CoChat
Sign Up
Cited by 5Cognitive Abilities and TestingCognitive Functions and MemoryNeural and Behavioral Psychology Studies

Examining Working Memory Training for Healthy Adults—A Second-Order Meta-Analysis

Open access
Syed, Maria, Lum, Jarrad A. G., Byrne, Linda K., Skvarc, David·2024·Journal of Intelligence
DOI 10.3390/jintelligence12110114PMC PMC11595675
Open this paper in CoChat — free account

Abstract

Background: Enhancing working memory performance in cognitively and physically healthy individuals is a popular area of research. The results from a large number of studies have now been summarized in multiple meta-analyses. In these reviews, various training methods have been examined, including mindfulness training, adaptive working memory training, physical activity training, and video game training, to examine whether working memory capacity can be improved. This report aggregated the results of these meta-analyses using second-order meta-analytic approaches to ascertain the extent to which working memory functioning can be enhanced in healthy adults. Methods: A total of six meta-analyses of randomized controlled trials that compared working memory interventions to a control group were included in the analyses. These studies were identified after systematically searching three electronic databases: APA PsycInfo, ERIC and Medline. Collectively, the meta-analyses investigated the effects of cognitive programs, mindfulness, video games and physical activity on working memory. Only meta-analyses undertaken with healthy adults aged between 18 and 55 years were included in the report. Results: The results revealed an average improvement in working memory across the included studies compared to the control groups. The findings indicated a small yet significant enhancement in working memory, with a standardized mean difference of 0.335 (95% CI [0.223; 0.447], p < .001). Further analysis tests for superiority effects between the different working memory training programs revealed no significant differences between intervention effect sizes. Conclusion: Based on the findings, it can be concluded that the working memory capacity of healthy adults can be improved through training. However, the effect size is small, so the utility of this type of training in real-life improvements in cognition may be minimal. The evidence does not indicate that one type of working memory training is superior to another.

AI snapshot

Written from the paper’s full text.

TL;DR

This second-order meta-analysis of six meta-analyses of randomized controlled trials found a small but statistically significant improvement in working memory among healthy adults following training. No one training type—working memory task training, mindfulness, physical activity, or video game/gamification—was superior to the others.

Key findings

  • The overall second-order average effect size was small and significant: SMD = 0.335, 95% CI [0.223; 0.447], p < .001.
  • Working memory task training was significantly different from controls: SMD = 0.370, SE = 0.07, 95% CI [0.223; 0.507], p < .001.
  • Mindfulness training (SMD = 0.248, 95% CI [−0.009; 0.506], p = 0.058), resistance exercise training (SMD = 0.350, 95% CI [−0.05; 0.750], p = 0.086), and gamification (SMD = 0.210, 95% CI [−0.230; 0.650], p = 0.350) were not significantly different from controls.
  • Head-to-head tests showed no significant differences between any pairs of interventions; heterogeneity was significant (I² = 67.9%, Q = 18.687, p = .005).

Methods

  • Second-order meta-analysis of meta-analyses of randomized controlled trials; databases APA PsycInfo, ERIC, and Medline were searched in February 2022, with two authors independently screening records.
  • Inclusion criteria required peer-reviewed meta-analyses of RCTs, healthy adults aged 18–55, and training types aimed at improving working memory; exclusions included neural-pattern studies, children/older adults, clinical/developmental/neurological populations, and dietary or pharmacological interventions.
  • From each included meta-analysis, the overall effect size and 95% CI were extracted as Cohen’s d or Hedge’s g, coded so positive values favored training; effect sizes were averaged with a random effects model, intervention moderation was examined using the I² statistic following Tamim et al. (2011), and analyses used Comprehensive Meta-Analysis; AMSTAR 2 assessed included meta-analyses.

