Research Brief · The Evidence, Plainly
Curiosity Aims Narrow
Wanting to know something helps you remember it. That claim is well supported. A second claim rode out of the same research and is now the more widely repeated of the two: that during a curious state your brain also encodes whatever else happens to be present. A 2026 meta-analysis pooling 47 independent studies found a moderate effect on recall of the information people were curious about and a small effect on incidental information encountered in the same state. Three separate experiments testing the pre-testing version of the effect found a reliable benefit for the target and no benefit at all for the material shown just before it. This brief sets out the imaging work, the meta-analysis, the replication with its null result on sleep, and the counter-evidence, and grades every study by tier.
What was actually studied
The best current summary of the size of the effect. A multilevel random-effects model integrated data from 47 independent studies on the relationship between state curiosity and recall. The result was a moderate effect (g = 0.44) of high curiosity states on recall of task-specific information, meaning the information people were curious about, and a small effect (g = 0.13) on recall of incidental information encountered during the same state. The effect was not moderated by schedule of recall, type of state curiosity, age group, or duration of exposure to the information. The authors report evidence that high curiosity states are associated with enhanced recall in delayed-recall conditions specifically, and read the pattern as consistent with reward-based learning, with facilitated encoding and consolidation, and with attentional and dopaminergic processes that may support incidental learning. A meta-analysis inherits the design limits of the studies it pools; most of these are laboratory trivia paradigms in convenience samples.
The mechanism study, and the origin of the popular version. Participants rated trivia questions for curiosity, then were scanned while they waited for answers, with unrelated neutral faces presented during the wait. Memory for the answers was better when curiosity had been high, in both an immediate test and a one-day delayed test, and memory for the incidental faces was also better when they had appeared during high-curiosity waits. Activity in the midbrain and the nucleus accumbens was enhanced during high-curiosity states. Individual variability in the incidental-memory benefit was supported by anticipatory activity in the midbrain and hippocampus and by functional connectivity between those regions, which is to say the spillover was not uniform across people even in the study that found it. The authors describe a link between the mechanisms supporting extrinsic reward motivation and intrinsic curiosity. This is a small functional imaging study and the dopaminergic account is inferred from the regions involved, not measured directly.
The earlier finding the 2014 work built on. Participants were scanned while reading trivia questions. Curiosity ratings correlated with activity in caudate regions previously implicated in anticipated reward. A behavioural follow-up showed that people spent scarce resources, either limited tokens or waiting time, to find out answers they were more curious about. The imaging also showed that curiosity increased activity in memory areas when participants guessed incorrectly, and a further behavioural study confirmed the prediction that followed from it: higher curiosity in the initial session correlated with better recall of surprising answers one to two weeks later. The willingness-to-pay finding and the delayed-recall finding are the two most useful results here, and both are correlational within subject.
The durability test, including a null result worth knowing. Participants learned answers to trivia questions they had previously rated for curiosity, with faces shown between each question and its answer. Memory for answers and faces was tested either immediately or after a twelve-hour delay containing either sleep or wakefulness, with polysomnography collected for a subset of the sleep group. The curiosity effect for both answers and incidentally learned faces was replicated in the immediate tests and after the twelve-hour delay. It was not affected by the presence of sleep in either case, and showed no relationship with total sleep time or time in slow-wave sleep. The authors conclude that curiosity may be a learning factor not subsequently modulated by sleep-dependent consolidation, and call for further work in other contexts. This study is the reason the issue tells readers to space the check rather than to count on an overnight gain.
The direct test of the spillover claim, and it did not survive. Guessing an answer to an unfamiliar question before seeing the answer improves memory for that answer, an effect known as the pre-testing effect, which some accounts attribute to induced curiosity. If pre-testing works by inducing a curiosity state during which encoding is generally enhanced, then it should also improve memory for incidental material presented at the same time. Three experiments varied the order, nature and timing of incidental material presented within a pre-testing context. All three produced a reliable pre-testing effect for the targets. None produced a benefit for the incidental material presented before the target. The authors conclude that the pre-testing effect is highly specific and is not consistent with a generalised state of curiosity. This is a boundary condition on the popular reading, not a refutation of the curiosity literature as a whole.
