Curiosity recruits the ventral tegmental area, the nucleus accumbens, and the hippocampus, a mesolimbic circuit documented in Neuron in 2014 and confirmed by a 2026 meta analysis of 47 studies. That mesolimbic pathway, the same dopaminergic system that values food, money, and social approval, also assigns value to the anticipation of new information, and the strength of that valuation signal predicts what a person will later remember. A functional magnetic resonance imaging study of adults published in Psychological Science in 2009 found that curiosity ratings for trivia questions correlated with activity in the caudate nucleus, and that questions eliciting stronger curiosity were recalled more reliably one to two weeks later. A separate 2014 Neuron study extended the finding: heightened curiosity increased functional coupling between the hippocampus and the midbrain dopaminergic circuit, and that coupling improved memory for the sought information, and improved memory too for unrelated faces encountered incidentally during the curious state. A 2026 meta analysis pooling 47 independent studies in Psychonomic Bulletin and Review found a moderate memory benefit for task relevant material, Hedges g of 0.44, and a small but reliable benefit for incidental material, Hedges g of 0.13. The principal scientific dispute concerns mechanism rather than existence: whether curiosity operates as a single unified drive, as the original 1994 information gap theory proposed, or as multiple partially dissociable systems for specific and diversive curiosity, as a 2015 Neuron review argued. For an employer, an educator, or a technologist building a learning product, the finding that survives the scrutiny of replication and meta analysis is narrower than the popular claim: curiosity reliably strengthens memory for the information it targets, the evidence for a broad halo effect on unrelated incidental learning is real but smaller, and no study yet demonstrates that manufacturing curiosity produces lasting gains in general cognitive ability. This report treats what a normally functioning brain does when curiosity is engaged; it is not a clinical evaluation and does not address any diagnosed cognitive or attentional condition.
B. The Question and the Population or Model It Applies To
What does curiosity do to brain activity and to memory, in whom has this been measured, and by what method. The construct under study is state curiosity, the transient motivational condition triggered by an information gap, distinguished throughout this report from trait curiosity, the relatively stable individual difference in a person’s general disposition to seek novelty and knowledge. The two are related but not interchangeable: a 2025 study of 1,218 adults ages 20 to 84, published in PLOS ONE, found that trait curiosity declined modestly across the adult lifespan while state curiosity, measured through responses to trivia questions, rose, a divergence that would be invisible to any study measuring only one construct.
The literature under review here drew its participants overwhelmingly from university subject pools in the United States, Germany, Japan, and the United Kingdom, healthy adults age 18 to 35 in the large majority of the foundational imaging studies, with the lifespan and aging literature extending the age range and the developmental literature extending downward into early childhood. The paradigm that recurs across the primary literature is the trivia question task: a participant rates curiosity about the answer to a factual question, waits through a delay, receives or does not receive the answer, and is tested on retention later, sometimes the same day and sometimes one to two weeks afterward. A second paradigm, drawn from behavioral economics, measures how much a participant will pay in money, time, or points to see an answer, information gap theory’s operational signature. No study in this review measured curiosity in a clinical population defined by a diagnosed psychiatric or neurological condition, and the animal literature relevant to mechanism used macaque monkeys in a single unit recording paradigm rather than a curiosity task in the colloquial human sense.
C. Neural Mechanism
What brain circuit does curiosity activate
Curiosity activates the mesolimbic dopaminergic circuit, the same ventral tegmental area, nucleus accumbens, and caudate pathway that computes the value of food, money, and other primary and secondary rewards. Single unit recordings in two macaque monkeys, published in Neuron in 2009, established the foundational mechanism before any human imaging study confirmed it: dopamine neurons in the midbrain fired in proportion to a monkey’s preference for advance information about an upcoming reward, independent of the reward’s own magnitude, demonstrating that the dopaminergic system treats information itself as a valued commodity rather than merely as a predictor of value. That finding reframes curiosity not as a separate cognitive faculty bolted onto the reward system but as the reward system operating on a different class of object.
The 2009 Psychological Science trivia study extended this logic to humans: caudate activation during question reading tracked self reported curiosity ratings, and the same reward related activation predicted which answers a participant would recall one to two weeks afterward. The 2014 Neuron study went further, showing that curiosity does more than activate the reward circuit in isolation: it increases functional connectivity between the hippocampus, the structure responsible for consolidating new episodic memories, and the ventral tegmental area and nucleus accumbens. That increased coupling is the mechanistic bridge between motivation and memory: dopamine released into the hippocampus during a curious state appears to enhance long term potentiation, the synaptic strengthening process that converts a transient experience into a stable memory trace. A 2019 study in the Proceedings of the National Academy of Sciences using functional imaging found that the ventromedial prefrontal cortex and the striatum represent monetary reward value and information value on a common neural scale, evidence for a shared currency rather than two parallel systems that happen to overlap anatomically.