Limitations

  • The paper states significant heterogeneity among included studies and that a moderator analysis could not be conducted because only six meta-analyses were included, below the Cochrane minimum of ten.
  • The paper states that some included meta-analyses contained participants under 18 or over 55, and three included healthy and clinical samples, despite the target population being healthy adults aged 18–55; it also notes mixed active and passive control conditions.
  • The paper states that nested effect sizes were not consistently reported across included meta-analyses and that Comprehensive Meta-Analysis cannot accommodate Robust Variance Estimation, which could introduce additional variability; the small number of meta-analyses also limited power and sensitivity for superiority tests.

Model-written summary. Check it against the paper before relying on it.

Integrity signals

Registry

Crossref · 10.3390/jintelligence12110114

Cites

71 of 71 references verified0 citation errors · 0 not checkable

Statistics

Nothing to recompute.

Claims the paper makes

19 of 35 cited sources checked · 14 supported · 2 flagged · accuracy not yet measured

  • Study design and conductown data

    Three included studies had healthy and clinical samples (Gathercole et al. 2019; Cásedas et al. 2020; Vermeir et al. 2020).

    “Three studies included healthy and clinical samples (Gathercole et al. 2019; Cásedas et al. 2020; Vermeir et al. 2020).”
    Quote in the paperNot a substantial claim2 of 3 sources checked
    • cites Gathercole et al. 2019 source
    • cites Cásedas et al. 2020

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Vermeir et al. 2020 source
    • Rests on an assumption: Including meta-analyses whose samples include some clinical participants, children, or older adults, as long as most participants are healthy adults aged 18–55, does not materially bias the second-order estimate for healthy adults aged 18–55.
  • Premisecitation

    Working memory training produces a near-transfer effect, a transfer effect on a task similar to the training task.

    “research has found that working memory training produces a near-transfer effect, a transfer effect on a task similar to the training task (Melby-Lervåg and Hulme 2013; Sala et al. 2019; Soveri et al. 2017).”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites Melby-Lervåg and Hulme 2013

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Sala et al. 2019 source
    • cites Soveri et al. 2017 source
  • Premisecitation

    Sala et al. (2018), in a third meta-analysis investigating the impact of video game training on general cognitive ability in 63 studies encompassing 3286 cognitively and physically healthy participants, found near-zero training effects.

    “However, the third meta-analysis investigated the impact of video game training on the general cognitive ability of 63 studies, encompassing 3286 cognitively and physically healthy participants, and found near-zero training effects.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Sala et al. (2018) source
  • Premisecitation

    Cochrane Collaboration guidelines require a minimum of ten studies to execute a moderator analysis.

    “According to Cochrane Collaboration guidelines (Cumpston et al. 2022), a minimum of ten studies is required to execute a moderator analysis”
    Quote in the paperSubstantial claim1 of 1 sources checked
    • cites Cumpston et al. 2022 source

      The 'Synthesis methods other than meta-analysis' section discusses alternatives when meta-analysis is not possible, but gives no minimum study count for moderator analysis. Searches for 'minimum of ten studies moderator analysis' and 'meta-regression subgroup analysis 10 studies minimum' found no passage establishing the claimed requirement. Page numbers were unavailable in our copy. (LLM, conf=0.75)

      Experimental — accuracy not yet measured
  • Premisecitation

    Multiple studies suggest a small, measurable, positive effect of working memory training on working memory capacity.

    “Multiple studies suggest a small, measurable, positive effect of training on WM capacity”
    Quote in the paperSubstantial claim1 of 3 sources checked
    • cites Melby-Lervåg and Hulme 2013

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Sala et al. 2019 source
    • cites Soveri et al. 2017 source

      The Main analyses and Conclusions (page unavailable) report small, statistically significant positive effects on untrained working-memory tasks across 33 n-back training studies. But the Mechanisms behind transfer section says the pattern mainly suggests task-specific improvement rather than an increase in underlying working-memory capacity. The measured task effects therefore support only part of the claim. (LLM, conf=0.86)

      Experimental — accuracy not yet measured
  • May combine two claims. 13
  • Classified as probably not a substantial claim. 15
  • The quote only partly states this claim. 1

  • May combine two claims.