The framework the field organises around. The PACE framework (Prediction, Appraisal, Curiosity, and Exploration) proposes that curiosity states relate to modulations in dopaminergic circuit activity, and that these modulations affect memory encoding and consolidation for both targets of curiosity and incidental information encountered during curiosity states. It is included here for orientation and because the pre-testing experiments above were designed as a test of one of its predictions. A framework paper generates testable predictions; it is not itself evidence for them, which is why it sits at Tier 2 in this brief.
What they found
People remember information they were curious about better than information they were not. This is the well-replicated core of the literature, present in the 2009 imaging and behavioural work (Kang et al., 2009), the 2014 imaging study at both immediate and one-day tests (Gruber et al., 2014), and the 2018 replication at immediate and twelve-hour tests (Stare et al., 2018).
The incidental or spillover effect is real in some paradigms and much weaker than the targeted one. The 2014 study found better memory for neutral faces presented during high-curiosity waits, and the 2018 study replicated it at twelve hours. The 2026 meta-analysis puts the pooled size of that effect at g = 0.13, against g = 0.44 for the information people actually wanted.
In the design that tested the spillover most directly, it did not appear. Three experiments varying the order, nature and timing of incidental material within a pre-testing context produced a reliable benefit for the target every time and no benefit for the incidental material presented before it. The authors concluded that the pre-testing effect is target-specific and not consistent with a generalised state of curiosity (Hollins et al., 2022).
Curiosity carries a behavioural price signal as well as a memory one. In 2009, participants gave up limited tokens and accepted waiting time to receive answers they were more curious about, and curiosity recruited memory regions most strongly when a guess had been wrong. Higher initial curiosity correlated with better recall of surprising answers one to two weeks later (Kang et al., 2009).
Sleep did not add to the effect. In the 2018 replication, the curiosity advantage for both answers and incidental faces survived a twelve-hour delay, but was no larger when that delay contained sleep, and showed no relationship with total sleep time or time in slow-wave sleep (Stare et al., 2018).
The meta-analysis found the effect held in delayed-recall conditions and was not moderated by recall schedule, curiosity type, age group, or exposure duration, which makes the advantage look more like an encoding and consolidation difference than a short-lived arousal effect (Radhakrishna & Padakannaya, 2026).
Where the evidence stands
The well-supported claim is narrow and useful: a high-curiosity state at encoding is associated with better later recall of the information the curiosity was about, at a moderate pooled effect size across 47 studies, and that advantage is still present at a delay. Three independent research lines point the same way, using different paradigms, and the largest synthesis available finds the effect stable across age groups and recall schedules.
The under-supported claim is the one that travelled furthest. Encountering unrelated material during a curious state is associated with only a small recall advantage in the pooled data, and when three experiments set out to test that spillover directly in a pre-testing paradigm, they found no benefit for the material presented before the target. The honest reading is that the general-state account is not supported as a practical mechanism, whatever residual incidental effect exists in particular laboratory setups.
What follows for a reader is a change of operation rather than a change of attitude. If the benefit attaches to the target, then the useful move is to create a target: name the question before opening the source, commit to an answer before checking it, and give material that matters but does not interest you its own question rather than parking it next to something more appealing. That is the practical content of this issue, and it is what the evidence supports. The evidence does not support arranging your day around being in a curious mood.
What this does not prove
This is a laboratory literature. The dominant paradigm is trivia questions rated for curiosity in a single session, usually with university-age convenience samples. Recall of a trivia answer is not the same task as understanding a technical paper, a legal brief, or a set of documentation, and no study here tested professional learning material.
The imaging studies are small and the dopamine account is inferred. The 2014 study reports enhanced midbrain and nucleus accumbens activity during high-curiosity states; it does not measure dopamine. The circuit is described as dopaminergic on the basis of its anatomy and prior work, which is a reasonable inference and not a direct observation.
The 2026 pooled numbers are effect sizes, not a ratio to be quoted. A g of 0.44 against a g of 0.13 is the difference between a moderate and a small effect. It is not a claim that one benefit is a fixed multiple of the other, and this brief does not make that claim.