Does curiosity resolution itself feel rewarding
Curiosity resolution, the moment an answer arrives, produces its own reward signature distinct from the anticipatory phase. A 2024 study in the Journal of Cognitive Neuroscience recorded event related potentials and found that receiving the answer to a high curiosity question elicited larger reward sensitive components than receiving the answer to a low curiosity question, a electrophysiological signature closely resembling the brain’s response to unexpected monetary gain. A 2016 study in the Journal of Experimental Psychology, General reported a parallel behavioral finding: participants weighted the positive prospect of learning something more heavily than the negative prospect of remaining ignorant when deciding whether to seek information, and prediction errors tied to receiving or being denied information, not merely the presence of a gap, drove learning in a manner formally analogous to a reward prediction error in reinforcement learning models.
Does curiosity operate through one mechanism or several
The evidence does not settle on a single mechanism, and a 2015 Neuron review is explicit that curiosity is unlikely to be a unitary construct. The review separates perceptual curiosity, the drive triggered by an ambiguous or incomplete sensory stimulus, from epistemic curiosity, the drive triggered by an information gap in conceptual knowledge, and separates specific curiosity, directed at a particular piece of missing information, from diversive curiosity, the more general appetite for novelty and stimulation. A 2013 theoretical synthesis in Trends in Cognitive Sciences proposed a computational account in which curiosity tracks learning progress, the rate at which a novel stimulus reduces prediction error, rather than tracking novelty or complexity directly, a formulation that connects the psychological construct to reinforcement learning models built to give artificial agents an intrinsic exploration drive. Which of these accounts, information gap, reward prediction error, or learning progress, best describes the dopaminergic signal in a given task remains an active empirical dispute rather than a settled finding, and the studies reviewed here support more than one of them under different task conditions.
D. The Evidence Base
Systematic evidence, memory benefit: Radhakrishna and Padakannaya, Psychonomic Bulletin and Review, 2026, Article 97. Meta analysis of 47 independent studies of state curiosity and memory. Pooled effect for task specific recall, Hedges g of 0.44, a moderate effect. Pooled effect for incidental recall, Hedges g of 0.13, a small effect. The authors report the effect held across recall delay, the specific curiosity measure used, and participant age group, and identify reward based valuation, encoding and consolidation support, and attentional and dopaminergic modulation as the three converging mechanisms proposed across the included studies.
Systematic evidence, intervention effect: Schutte and Malouff, Current Psychology, volume 42, issue 24, pages 20374 to 20384, 2023. Meta analysis of 41 randomized controlled trials of curiosity enhancing interventions, a combined sample of 4,496 participants. Weighted effect, Hedges g of 0.57, ninety five percent confidence interval 0.44 to 0.70. After trim and fill adjustment for publication bias, the estimate fell to 0.49, confidence interval 0.36 to 0.62, still a moderate effect. The strongest results came from interventions using mystery or game based framing, and the effect held across participant age group and measurement method.
Original study, memory and reward circuitry: Kang, Hsu, Krajbich, Loewenstein, McClure, Wang, and Camerer, Psychological Science, volume 20, issue 8, pages 963 to 973, 2009. Functional imaging during a trivia question task in healthy adults. Curiosity ratings correlated with caudate nucleus activation, and curiosity level predicted delayed recall accuracy one to two weeks after the imaging session. Participants also allocated more of a fixed resource to see answers to questions that produced stronger curiosity, a behavioral economics measure of information value.
Original study, hippocampal dopaminergic coupling: Gruber, Gelman, and Ranganath, Neuron, volume 84, issue 2, pages 486 to 496, 2014. Functional imaging in healthy adults during a trivia paradigm with an incidental face learning task inserted during a fourteen second anticipatory delay. High curiosity states increased activation in the ventral tegmental area and nucleus accumbens, increased hippocampal to midbrain functional connectivity, and improved recognition memory for the incidentally viewed faces at a twenty four hour retest, evidence that a curious state broadens its memory benefit beyond the target information itself.
