---
title: "The eye must never hold still — fixational movements counteract neural adaptation, and a stabilized retinal image fades to a blank field"
type: "claim"
status: "seedling"
writer_model: "claude-opus-4-8"
source_url: "https://smc.neuralcorrelate.com/files/publications/martinez-conde_pbr06.pdf"
source_author: "Susana Martinez-Conde"
source_date: 2006
source_venue: "Progress in Brain Research, Vol. 154, Ch. 8 (Fixational eye movements in normal and pathological vision)"
source_quote: "the goal of oculomotor fixational mechanisms may not be retinal stabilization, but rather controlled image motion"
source_tier: 1
audit_status: "capture-verified — the batch worker read the Martinez-Conde 2006 PDF directly and preserved the grounding quotes (Tier 1); no web tool in this headless run, so no independent re-check. The un-paywalled PDF is a clean verified-verbatim re-check target. The ~80 ms entoptic figure is Coppola & Purves (1996) as reported inside this Tier-1 review, not re-extracted from that primary."
provenance: "Promotion from 10-inbox/raw/2026-07-11-hop-redundancy-reduction-retina-to-barlow-twins.md, 2026-07-12 (headless)"
origin: "hop-batch"
derived_from: "10-inbox/raw/2026-07-11-hop-redundancy-reduction-retina-to-barlow-twins.md"
date_created: "2026-07-12T00:00:00.000Z"
tags: ["neuroscience","perception","microsaccades","fixational-eye-movements","retinal-adaptation","efficient-coding","redundancy-reduction"]
drafted_in: ["2026-07-13-the-name-is-the-citation","the-name-is-the-citation"]
---


During visual fixation the eye is never truly still: microsaccades, slow drift, and tremor keep the retinal image in constant micro-motion. That motion is not tolerated noise — it is functionally required. Neural adaptation continuously "normalize[s] responses across neurons in the face of unchanging or uniform visual stimulation," so an image held perfectly stationary on the retina adapts away and disappears. Under artificial retinal stabilization, perception "fade[s] to a homogeneous field"; the naturally stabilized entoptic shadows of the eye's own retinal blood vessels vanish especially fast — Coppola and Purves (1996), as reported in this review, found them gone "in as little as 80 ms." Fixational eye movements continually refresh the image and prevent this fading.

The reframe Martinez-Conde draws is the surprising part: the oculomotor system is not fighting to *hold* the image steady. "The goal of oculomotor fixational mechanisms may not be retinal stabilization, but rather controlled image motion." A visual system built to detect change discards what stays constant, so it must manufacture change to keep seeing.

This is the perceptual instance of a principle the vault meets elsewhere as prediction and conflict: no change, no signal. It is the sensory-neuroscience "why" beneath [[claim-barlow-1961-efficient-coding-removes-sensory-redundancy]] — the constant is erased because it is statistically redundant — which in turn resurfaces by name in the 2021 loss function [[claim-barlow-twins-loss-minimizes-embedding-redundancy]]. It rhymes with the vault's contradiction discipline in [[claim-cognitive-conflict-prerequisite-for-restructuring]] (congruent input triggers nothing) and with the prediction-error framing of [[claim-p3-indexes-prediction-error-schema-updating]]. The visual cortex whose wiring encodes this input is the same one rewired by early deprivation in [[claim-monocular-deprivation-permanently-rewires-visual-cortex]].

> [!note] Seek's commentary:
> I came in expecting fixational tremor to be a bug the brain tolerates and found it is the feature the brain depends on. "Controlled image motion, not stabilization" is the whole hop in one phrase: the eye jitters *so that* the redundant constant can be thrown away. That is the retina running efficient coding in hardware, forty years before Zbontar and [[entity-yann-lecun|LeCun]] wrote it as a loss. Held at seedling because I have the review, not the Coppola–Purves primary behind the 80 ms. — Seek