Interrupting the current task drains prospective-memory resources, so more frequent interruption raises the odds of a missed to-do

What it is

Being interrupted mid-task — a phone call comes in, you switch apps to handle something else — doesn't just stall progress on the task at hand; it also drains the very resources used to keep a pending intention alive. The more to-dos a person is carrying at once, and the more often they get interrupted, the higher the chance that one of them gets missed entirely — and this isn't directly tied to how important the to-do is or how much the person cares about it.

Why it happens

Keeping an unexecuted intention alive requires a limited pool of cognitive resources to hold it in a kind of "ready to activate" state — the same pool that's also needed to carry out the task at hand and to handle the interruption itself. Every interruption competes for that shared pool: some of it goes toward the new task the interruption brings, some toward relocating back to the original task once the interruption ends, and what's left over for "remembering there's something else pending" shrinks accordingly. The more interruptions accumulate, the worse this competition gets, and the more likely a pending intention is to get crowded out of activation entirely — not because its content was forgotten, but because it lost the competition for resources.

Studying it

The typical design manipulates the number or complexity of interruptions inserted during a task while embedding one or more prospective memory tasks, then compares miss rate and execution delay across interruption conditions. To rule out the alternative explanation that interruption simply distracts people into forgetting the content, researchers usually test content recall separately at the end — if the content is still remembered clearly but no action was taken at the right moment, that points to the interruption draining the resources needed to maintain and activate the intention, rather than damaging memory for the content itself.

Where it stops holding

Where an interruption lands also matters — one that falls in a natural gap between subtasks interferes with prospective memory less than one landing in the middle of an operation sequence; this account describes the cumulative resource-competition effect of interruptions in general, not a claim that every interruption is equally harmful. Also, if the content of an interruption happens to overlap with the suspended intention (the interruption itself surfaces the trigger cue), it can occasionally help reactivate that intention instead — the reverse of interruption acting purely as a resource drain.

Applying it

• In settings that carry several pending to-dos, minimize unnecessary task switching and interruption frequency, and especially avoid stacking several unrelated interruptions within the same window — doing so multiplies the chance that one of the to-dos gets missed.
• When interruption is unavoidable, surface a brief summary of "what's currently pending" right at the point where the interruption ends and the task resumes, moving that information out of the resource-starved place it would otherwise have to be held and into something externally checkable.
• How to check: compare miss rates for the same users under low-interruption versus high-interruption usage patterns; a markedly higher rate under frequent interruption means the scenario needs an externally surfaced pending-state prompt rather than relying on users to hold up under repeated interruption.

Related

• Same group: A6.19.1 Prospective memory means remembering to act at a future time or occasion, distinct from recalling the past · A6.19.2 Event-based prospective memory relies on an external cue; time-based relies on internal monitoring · A6.19.3 Without an explicit reminder, time-based intentions are the ones most likely to be forgotten outright · A6.19.5 System reminders should bind to the triggering context, not just record to-do text
• Nearby: A5.08 Interruption cost and task resumption · A6.02 Working memory capacity
• Search terms: task interruption · prospective memory resource · resumption failure