    Study design and conductnone given

    Three of the included studies had participants less than 18 years of age (Soveri et al. 2017; Vermeir et al. 2020; Gathercole et al. 2019), and all studies had participants who were older than 55 years of age.

    “Three of the included studies had participants less than 18 years of age (Soveri et al. 2017; Vermeir et al. 2020; Gathercole et al. 2019), and all studies had participants who were older than 55 years of age.”
    Quote in the paperNot a substantial claim3 of 3 sources checked
    • cites Soveri et al. 2017 source

      Split into 2 atoms — 0 confirmed, 0 partially supported, 2 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured
    • cites Vermeir et al. 2020 source

      Split into 2 atoms — 0 confirmed, 1 partially supported, 1 disputed, 0 misattributed, 0 unverifiable. Expand the claim to see per-atom verdicts.

      Checked by a reasoning model — may be wrong.

      Experimental — accuracy not yet measured
    • cites Gathercole et al. 2019 source
    • Rests on an assumption: Including meta-analyses whose samples include some clinical participants, children, or older adults, as long as most participants are healthy adults aged 18–55, does not materially bias the second-order estimate for healthy adults aged 18–55.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Regular physical activity positively influences working memory performance more than acute physical activity.

    “Regular physical activity positively influences WM performance more than acute physical activity (Rathore and Lom 2017; Haverkamp et al. 2020).”
    Quote in the paperNot a substantial claim2 of 2 sources checked
    • cites Rathore and Lom 2017 source
    • cites Haverkamp et al. 2020 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Acute physical activity interventions, such as training sessions lasting less than 60 min, have been found to positively influence attention and cognitive flexibility, but not working memory.

    “Conversely, acute physical activity interventions, such as training sessions lasting less than 60 min, have been found to positively influence attention and cognitive flexibility, but not working memory (Haverkamp et al. 2020).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Haverkamp et al. 2020 source

      The Abstract (p. 3) reports improved attention after acute interventions, and the results (p. 7) show no significant effect on working memory. However, the Interpretation (p. 11) says acute interventions did not show a significant overall effect on cognitive flexibility; the finding was instead that shorter sessions were associated with greater cognitive-flexibility effects (p. 12). The studied population was adolescents and young adults. (LLM, conf=0.90)

      Experimental — accuracy not yet measured

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Powers et al. (2013), in a second meta-analysis examining the effects of video game training on 46 experimental studies, found a small to medium effect size indicating that video game training enhances information-processing skills.

    “Their second meta-analysis (Powers et al. 2013) examining the effects of video game training on 46 experimental studies found a small to medium effect size, indicating that video game training enhances information-processing skills.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Powers et al. (2013) source

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Premisecitation

    Long-term physical activity, such as chronic physical activity interventions, has been found to have a positive effect on working memory.

    “Long-term physical activity, such as chronic physical activity interventions, has been found to have a positive effect on working memory (Haverkamp et al. 2020).”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Haverkamp et al. 2020 source

  • May combine two claims.

    Premisecitation

    Rathore and Lom (2017) found a significant, small effect for chronic physical activity intervention and a non-significant effect for acute intervention on working memory.

    “They found a significant, small effect for chronic intervention and a non-significant effect for acute intervention.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Rathore and Lom 2017 source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Powers et al. (2013) compared video game players with non-video game players in 71 quasi-experimental studies and reported a moderate to large effect size indicating that video game players have superior information-processing skills compared to non-video game players.

    “For example, in their meta-analysis, Powers et al. (2013) compared video game players with non-video game players in 71 quasi-experimental studies and reported a moderate to large effect size indicating that video game players have superior information-processing skills compared to non-video game players.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Powers et al. (2013) source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Premisecitation

    Sala et al. (2018), in a first meta-analysis of 66 correlational studies involving 8141 cognitively and physically healthy participants, found a weak positive correlation between memory and video game skill and a significant correlation with spatial ability.