Original study, information as reward in a nonhuman primate: Bromberg-Martin and Hikosaka, Neuron, volume 63, issue 1, pages 119 to 126, 2009. Single unit recording of midbrain dopamine neurons in two rhesus macaques. Dopamine neuron firing tracked the animals behavioral preference for advance information about an upcoming reward, independent of the reward magnitude itself, the animal model evidence that the dopaminergic valuation system treats information as an object of value in its own right rather than only as a cue that predicts value.
Original study, common neural currency: Kobayashi and Hsu, Proceedings of the National Academy of Sciences, volume 116, issue 26, pages 13061 to 13066, 2019. Functional imaging in healthy adults comparing choices for monetary reward against choices for information. Overlapping representations in the ventromedial prefrontal cortex and the striatum for reward value and information value, evidence for a shared valuation currency rather than two independent systems.
Original study, information prediction error: Marvin and Shohamy, Journal of Experimental Psychology, General, volume 145, issue 3, pages 266 to 272, 2016. Behavioral choice task in healthy adults. Positive framing of an information seeking choice, the prospect of learning something, predicted choice more strongly than negative framing, the prospect of avoiding ignorance, and prediction errors attached to receiving or being denied information enhanced subsequent learning in a pattern formally parallel to reward prediction error.
Original study, real world information seeking: Eschmann, Pereira, Valji, Dehmelt, and Gruber, Social Cognitive and Affective Neuroscience, volume 18, issue 1, article nsac050, 2023. Functional imaging combined with an ecological momentary assessment protocol tracking information seeking outside the laboratory. Resting state mesolimbic functional connectivity predicted the frequency of real world information seeking behavior recorded over the following days, extending the laboratory finding to naturalistic behavior.
Original study, decisions under risk: Lau, Ozono, Kuratomi, Komiya, and Murayama, Nature Human Behaviour, volume 4, pages 531 to 543, 2020. Functional imaging during a task in which participants chose whether to satisfy curiosity or hunger at the risk of a mild electric shock. Shared striatal activity across the two decision types, evidence that the same valuation architecture that governs a primary biological drive also governs the drive to resolve an information gap, including when resolution carries a cost.
Original study, reward sensitive electrophysiology: Rüterbories and colleagues, Journal of Cognitive Neuroscience, volume 36, issue 5, pages 888 to 903, 2024. Event related potential recording during a trivia paradigm. Reward sensitive components were larger following delivery of answers to high curiosity questions than low curiosity questions, a millisecond resolution electrophysiological correlate of the reward circuit finding from the imaging literature.
Theoretical and computational synthesis: Gottlieb, Oudeyer, Lopes, and Baranes, Trends in Cognitive Sciences, volume 17, issue 11, pages 585 to 596, 2013. A synthesis proposing that curiosity tracks learning progress rather than novelty or complexity directly, linking psychological and neuroscientific findings to computational reinforcement learning models built to give artificial agents an intrinsic exploration drive.
Foundational review and theory: Kidd and Hayden, Neuron, volume 88, issue 3, pages 449 to 460, 2015. A Neuron perspective article proposing that curiosity is not a unitary construct, distinguishing perceptual from epistemic curiosity and specific from diversive curiosity, and reviewing developmental and comparative evidence for each.
Foundational theory: Loewenstein, Psychological Bulletin, volume 116, issue 1, pages 75 to 98, 1994. The information gap theory, proposing that curiosity is an aversive state of felt deprivation produced by the perception of a gap between what a person knows and what a person wants to know, and that the intensity of curiosity scales with the salience of the gap rather than with the novelty of the missing information alone.
Lifespan study: Whatley and colleagues, PLOS ONE, 2025, article e0320600. A preregistered study of 1,218 adults age 20 to 84, following a pilot of 193, using a trivia rating task. Trait curiosity, measured by a standardized epistemic curiosity scale, declined modestly with age. State curiosity, measured by responses to the trivia task, increased with age, with an exploratory quadratic pattern showing a dip in middle age.
E. Measurement, Method Validity, and the Reverse Inference Problem
Every instrument carrying weight in this literature has a specific temporal or spatial resolution, and none measures curiosity directly, each measures a proxy the field treats as valid to varying degrees. Functional magnetic resonance imaging, the dominant method in the primary studies reviewed here, offers spatial resolution on the order of millimeters but temporal resolution on the order of seconds, adequate to localize the ventral tegmental area, nucleus accumbens, caudate, and hippocampus but too coarse to separate the anticipatory phase of curiosity from the resolution phase without a carefully staggered trial design. Event related potential recording, used in the 2024 Journal of Cognitive Neuroscience study, inverts that trade off, offering millisecond temporal resolution but only coarse localization to broad scalp regions rather than specific subcortical structures, which is why the imaging and electrophysiological literatures are treated as complementary rather than redundant in this report.
