Local adaptation completes faster than overall light/dark adaptation

What it is

The dark- and light-adaptation time constants covered elsewhere describe how long the retina as a whole takes to switch to an entirely new average illumination level — anywhere from seconds to tens of minutes. But the visual system has another, much faster mechanism: different regions within the visual field can independently adjust their own local gain to match that region's own local illumination level, on a much shorter timescale (typically hundreds of milliseconds to a second or two). This is local adaptation.

It's easy to conflate this with overall light/dark adaptation, but the two answer different questions. Overall adaptation answers "how long does the retina take to switch to a new average illumination level?" Local adaptation answers "at the same instant, can a brighter patch and a darker patch within the same field of view both be seen clearly at once?" The latter doesn't require waiting — in most cases it happens near-instantaneously, and it's the reason the visual system can handle high-dynamic-range scenes simultaneously.

Why it happens

Gain control in the retina doesn't happen only at one global level. The retina and early visual pathway contain neural circuits that can independently regulate response gain over a smaller spatial extent: when a local region receives a rise in illumination, the neurons corresponding to that region relatively quickly turn down their own response gain, preventing saturation from excess photons; a neighboring region at a different illumination level maintains its own independent gain setting, unaffected by the first. This local rather than global gain regulation lets the retina handle sharply different illumination levels within the same visual field at once, without first bringing the entire field to a uniform state.

This mechanism runs much faster than overall adaptation precisely because it adjusts the gain state of relatively local, relatively small neural circuits, rather than depending on the slower, retina-wide change in photopigment concentration that underlies light/dark adaptation's biochemistry. Local adaptation is closer to a rapid neural gain reset, while overall adaptation depends on a slower photochemical regeneration process — that difference of several orders of magnitude in time constant is exactly why.

Studying it

The standard paradigm is a simultaneous-contrast / local-adaptation psychophysics experiment: participants view a single field of view containing multiple regions at different background illumination levels, with a target stimulus to be detected or matched embedded in each region; the measure is whether detection thresholds or matching results vary systematically with that region's local background illumination rather than with the average illumination across the whole field. The typical result is that detection threshold depends mainly on the illumination level of the target's local region, demonstrating a gain-regulation mechanism independent of global adaptation state. The independent variables are local background illumination level (which can differ across regions within a single presentation) and target stimulus properties; the dependent variable is local detection threshold or matched brightness.

This behavioral phenomenon corroborates retinal physiology (direct recordings of gain properties in retinal ganglion cell receptive fields) — the behavioral and physiological evidence were obtained independently and support each other.

Where it stops holding

• The dynamic range local adaptation can handle is limited. If a local region itself contains extremely disparate brightness levels internally (a small window containing both direct light and deep shadow at once), local gain regulation cannot make every detail within that small area visible simultaneously either — local adaptation solves "different regions adapt independently," not "unlimited dynamic-range compression."
• Local adaptation is mainly a relatively fast retinal-level adjustment; it cannot substitute for genuine overall dark or light adaptation. If a user needs to reach full dark-vision sensitivity (stargazing, say), local adaptation won't speed that process up — the two mechanisms address different problems.
• A transition effect appears at region boundaries. Local adaptation isn't a fully independent, island-like regulation — the gain settings of neighboring regions influence each other to some degree, and perception near a boundary can sit in a state intermediate between the two regions.

Applying it

• When designing a single screen containing regions of clearly different brightness (a dark content panel next to a bright media preview, say), it's reasonable to expect the user's visual system to see content on both sides nearly simultaneously, without needing extra buffer time for "switching adaptation" — this is a different design consideration from switching to another screen with a completely different average brightness, which does need overall adaptation time.
• Don't rely on local adaptation to paper over a single small region that itself has too wide a contrast range — if one card contains both extreme highlights and extreme shadows internally, local adaptation can't help with that; the region's own dynamic range still needs to be compressed.
• Verification: have testers look back and forth between differently lit regions on the same screen and confirm they can see each region's content clearly without a noticeable wait; if some region still has high-contrast details that are hard to make out simultaneously, the problem is that region's own dynamic range, not the speed of local adaptation.

Related

• Same group: A1.07.1 Dark adaptation takes far longer than light adaptation · A1.07.2 Sudden brightness increases cause brief disability · A1.07.3 Night-mode peak luminance needs to be set independently
• Nearby: A1.02 The division of labor between the fovea and peripheral vision
• Search terms: local adaptation · gain control · retinal local gain · simultaneous contrast