    “The studies involved 8141 cognitively and physically healthy participants, revealing a weak positive correlation between memory and video game skill and a significant correlation with spatial ability.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Sala et al. (2018) source

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims. The quote only partly states this claim.

    Premisecitation

    Sala et al. (2018), in a second meta-analysis comparing video game players with non-players in 98 quasi-experimental studies involving 6166 cognitively and physically healthy participants, found that video game players outperformed non-players on memory and spatial ability measures.

    “The results showed that video game players outperformed non-players on memory and spatial ability measures.”
    Quote in the paperNot a substantial claim1 of 1 sources checked
    • cites Sala et al. (2018) source

  • May combine two claims.

    Premisecitation

    Since 2016, at least ten meta-analytic studies have examined the effects of mindfulness-based interventions on working memory performance in healthy adults, but the evidence is generally inconclusive.

    “Since 2016, at least ten meta-analytic studies have examined the effects of mindfulness-based interventions on WM performance in healthy adults, but the evidence is generally inconclusive.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Premisecitation

    Sala et al.'s (2019) analysis included both clinical and non-clinical populations as well as children and older adults.

    “However, their analysis included both clinical and non-clinical populations as well as children and older adults.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Sala et al. (2019) source

  • May combine two claims.

    Premisecitation

    Yakobi et al. (2021), in a meta-analysis of 27 randomized controlled trials examining mindfulness training in healthy adults, found limited improvements in attention and executive functions but no improvement in working memory.

    “Yakobi et al. (2021), in their meta-analysis of 27 randomized controlled trials examining the efficacy of mindfulness training in healthy adults, found limited improvements in attention and executive functions but no improvement in WM.”
    Quote in the paperNot a substantial claim0 of 1 sources checked
    • cites Yakobi et al. (2021) source

  • Classified as probably not a substantial claim. May combine two claims.

    Premisecitation

    Randomized controlled trials are widely regarded as the gold standard in clinical research because they help reduce selection bias and confounding factors, leading to a more precise determination of causality.

    “Randomized controlled trials are widely regarded as the gold standard in clinical research as they help reduce selection bias and confounding factors, leading to a more precise determination of causality (Chandler et al. 2019; Higgins et al. 2008).”
    Quote in the paperSubstantial claim0 of 2 sources checked
    • cites Chandler et al. 2019

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

    • cites Higgins et al. 2008

      Couldn’t be matched to a paper in Crossref, so it can’t be checked.

  • Classified as probably not a substantial claim.

    Premisecitation

    Sala et al. (2019) conducted a second-order meta-analysis to evaluate the effectiveness of various working memory training programs in producing near- and far-transfer effects.

    “Sala et al. (2019) conducted a second-order meta-analysis to evaluate the effectiveness of various working memory training programs in producing near- and far-transfer effects.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Sala et al. (2019) source

  • May combine two claims.

    Premisenone given

    Sala et al.'s (2019) analysis did not specify randomized controlled trials in their inclusion criteria, and so the analyses included quasi-experimental studies.

    “Furthermore, their analysis did not specify randomized controlled trials in their inclusion criteria, and so the analyses included quasi-experimental studies.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Sala et al. (2019) source

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    Only meta-analyses undertaken with healthy adults aged between 18 and 55 years were included in the report.

    “Only meta-analyses undertaken with healthy adults aged between 18 and 55 years were included in the report.”
    Quote in the paperSubstantial claim
    • Rests on an assumption: Including meta-analyses whose samples include some clinical participants, children, or older adults, as long as most participants are healthy adults aged 18–55, does not materially bias the second-order estimate for healthy adults aged 18–55.

  • Probably not a substantial claim: may not pass S3 (load-bearing).

    Study design and conductown data

    Two authors independently completed the screening process.

    “Two authors (MS and JJ) independently completed the screening process.”
    Quote in the paperNot a substantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing). May combine two claims.

