Dynamic Remapping for Visual Stability Across Saccades

You are reading version 1 of this direction, frozen 2026-10-04. The latest version is version 2.

Version 1 of 224 papers36 conceptsfrozen 2026-10-04

Every time you move your eyes—which happens about three to four times per second—the image projected onto your retina shifts violently. If you were to move a handheld camera with the same frequency and speed, the resulting footage would be an unwatchable, jittery mess. Yet, our perception of the world remains perfectly stable. Objects do not appear to jump; the room does not appear to shake.

For decades, neuroscientists assumed the brain solved this "shaky camera" problem by converting eye-centered (retinotopic) input into a stable, world-centered (spatiotopic) map. However, recent evidence suggests a more dynamic and counterintuitive reality: our internal maps remain stubbornly eye-centered, and stability is a hard-won computational result of constant, active updating.

The Retinotopic Reality

The most intuitive way to achieve visual stability would be to build a master map of the environment where every object has a fixed coordinate. But research by Justin L. Gardner, Elisha P. Merriam, and J. Anthony Movshon has shown that early visual maps in the human occipital cortex are strictly retinotopic. When your eyes move, the neural activity representing the world moves with them Gardner et al..

This "eye-centeredness" persists even into higher-order regions. Saloni Sharma, Gabriel Kreiman, and Margaret S. Livingstone found that neurons in the ventral visual stream—the "what" pathway responsible for object recognition—reset their activity with every single fixation Sharma et al.. They do not build a cumulative, gaze-independent scene. Instead, they process only what is currently landing on the retina, effectively "forgetting" the previous snapshot to make room for the new one.

Bridging the Snapshots with Gain Fields

If the maps themselves aren't stable, how does the brain know that an object at the left of the retina after a saccade is the same object that was at the right of the retina before the saccade?

The answer lies in spatial gain fields. In a seminal 1991 paper, Alexandre Pouget and Terrence J. Sejnowski demonstrated that the brain can transform coordinates by modulating visual signals with eye position Pouget et al.. Later fMRI work by Merriam and Gardner confirmed that the human visual cortex "labels" retinal images with current gaze direction Merriam & Gardner. By combining the "where" of the retina with the "where" of the eye, the brain can mathematically infer the "where" of the world without needing a physical, world-centered map.

Predictive Remapping and Pre-loading

Stability isn't just about reacting to eye movements; it's about anticipating them. Before your eyes even finish moving to a new target, your brain performs trans-saccadic remapping.

Grace Edwards and Patrick Cavanagh used EEG to show that the brain "pre-loads" information about a target before the eyes land on it Edwards & Cavanagh. This predictive update ensures that when the new retinal input arrives, it matches the brain's expectation, creating a seamless transition. This process is so fundamental that it happens rhythmically; Douglas McLelland and Rufin VanRullen found that the success of this spatial updating is tied to the phase of 7 Hz theta oscillations in the brain McLelland et al..

The Cognitive Blackboard

If early visual areas are just transient buffers, where does the "feeling" of a stable scene come from? Floris P. de Lange proposes that the early visual cortex acts as a multiscale cognitive blackboard de Lange. Higher-level regions use feedback connections to "write" task-relevant information—like the identity of an object—back onto these low-level maps.

This feedback helps maintain "objecthood" across saccades. Research by Paolo Papale suggests that once the brain identifies an object, it remaps that representation across eye movements to reduce the time it takes to re-recognize the object in the next fixation Papale et al..

Suppression and Reset

Finally, the brain manages the "motion blur" of the eye movement itself through saccadic suppression. James M. McFarland and colleagues discovered a biphasic modulation in the primary visual cortex: a sharp suppression of activity during the saccade, followed by a boost in firing immediately after McFarland et al.. This effectively "blanks" the shaky footage and resets the system for a fresh, clean snapshot.

In summary, visual stability is not a static property of our brain's architecture. It is a high-speed, dynamic performance involving predictive remapping, coordinate transformation via gain fields, and rhythmic suppression—all working together to stitch a series of disjointed, eye-centered snapshots into the illusion of a continuous, stable world.

Papers in this version

History & citationVersion 2 · 2026-10-04

Current · Permanent link

  1. Version 2 (latest) · 2026-10-04 · 46 papers
  2. Version 1 · 2026-10-04 · 24 papers
Cite this direction

Latest version (a living page):

@misc{lacuna_perceptual_visual_stability_2e24c494,
  title = {perceptual visual stability},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/perceptual-visual-stability/rdr_2e24c4942386442a90b0aae0f6e2e359},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-22},
}

This exact version (version 1, frozen text):

@misc{lacuna_dynamic_remapping_for_visual_stability_a_2e24c494_v1,
  title = {Dynamic Remapping for Visual Stability Across Saccades},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/dynamic-remapping-for-visual-stability-across-saccades/rdr_2e24c4942386442a90b0aae0f6e2e359/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-22},
}