Two spatial-frequency channels can fatigue independently; blur compensation in one cannot substitute for the other

What it is

Staring at a pattern of one particular "coarseness" for a long stretch makes the eye specifically less sensitive to that spatial frequency, while sensitivity to other spatial frequencies stays essentially unaffected — a phenomenon called selective adaptation. It shows that the channels handling different spatial frequencies fatigue independently, rather than the eye going generically "tired" all over.

A counterintuitive consequence follows: if a user has become desensitized in one band because of prolonged exposure to dense fine content (small-size body text, say), making the whole interface sharper or crisper overall will not fix it — the fatigue sits in one specific channel, and compensating a different attribute does nothing for it. Only letting that specific spatial frequency rest lets sensitivity recover.

Why it happens

Visual cortex contains a set of neuron populations each narrowly tuned to a particular spatial frequency band, and these channels are relatively independent of each other physiologically. Sustained exposure to a high-contrast pattern at one spatial frequency drives the neuron population tuned to that band into prolonged high firing, causing a response decrement (adaptation) that only recovers after the stimulus is removed for a while. Because this adaptation happens in a band-specific tuned population rather than in some unified, frequency-blind "overall fatigue pool," the resulting sensitivity loss, once measured, clusters around the exposed band — sensitivity at other frequencies tests essentially normal. That is exactly where the term "selective" comes from.

And precisely because the fatigue is localized to a specific band rather than to overall image sharpness, any compensation that doesn't target that band — sharpening, blurring, or adjusting contrast of the whole scene — never touches the actually-fatigued neuron population, and so cannot substitute for genuine rest or a switch away from that band.

Studying it

The classic paradigm is Blakemore and Campbell's selective adaptation experiment: participants fixate a sinusoidal grating at one spatial frequency (the adapting stimulus) for an extended period (tens of seconds to a few minutes), then have their contrast detection thresholds measured immediately afterward across a range of test spatial frequencies, producing a post-adaptation sensitivity curve relative to a pre-adaptation baseline. The typical result is a localized dip in sensitivity centered on the adapting frequency and its near neighbors, with sensitivity at frequencies far from it essentially unchanged. The independent variables are the adapting stimulus's spatial frequency and exposure duration; the dependent variable is the change in contrast sensitivity at each test frequency relative to baseline.

This paradigm was originally key evidence that the human visual system contains multiple independent spatial-frequency channels rather than a single, generic "sharpness detector" — it belongs to fairly basic visual psychophysics. Applied research specifically targeting interface-use scenarios (e.g., measuring band-specific sensitivity change after prolonged reading of fine text) is comparatively scarce; current understanding is mostly an extrapolation of the basic finding to interface contexts, with limited direct measurement.

Where it stops holding

• The adaptation effect is temporary and self-resolving. Sensitivity returns to baseline after the specific spatial frequency is no longer being viewed; this describes a short-term functional state, not permanent visual damage or a long-term eye-health issue.
• It only holds for sustained, high-contrast exposure to one specific band. Everyday content usually mixes spatial-frequency components rather than presenting a single frequency continuously; there is a gap between the clean single-frequency-grating adaptation paradigm in the lab and real usage scenarios, so effect size should not be extrapolated directly.
• It does not explain generic "tired eyes." General visual fatigue and accommodative muscle fatigue have their own mechanisms and causes; selective adaptation describes a more specific, more localized phenomenon, and the two should not be used to explain each other.
• Individual channel tuning width and adaptation speed vary. The adaptation curve shape measured in a lab is a population-level regularity, not a guarantee that every individual's affected frequency range and recovery speed match it exactly.

Applying it

• If an interface has a large block of persistently high-spatial-frequency dense content (small type, fine grid lines, fine-grained texture backgrounds), don't expect making the rest of the interface sharper or blurrier to relieve the user's reduced sensitivity there — the fatigue sits in that specific band, and adjusting a different attribute doesn't address it.
• When users must stare at content of one fixed coarseness for a long time (long tables, dense code, fine drawings), give the content's spatial-frequency composition some variation — different blocks with different densities — rather than keeping a single fine-grained density throughout, so the same channel isn't continuously driven.
• When users report that a section "looks blurry after staring at it a while," suspect channel-specific adaptation from prolonged fixation on one particular coarseness first; the fix is letting the eyes leave that band for a moment (look elsewhere, at a different density), not re-rendering the block sharper.
• Verification: have test participants fixate the target dense content continuously for a period, then immediately test their ability to discriminate content at that same spatial frequency versus content at a clearly coarser or finer frequency measured in the same session; a pattern of decline only at the original frequency confirms selective adaptation rather than general visual fatigue.

Related

• Same group: A1.04.2 Contrast sensitivity varies with spatial frequency; fine strokes need more contrast · A1.04.4 Texture segregation depends on local spatial-frequency differences, not color or luminance
• Nearby: A1.15 Visual fatigue and accommodative load · A1.31 Mach bands and edge enhancement
• Search terms: selective adaptation · spatial frequency channel · contrast adaptation · Blakemore Campbell