A caudate or ventral tegmental area activation during a curiosity task does not, by itself, prove that the region’s involvement is specific to curiosity rather than to reward processing in general, the reverse inference problem named in a 2006 Trends in Cognitive Sciences analysis. Most subcortical reward structures participate in dozens of cognitive and motivational processes, and a popular account that treats caudate activation as a curiosity signature, full stop, has already committed the error the 2006 analysis warned against. The studies reviewed here largely avoid this error by design rather than by caveat: the 2009 and 2014 studies used within subject curiosity ratings as a continuous covariate rather than treating any region’s activation as a categorical marker of curiosity, and the 2019 Proceedings of the National Academy of Sciences study directly tested for overlap between reward and information value representations rather than assuming it. Where a popular summary of this literature claims that a specific brain region is the seat of curiosity, that claim outruns what a correlational activation finding can support.
Multiple comparisons correction matters specifically in whole brain imaging analysis, where testing tens of thousands of voxels without correction inflates the false positive rate well past the nominal five percent threshold. The 2014 Neuron study and the 2019 Proceedings of the National Academy of Sciences study both report a family wise error or cluster based correction for their primary findings, and this report treats an imaging claim as provisional wherever the correction method is not stated.
F. Sample, Generalizability, and Replication
The sample the strongest of these findings depend on is narrow. The 2009 and 2014 imaging studies, along with the majority of the trivia paradigm literature this report draws on, recruited healthy university age adults, predominantly undergraduate subject pool participants at research universities in the United States. This is the sample the 2010 Behavioral and Brain Sciences analysis of psychology’s evidence base described as western, educated, and drawn from industrialized, rich, and democratic societies, a population that analysis showed to be a statistical outlier on a range of psychological and perceptual measures relative to the rest of the world’s population, not a representative default. A curiosity finding demonstrated in a sample of American or German undergraduates in their twenties does not, on its own, establish how curiosity operates in a sixty year old factory worker in Jakarta or a subsistence farmer in rural Kenya, and the literature reviewed here has not closed that gap.
Replication status varies across the findings reported here rather than clustering at one level of confidence. The core behavioral finding, that self reported curiosity predicts subsequent memory for the target information, has now been confirmed at the level of a 2026 meta analysis pooling 47 independent studies, the strongest form of replication evidence a psychological or neuroscientific claim can carry. The narrower claim that curiosity broadens memory for incidental, unrelated information carries a smaller and more variable effect, Hedges g of 0.13 in the same meta analysis, and should be treated as a real but modest finding rather than the dramatic halo effect some popular accounts describe. The single unit dopamine recording finding in macaques has not, to this reviewer’s knowledge, been directly replicated in an independent laboratory using the identical paradigm, though its central claim, that dopamine neurons value information independent of reward magnitude, has since been corroborated by the human imaging evidence for a shared reward and information currency.
The wider replication crisis in psychology bears directly on how much confidence any single study in this literature deserves standing alone. A 2015 replication effort published in Science attempted to reproduce findings from one hundred studies drawn from three major psychology journals and found that a minority produced a statistically significant effect on replication, with replicated effect sizes running smaller than the original reports on average. A 2013 analysis in Nature Reviews Neuroscience calculated that the median statistical power across a broad sample of neuroscience studies sits between roughly eight percent and thirty one percent, well below the eighty percent power convention that a well designed study should target, meaning a substantial share of published neuroscience findings are underpowered enough that a significant result may not reflect a true effect at all. Against that backdrop, a single small sample fMRI study of curiosity, no matter how striking its images, is preliminary evidence, and the meta analytic and multi laboratory findings summarized in Section D carry the analytical weight in this report.
