage reached about 94 percent of the actual eye movement,” said lead author Richard Schweitzer. “In practical terms, perception follows eye movements very closely, but not perfectly.”
Notably, hypometria, the tendency to undershoot the correct location, was common across individuals and regardless of eye movement direction and distance. This indicates that the issue is systemic and not random or individual. Most people don’t notice the minor difference, and only an investigation into how the brain updates its understanding of space following an eye movement can account for it.
“If one individual has strong eye muscles, then only a small motor output is needed to achieve a certain eye movement,” Schweitzer told The Debrief. “The afterimage moves based on the size of motor output (and in total darkness, this is the only source of information the system has access to), so for small motor output, the afterimage’s movement should be shorter.“
“However, due to the strong eye muscles in our example, the eye moves further,” he added. “Therefore, when expressed as a fraction of the size of the eye movement, the afterimage gain is lower.”
How the Brain Predicts
The team hypothesized several ways the brain may be positioning the afterimage. The team considered that visual feedback after each eye movement might be driving the placement. To test this, they kept the saccade target in place after the eye reached its target position in some rounds of testing, whereas in others, they shifted it slightly to interrupt the feedback.
In both variations, participants perceived the afterimage in the same place. From this, the team suggests that an efferent copy—a copy of the command the brain sends to the eye muscles—is used to predict how the scene should shift.
Those signals tell the brain how far the eyes have moved, allowing it to adjust in advance rather than waiting for feedback. This demonstrates that there is a mismatch between the efferent copy and the eye’s true movement.
“Afterimages become a useful tool for studying how the brain keeps the visual world stable by predicting the sensory consequences of its own movements,” Schweitzer concludes.
“Understanding these predictive mechanisms may provide insights beyond basic vision science, for example, in robotics, virtual reality, and clinical studies of eye-movement disorders, where linking movement with sensory consequences reliably is essential,” Schweitzer said.
The paper, “High-fidelity but Hypometric Spatial Localization of Afterimages Across Saccades,” appeared in Science Advances on March 13, 2026.
Ryan Whalen covers science and technology for The Debrief. He holds an MA in History and a Master of Library and Information Science with a certificate in Data Science. He can be contacted at ryan@thedebrief.org, and follow him on Twitter @mdntwvlf.