Attention can be allocated to a region of space or to a whole perceptual object

What it is

Attention doesn't only carve out a patch of the screen to process more heavily. It has a second, distinct way of allocating itself: by perceptual object — whatever the visual system has grouped into "one thing" — rather than by coordinate location. The first is space-based attention; the second is object-based attention. The two get conflated because most of the time an object happens to occupy one contiguous patch of space, so they look identical in effect. Only when an object's boundary and a spatial region's boundary can be pulled apart experimentally does it become clear which one attention is actually tracking.

Why it happens

Space-based attention behaves roughly like a movable, zoomable spotlight: it boosts everything falling inside a given region, regardless of how many objects sit there or whether those objects have anything to do with each other. Object-based attention works differently. The visual system first segments the scene into distinct wholes — using contour, connectedness, common motion, and similar cues to do the grouping — and only then does attention land on one of those already-formed objects. Once attention locks onto an object, every property belonging to that object benefits simultaneously, even if those properties sit in different parts of the object and would otherwise fall into regions that space-based attention could tell apart. Conversely, two objects sitting right next to each other and sharing the same patch of space don't get attention automatically averaged or spilled between them just because they're spatially adjacent.

Studying it

The classic paradigm for telling the two apart places two objects overlapping or immediately adjacent, and varies "where the target property sits" and "which object the target property belongs to" independently, comparing judgment speed for two properties on the same object against two properties at an equal spatial distance but on different objects. If the same-object condition is reliably faster, that shows attention is tracking object boundaries, not just spatial distance.

Common independent variables: spatial distance between target properties, and whether the properties belong to the same object. Common dependent variables: response time and judgment accuracy.

This paradigm is often used in interface research to judge how visual elements are actually being grouped — by spatial proximity alone, or by shared boundary and connected contour — and which grouping actually gets users to process several pieces of information as one unit.

Methodological caveat: object identity in the lab is usually manufactured with the boundaries of simple geometric shapes; "objects" in real interfaces are rarely that clean (translucency, occlusion, dynamically generated cards). Before applying lab findings to complex interface elements, first confirm whether the target element would actually be treated as one object by the perceptual system.

Where it stops holding

• The object-based effect depends on the perceptual system being able to cleanly identify the object's boundary; when boundaries are ambiguous or visual separation between elements is weak, attention is more likely to fall back to space-based allocation.
• The two modes are not mutually exclusive — both can operate at once, with space-based attention roughly circling a region first and object boundaries then determining how information within that region gets further organized.
• This entry only establishes that two allocation modes exist. How within-object properties get jointly boosted, and how object boundaries steer attention as an object moves, are separate claims.

Applying it

• To decide whether a set of information should be built as separate elements occupying different spatial positions, or merged inside one recognizable object boundary (the same card, the same container), ask whether that set needs to be read as a whole by the user. Information requiring joint processing belongs inside one object boundary — that is more reliable than relying on spatial proximity alone.
• Don't expect shrinking the gap between elements alone to make users treat them as a group. Without a clear visual object boundary (an outline, a background block, a connected container), mere spatial closeness may not trigger the holistic-processing benefit that object-based attention provides.
• Verification: build a simple judgment task comparing reading or comparison speed for "two pieces of information inside the same container" versus "the same spatial distance but split across two containers." The speed gap directly tests whether the current visual grouping actually forms a perceptual object.

Related

• Same group: A5.15.2 Two properties within the same object both benefit from attention, but properties across objects do not · A5.15.3 The spread of spatial attention can narrow or widen like a spotlight · A5.15.4 Once an object boundary forms, it automatically steers attention to follow the whole object rather than staying fixed in place
• Adjacent: A5.05 Inattentional blindness · A5.07 Attentional capture
• Search terms: space-based attention · object-based attention · perceptual grouping · attentional selection