G. Developmental, Age, and Individual Difference Considerations
Curiosity changes shape across the lifespan rather than simply declining or rising in a straight line. The developmental literature on children establishes that epistemic and information seeking behavior is present from the preschool years, expressed through question asking and exploratory play well before a child can complete a self report curiosity inventory, though the imaging literature reviewed in Section C has not extended its methods down into early childhood, so the dopaminergic and hippocampal mechanism demonstrated in adults is inferred rather than directly measured in young children. The 2025 PLOS ONE lifespan study, drawing on 1,218 adults age 20 to 84, found a divergence between the two forms of curiosity this report distinguishes in Section B: trait curiosity, the general disposition, declined modestly across the adult lifespan, while state curiosity, the response to a specific trivia question, rose with age, with an exploratory quadratic pattern suggesting a dip in midlife before a rise in older adulthood. That divergence matters because a study measuring only trait curiosity would report that older adults are less curious, while a study measuring only state curiosity would report the opposite, and neither statement alone describes what the literature actually shows.
Individual differences beyond age remain underexplored in the imaging literature specifically. The behavioral literature identifies stable individual differences in trait curiosity using validated inventories, and a small but growing literature connects higher trait curiosity in older adults to better preserved cognitive function, though this report treats that association as correlational until a study demonstrates the causal direction, since it remains equally plausible that preserved cognitive function supports sustained curiosity rather than the reverse.
H. Practical Application
An educator building a curriculum around curiosity gains the strongest support from the intervention literature rather than the imaging literature directly: the 2023 Current Psychology meta analysis of 41 randomized controlled trials found a moderate effect, Hedges g of 0.49 after adjustment for publication bias, for interventions that deliberately induce curiosity before instruction, with mystery and game framing outperforming other approaches. The magnitude of that effect, a moderate improvement rather than a transformation of learning capacity, is the number a school administrator or corporate training designer should carry into a budget conversation, not the more dramatic claim that curiosity training rewires the brain for general learning.
A technologist building a product on this evidence should distinguish two claims the literature supports from a third claim it does not. The literature supports that curiosity, once engaged, reliably improves memory for the specific content that produced it, and supports a smaller but real spillover benefit to incidental content encountered during the same window. The literature does not support the claim that repeated engagement of curiosity produces a lasting, general improvement in memory capacity or attentional control, the claim that distinguishes a legitimate curiosity based learning feature from an unsubstantiated brain training marketing claim, a distinction with direct commercial consequence addressed in Section K. An employer designing a training program can reasonably expect a state curiosity manipulation, a well framed question preceding new material, to modestly improve retention of that specific material; the employer should not expect it to substitute for spaced repetition or retrieval practice, the two learning techniques with a larger and more consistently replicated evidence base in the cognitive science literature.
I. Boundary with Clinical and Psychiatric Practice
This report addresses curiosity as a normally functioning motivational and memory process in healthy adults and, to a lesser extent, in healthy children and older adults. It does not address anhedonia, apathy, or reduced information seeking behavior as a symptom of a diagnosed psychiatric or neurological condition, and nothing in this report should be read as a description of any named individual’s cognitive or motivational state. Where a reduction in curiosity or exploratory behavior arises as a clinical concern, a diagnostic question, or a treatment question, that question belongs with a qualified clinician and, for a question about the diagnosis or treatment of a specific condition affecting motivation or memory, with Project Health rather than with this report. This report summarizes published research on brain and behavior; it does not constitute medical, psychiatric, or psychological advice.
J. Neurotechnology, Computational Modeling, and the AI Intersection
The computational reinforcement learning literature has, since roughly 2013, built artificial agents with an intrinsic curiosity signal modeled directly on the learning progress account described in Section C: an agent earns a synthetic reward proportional to the reduction in its own prediction error rather than only for reaching an externally defined goal, and that synthetic reward drives exploration in environments with sparse external rewards. The traffic between the two fields runs in both directions. The computational models borrowed their basic architecture, a prediction error signal driving approach behavior, from the dopaminergic reward prediction error literature that the human and primate studies in Section D describe directly, and the computational literature has in turn sharpened the psychological field’s own theoretical vocabulary, giving the loose intuitive idea of an information gap a formal, testable mathematical definition.
No commercial brain computer interface or neurotechnology device reviewed for this report claims to measure curiosity directly as a clinical or consumer product, and this report accordingly makes no finding regarding the validation of any such device. Where a consumer brain training or cognitive enhancement product markets a claim about curiosity, motivation, or learning capacity, the applicable standard is the evidentiary one described in Section H and enforced in the matter described in Section K, not a claim about neurotechnology measurement accuracy specifically.