The counter-evidence is specific in scope. The three experiments that found no spillover used a pre-testing manipulation, which induces curiosity indirectly. They are strong evidence against a generalised-state account of pre-testing. They do not refute the incidental findings in the studies that measured self-rated curiosity directly, and the honest position is that the two literatures disagree about how far the effect reaches.
The sleep result is a single null in a single design. The 2018 study found no sleep benefit over a twelve-hour delay with a subset polysomnography sample. That is a reason not to promise an overnight gain. It is not established that sleep is irrelevant to curiosity-driven memory in every context, and the authors themselves say more work is needed.
Individual variation is real and unaddressed by the practice advice. Even in the study that found the incidental benefit, its size tracked anticipatory midbrain and hippocampal activity and the connectivity between those regions, meaning it was not uniform across participants. Nothing here predicts how large the effect will be for any particular reader.
None of this is medical advice, and none of it is a treatment for an attention or memory difficulty. Persistent trouble with memory, concentration, or learning is a clinical question for a physician or a qualified specialist, not a study-technique question. No supplement is recommended anywhere in this issue.
What it means for you
Write the question before you open the source, because across three research lines the reliable advantage attaches to the information a person was trying to retrieve, and in the pre-testing experiments it attached to nothing else. Guess first and then check, because curiosity recruited memory regions most strongly when a guess had been wrong, and surprising answers were the ones still recalled one to two weeks later. Stop parking dull material beside interesting material, because that is precisely the arrangement the three-experiment test examined and found no benefit from. Spend the interest you already have, because participants demonstrably gave up tokens and waiting time for answers they wanted, which makes existing curiosity a resource worth routing rather than a mood worth waiting for. And check yourself at a delay rather than at the end of the session, because the advantage was present at one day, at twelve hours, and in the meta-analysis under delayed-recall conditions, but was no larger when the delay contained sleep. The newsletter walks through those five moves. The research above is why they are the five, and where each one runs out.
Go to the source
- Tier 1 · Meta-analysis Radhakrishna, P., & Padakannaya, P. (2026). Mnemonic benefits of state curiosity: A meta-analysis. Psychonomic Bulletin & Review, 33(3).
doi.org/10.3758/s13423-025-02800-8 - Tier 1 · Functional MRI study Gruber, M. J., Gelman, B. D., & Ranganath, C. (2014). States of curiosity modulate hippocampus-dependent learning via the dopaminergic circuit. Neuron, 84(2), 486-496.
doi.org/10.1016/j.neuron.2014.08.060 - Tier 1 · Functional MRI and behavioural studies Kang, M. J., Hsu, M., Krajbich, I. M., Loewenstein, G., McClure, S. M., Wang, J. T., & Camerer, C. F. (2009). The wick in the candle of learning: Epistemic curiosity activates reward circuitry and enhances memory. Psychological Science, 20(8), 963-973.
doi.org/10.1111/j.1467-9280.2009.02402.x - Tier 1 · Replication with polysomnography Stare, C. J., Gruber, M. J., Nadel, L., Ranganath, C., & Gómez, R. L. (2018). Curiosity-driven memory enhancement persists over time but does not benefit from post-learning sleep. Cognitive Neuroscience, 9(3-4), 100-115.
doi.org/10.1080/17588928.2018.1513399 - Tier 1 · Three experiments, counter-evidence Hollins, T. J., Seabrooke, T., Inkster, A., Wills, A., & Mitchell, C. J. (2022). Pre-testing effects are target-specific and are not driven by a generalised state of curiosity. Memory, 31(2), 282-296.
doi.org/10.1080/09658211.2022.2153141 - Tier 2 · Theoretical review Gruber, M. J., & Ranganath, C. (2019). How curiosity enhances hippocampus-dependent memory: The prediction, appraisal, curiosity, and exploration (PACE) framework. Trends in Cognitive Sciences, 23(12), 1014-1025.
doi.org/10.1016/j.tics.2019.10.003
Tier 1 means peer-reviewed primary research or meta-analysis, the strongest evidence. Tier 2 means an expert framework or smaller study that traces to peer-reviewed work. We grade every source so you can see the weight behind each claim.
Pegasus Realm publishes Wellness Team Culture and the practice resources behind it.