Registering a scene briefly does not mean any part of it survives without attention
Aliases: sensory memory decay · attentional gating · attentional selection
What it is
A whole scene being briefly and fully registered in sensory memory does not mean every part of that registration gets a chance to be used. What actually determines which content gets "pulled out" of this raw record and passed on to further processing is whether attention lands on that content before sensory memory fades. Content that attention never selects disappears along with sensory memory's own natural decay — it isn't copied anywhere more lasting, and it can't be recovered by "trying to recall what was just seen," because the raw record itself no longer exists.
Why it happens
There is an attention-formed gate between sensory memory and the next stage of memory: within the extremely brief window before sensory memory fades, attention transfers or consolidates whatever it selects into a more stable processing stage, while whatever it doesn't select stays behind and decays along sensory memory's own natural trajectory. This isn't "understand everything first, then decide what to keep" — selection happens before understanding is even complete. Unattended content may vanish from the raw record before it's ever identified at all. This is exactly why the experience of "something was definitely on screen, but I have no memory of it whatsoever" happens: it's not a memory malfunction — that content simply was never selected by the gate in the first place.
Studying it
Verifying this typically relies on manipulating which parts of a display can obtain attention — using a visual cue to direct a participant's attention to a region, or having participants perform a concurrent attention-demanding task — and comparing whether the attended/cued region versus the unattended region can be correctly reported or recognized once sensory memory has faded. If the unattended region cannot be reported even with ample display duration, that shows the loss is due to not being selected by attention, not insufficient display time.
Common independent variables: whether attention is directed to the target region, and whether attentional resources are occupied by a concurrent task. Common dependent variables: report accuracy and recognition accuracy for the target region's content.
This method is often used in interface research to verify whether a visual element actually entered a user's processing scope — the mere fact that something "appeared on screen" doesn't prove the user acquired that information; whether the region received attention needs separate verification.
Methodological caveat: judging "whether content was attended to" itself relies on indirect measures (after-the-fact report, recognition tests), which measure the output of memory rather than directly observing attentional allocation. Keep this inferential chain in mind when interpreting results.
Where it stops holding
- This entry covers the mechanism itself — that unattended content decays along with sensory memory. Exactly how much and how complex a piece of content gets missed this way depends on the overall information density of the display and the capacity ceiling of sensory memory itself.
- Content vanishes completely only when it receives no attentional resources at all. If attention is partially allocated (the content sits at the edge of the attended region, say), a residual trace may be enough to support some low-level judgment (sensing "something changed" without being able to say what), so this is not strictly binary.
- This entry does not cover how attention itself gets drawn to or allocated toward a given region — that's a question of attentional mechanisms; here the focus is only on what happens to sensory memory in attention's absence.
Applying it
- Don't assume that a user "looked at" a region means they acquired the information in it; if the target information appears at a location the user's attention wasn't directed toward at the time, they may well have no memory of it afterward, even though it was fully present on screen.
- For critical information that users need to notice, actively direct attention to its location (visual cues, motion prompts) rather than relying on the information "just being there" to be automatically captured.
- Verification: at the moment the target information appears, use eye tracking or a simple after-the-fact recognition test to confirm whether the user's attention actually reached that region. If attention never got there but the design assumed "the user should have seen it," the presentation needs adjusting to add clearer attentional guidance.