K. Legal, Regulatory, Privacy, and Commercial Dimension
What law governs a brain training company’s claims about curiosity and memory
The Federal Trade Commission’s 2016 action against a prominent brain training company set the operative standard for any curiosity or cognition based learning product marketed in the United States. The company agreed to pay two million dollars and to stop claims that its program would improve everyday cognitive performance, delay age related cognitive decline, or reduce the cognitive impact of a health condition, after the Commission found the company lacked the competent and reliable scientific evidence the Federal Trade Commission Act requires to support those claims. The evidentiary gap the Commission identified in that matter is the same gap this report identifies in Section H: evidence that an intervention improves performance on the trained task, or even on the specific content a curiosity manipulation targets, does not by itself establish the broader, more marketable claim that the product improves general cognitive ability. A company building a curiosity based learning product should scope its marketing claims to what Section D’s meta analytic evidence actually supports, task specific and modest incidental memory benefit, and should treat any claim about general cognitive enhancement, attentional control, or protection against cognitive decline as requiring its own dedicated clinical evidence, independent of the memory literature reviewed here.
What privacy law governs neural data collected from a curiosity or attention sensing product
A product that infers curiosity or engagement from neural signals, electroencephalography, functional near infrared spectroscopy, or a comparable sensor, has entered a regulatory area that expanded materially in 2024. Colorado amended its Colorado Privacy Act that year to add neural data and biological data to the statute’s definition of sensitive data, the first explicit statutory recognition of neural data as a protected category in the United States, triggering the statute’s heightened consent and processing requirements for any company that collects it. Illinois has separately maintained its Biometric Information Privacy Act since 2008, a statute that already reaches biometric identifiers and has been the basis of substantial private litigation exposure for companies handling biometric data without the statute’s required consent and retention disclosures, and a 2026 legislative proposal in Illinois would extend that statute’s protections explicitly to neural data, following Colorado’s lead. A company building a curiosity sensing feature into a consumer product should treat neural signal data as sensitive data under an expanding, not a settled, legal category, and should build consent and retention practices to the more protective standard rather than to the jurisdiction with the least current coverage.
What the market pays for curiosity based learning products
Commercial cognitive training subscriptions in the United States market in the range of eight to fifteen dollars per month as of 2026, a price the underlying evidence in Section H supports only for the specific, modest, task bound memory benefit the literature demonstrates, not for the broader cognitive enhancement claim the 2016 Federal Trade Commission matter found unsupported. A buyer, whether an individual consumer or a corporate training department, should price the product against the moderate effect size the evidence actually supports rather than against the more expansive marketing claim.
L. What the Evidence Does Not Yet Settle
The mechanism question is not settled. Whether a single dopaminergic valuation signal underlies all forms of curiosity, or whether perceptual, epistemic, specific, and diversive curiosity reflect partially distinct circuits, as the 2015 Neuron review proposed, remains an open empirical dispute rather than a resolved finding, and answering it would require a within subject imaging study that directly contrasts multiple curiosity types in the same participants rather than comparing across separate studies with different paradigms.
The incidental memory benefit’s mechanism is not settled. The 2026 meta analysis confirms the incidental benefit is real but small, Hedges g of 0.13, yet the studies it pools do not agree on whether that benefit reflects a broad, nonspecific arousal effect, a spillover of dopaminergic modulation onto nearby hippocampal encoding regardless of content, or a narrower attentional effect specific to stimuli that share a temporal or spatial context with the curiosity inducing question. A study manipulating the physical and temporal proximity of incidental material to the curiosity inducing question, while holding arousal constant, would separate these accounts.
Generalizability across culture, socioeconomic status, and education level is not established. Because the great majority of the foundational imaging studies drew participants from university subject pools in a small number of wealthy countries, this report cannot state with confidence that the effect sizes reported in Section D apply outside that population, and a multi site study recruiting outside university subject pools across a range of countries and educational backgrounds would close this gap.
Whether curiosity training produces any lasting, general improvement in memory or attentional capacity, as opposed to a specific, session bound benefit, is not settled and the current evidence points against it. A preregistered, adequately powered trial measuring transfer to untrained cognitive tasks weeks or months after a curiosity based intervention would give a company or a school district and, more directly, a regulator evaluating a marketing claim under the standard described in Section K, the evidence current products lack.
The developmental mechanism is not settled. Because the imaging and electrophysiological methods reviewed in Section C have been applied almost exclusively to adults, the dopaminergic and hippocampal account of curiosity in Section C is inferred rather than directly measured in children, and a pediatric neuroimaging study using an age appropriate paradigm would establish whether the same circuit is active as early as the preschool behavior described in Section G suggests.
