# Eye movements efficiently expose single cone photoreceptors to global scene color statistics

**Authors:** Takuma Morimoto, Luna Wang, Kinjiro Amano, David H. Foster, Sérgio M.C. Nascimento

PMC · DOI: 10.1016/j.isci.2026.114948 · iScience · 2026-02-09

## TL;DR

Eye movements help individual cone cells in the eye quickly adapt to the overall color of a scene.

## Contribution

The study shows that natural eye movements rapidly align single cone photoreceptor inputs with global scene color statistics.

## Key findings

- Local cone excitation histograms converge with global statistics within 10 seconds of viewing.
- Local von Kries scaling from sampled cones closely matches global adaptation.
- Natural gaze patterns outperform random-walk and chromatic-salience models in sampling scene color.

## Abstract

Our visual experience is a dynamic consequence of our actions, most notably our continuously shifting gaze. These shifts directly influence the spectral input received by individual cone photoreceptors. The present study tested how gaze shifts shape the chromatic diet of single cones and their relationship to global adaptation. Eye movements were recorded from observers freely viewing natural scenes outdoors and from observers freely viewing hyperspectral images of the same scenes indoors on a computer-controlled laboratory display. From the hyperspectral data, spatially local histograms of excitations in long-, medium-, and short-wavelength-sensitive cones were accumulated over time. A simulated global illuminant change was then introduced into the images to test how well local retinal adaptation might discount its effects. Despite the highly non-uniform pattern of natural gaze behavior, the results suggested that individual cones may experience the global color statistics of full scenes within the first several seconds of viewing. This effective sampling could support robust adaptation, allowing local adaptation mechanisms to compensate for illumination changes.

•Natural gaze makes single-cone chromatic diets converge to scene-wide statistics•Local vs. global L/M/S cone histograms correlate strongly within ∼10 s of viewing•Local von Kries scaling from cone samples nearly matches global adaptation•Random-walk sampling approximates observers, but chromatic-salience underperforms

Natural gaze makes single-cone chromatic diets converge to scene-wide statistics

Local vs. global L/M/S cone histograms correlate strongly within ∼10 s of viewing

Local von Kries scaling from cone samples nearly matches global adaptation

Random-walk sampling approximates observers, but chromatic-salience underperforms

Biological sciences; Physiology; Biophysics

## Full-text entities

- **Diseases:** cataract (MESH:D002386)
- **Chemicals:** CIECAM16 (-)
- **Species:** Mus musculus (house mouse, species) [taxon 10090], Homo sapiens (human, species) [taxon 9606]

## Full text

_Full body text omitted from this summary view._ Fetch the complete paper as Markdown: https://tomesphere.com/paper/PMC12955238/full.md

## Figures

8 figures with captions in the complete paper: https://tomesphere.com/paper/PMC12955238/full.md

## References

93 references — full list in the complete paper: https://tomesphere.com/paper/PMC12955238/full.md

---
Source: https://tomesphere.com/paper/PMC12955238