You learn something good — a capital, a date, the answer to today's question — and immediately pick up your phone. A game, a feed, a quick puzzle. Five minutes later the fact feels smudged, and by tomorrow it may be gone. A study published in Current Psychology in 2026 (Fisher and Mercer) suggests that smudge is not your imagination: what you do in the first few minutes after learning can measurably change how much of it you keep — even when the thing you do next has nothing to do with what you learned.
The effect has a name, non-specific retroactive interference: new mental activity disrupting a memory that has only just been formed. And the 2026 study is the strongest test yet of whether it applies to the kind of material trivia players actually deal in — meaningful facts, not word lists.
The idea is not new. In 1900, German researchers Georg Müller and Alfons Pilzecker reported that people who memorised verbal material and then completed a picture-description task recalled far fewer of the original items than people given an unfilled delay — about 24 percent recovered after the picture task, versus 56 percent in the interference-free condition. The interfering task did not resemble the learned material at all. It was simply effortful and different, and that was enough.
For the next century the finding refused to settle. Some studies found interference from visual search tasks, spot-the-difference puzzles, video games, and even daydreaming. Others found nothing. A 2026 review of the record would call it inconsistent — which is exactly why Fisher and Mercer designed their study the way they did: bigger samples than the field's usual few dozen, and real factual material instead of artificial lists.
Across three experiments, Luke Fisher (Aston University) and Tom Mercer (University of Wolverhampton) had participants learn sets of trivia facts — the example given is that the word "squirrel" comes from a Greek term meaning "shadow-tail" — then spend a five-minute delay either in quiet wakeful rest or playing spot-the-difference puzzles, then recall the facts.
Experiment 1 (121 undergraduates, 20 facts) produced the honest complication first: averaged across everyone, there was no overall interference effect. Recall sat around a quarter of the facts in both conditions. But a significant interaction with condition order appeared — when the puzzle task came second in the session, recall dropped markedly, while the rest condition showed no such decline. Interference, it seemed, hit harder later in a session, when mental fatigue and earlier learning had built up.
Experiment 2 (100 new participants) tightened the design: shorter lists of 14 facts, equated for interest, and no interest-rating probes during learning (the researchers suspected that rating each fact had forced deeper processing in Experiment 1, armouring the memories). This time the effect was clear: recall of roughly 42 percent after rest versus 35 percent after puzzles — and again, distraction arriving late in the session produced the worst performance of all.
Experiment 3 (135 participants, online) removed the order confound entirely by giving each person only one condition, twice. The puzzle group recalled consistently fewer facts than the rest group, and the gap grew: the interference group's recall declined by a median of about 21 percent from the first test to the second, versus roughly 6 percent for the rest group. Both groups forgot over time — distraction made the slope steeper.
The researchers are careful here, and the two classic explanations both survive contact with the data only partially.
The consolidation account says a fresh memory trace is fragile for its first minutes and needs quiet biological stabilisation time. Divert attention to an effortful task in that window and stabilisation is interrupted; the trace weakens. This fits the older wakeful-rest literature well: in experiments by Michaela Dewar and colleagues (2012, 2014), ten minutes of quiet rest after learning stories or word-like material protected memory better than the same ten minutes of spot-the-difference — and in the 2012 study the benefit was still measurable seven days later, even when rehearsing the material during rest was made impossible.
The temporal distinctiveness account says the problem is crowding, not fragility: a filled delay surrounds the new memory with psychological neighbours, making it harder to pick out later. This fits a detail consolidation struggles with — rest before learning also seems to help, plausibly by isolating what comes next.
Fisher and Mercer conclude that neither theory alone explains the whole pattern, including the order effects. Their study was not designed to pick a winner, and they say so.
Three limits deserve stating plainly, because this topic attracts oversold headlines.
First, the effects are modest and context-sensitive. Experiment 1 found no overall effect at all; the interference appeared reliably only under particular conditions (cleaner encoding in Experiment 2, between-subjects design in Experiment 3, distraction late in a session). Second, the trivia facts were not controlled for truthfulness or for what participants already knew — prior knowledge varies, and known facts behave differently from new ones. Third, the headline proportional-decline result in Experiment 3 reached significance under a one-tailed test, the more lenient standard the authors had pre-committed to. The direction of the evidence is consistent across three experiments and with the older rest literature, but "a puzzle erases your memory" is a stronger sentence than the data supports. "A busy five minutes after learning costs you some facts, sometimes" is closer.
For trivia players and anyone studying facts, the practical version is cheap and low-risk:
None of this requires an app, a supplement, or a belief about brain waves. It requires doing nothing, briefly, on purpose — which is presumably why so few people do it.
The lab distractor was spot-the-difference, but the theory is about effortful new activity in general — earlier studies found similar interference from visual search and video games, and some found none at all. The 2026 contribution is showing the effect with meaningful facts and large samples, not proving every distraction is equal.
The 2026 experiments used five minutes; the Dewar rest studies used ten. There is no established minimum dose. A few genuinely input-free minutes is the tested idea, not a magic number.
No. A tip-of-the-tongue state is a retrieval jam on a fact you still hold. Retroactive interference is about the fact never stabilising in the first place — the loss happens at encoding's exit, not at recall.
No — the studies used learning sets followed by one delay. Rest after a round, a page, or a study block. Resting after each individual fact would make learning slower than the interference ever could.
This article is the starting point. Copy any of these prompts into your favorite AI assistant to learn more about retroactive interference and wakeful rest:
🔎 Build a study session with quiet minutes engineered in
I am studying [topic, e.g. "European geography facts for a pub quiz"] in sessions of about [30] minutes. Design the session using the wakeful-rest research: split it into learning blocks, tell me exactly where to put 5-minute no-input rest breaks (no phone, no music, no reading), and end with a self-test. Explain in one line why each break sits where it does, referencing retroactive interference.
🧭 Stress-test the 2026 study like a reviewer
Summarise Fisher and Mercer's 2026 Current Psychology study on non-specific retroactive interference for meaningful material (three experiments, trivia facts, spot-the-difference vs wakeful rest). Then argue both sides: the strongest case that the effect is real and matters outside the lab, and the strongest skeptical case (including Experiment 1's null overall effect and the one-tailed test in Experiment 3). End with what experiment would settle it.
💡 Run a one-week experiment on yourself
Help me test the wakeful-rest effect on myself. I will learn two matched sets of 15 trivia facts on different days: after one set I rest quietly for 5 minutes, after the other I play a phone game for 5 minutes. Give me: how to match the sets for difficulty, the exact daily protocol, a scoring sheet for a recall test 24 hours later, and how to interpret the result honestly given that my sample size is one person.
Tip: replace the bracketed parts with your own situation — the more specific your prompt, the more useful the answer.
A 2026 Current Psychology study by Fisher and Mercer — three experiments, hundreds of participants, real trivia facts — found that five minutes of spot-the-difference puzzles after learning cost people facts compared with five minutes of quiet wakeful rest (about 42% vs 35% recall in the cleanest within-subject test), with interference hitting hardest when it arrived late in a session. The effect is modest, condition-sensitive, and honestly caveated by its own authors, but it lines up with a decade of wakeful-rest research reaching back to Müller and Pilzecker in 1900.
The single most important takeaway: the minutes right after learning are part of the learning. Protect them — five quiet, input-free minutes after a study block or quiz round is the cheapest memory upgrade the research currently offers.
One sharper AI prompt + one new daily trivia topic every week. Free forever, unsubscribe anytime.
We never share your email. One email a week, that's the deal.
Published October 31, 2026