    Study design and conductown data

    A total of 8250 records were found through a database search, and after screening, six studies were included in the final sample for analysis.

    “A total of 8250 records were found through a database search, and after screening, six studies were included in the final sample for analysis.”
    Quote in the paperNot a substantial claim

  • Classified as probably not a substantial claim.

    Study design and conductnone given

    The Comprehensive Meta-Analysis (CMA) software does not accommodate more advanced methods like Robust Variance Estimation (RVE).

    “We utilized the Comprehensive Meta-Analysis (CMA) software to manage nested effect sizes, but this software does not accommodate more advanced methods like Robust Variance Estimation (RVE).”
    Quote in the paperSubstantial claim

  • Classified as probably not a substantial claim.

    Study design and conductown data

    The findings were based on the overall effect sizes provided by each meta-analysis, which employed a random effects inverse variance model.

    “Our findings were based on the overall effect sizes provided by each meta-analysis, which employed a random effects inverse variance model.”
    Quote in the paperSubstantial claim

  • Probably not a substantial claim: may not pass S3 (load-bearing), S4 (based).

    Interpretationnone given

    The current study considered more recent studies that may have yielded improved training effects due to improved training approaches.

    “Additionally, our research considered more recent studies that may have yielded improved training effects due to improved training approaches.”
    Quote in the paperNot a substantial claim

  • May combine two claims.

    Study design and conductnone given

    The current study exclusively focused on randomized controlled trials, whereas Sala et al. (2019) included quasi-experimental studies.

    “our study exclusively focused on randomized controlled trials, whereas Sala et al. (2019) included quasi-experimental studies.”
    Quote in the paperSubstantial claim0 of 1 sources checked
    • cites Sala et al. (2019) source
22 of 22

  • The paper relies on this without support.

    The effect sizes extracted from the included first-order meta-analyses are comparable and can be validly averaged using a random effects model.

    “First, the effect sizes from each meta-analysis were averaged using a random effects model.”

  • The paper relies on this without support.

    Including meta-analyses whose samples include some clinical participants, children, or older adults, as long as most participants are healthy adults aged 18–55, does not materially bias the second-order estimate for healthy adults aged 18–55.

    “Studies where we noticed that the majority of the individual studies had healthy participants in the age range between 18 and 55 years were included in our analysis, while a minority of studies had participants who had clinical presentation and were over the age of 55 years or less than 18 years of age.”

  • The paper relies on this without support.

    Randomization of participants enhances validity, reduces the risk of selection bias, and helps account for confounding variables.

    “Randomization of participants enhances validity, reduces the risk of selection bias, and helps account for confounding variables”
3 of 3

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

Each contradiction between two passages listed here was confirmed by an independent check.

  • Included meta-analyses do not meet the stated age restriction

    Major

    The abstract says only meta-analyses of adults aged 18-55 were included, but the results say some included participants were under 18 and every study included participants over 55.

  • Excluded studies do not reconcile with the reported count

    Minor

    Of 25 studies with working memory as an outcome, excluding three unpublished studies leaves 22, not 23.

    Reported 23 · recomputed 22

  • Included meta-analyses do not meet the stated health restriction

    Minor

    The eligibility criterion requires healthy participants, but the results state that three included meta-analyses contained clinical samples.

3 of 3

Cites

Flagged first, then by citations

4 of the 75 references couldn’t be matched to a known paper.

Research works printed but not linked: 3

  1. Baddeley, Alan D. (1986). Working Memory, Oxford University Press.
  2. Chandler, Jacqueline, Cumpston, Miranda, Li, Tianjing, Page, Matthew, and Welch, Vivian (2019). Cochrane Handbook for Systematic Reviews of Interventions, Wiley.
  3. Higgins, Julian P. T., Thomas, James, Chandler, Jacqueline, Cumpston, Miranda, Li, Tianjing, Page, Matthew J., and Welch, Vivian A. (2008). Cochrane Handbook for Systematic Reviews of Interventions, John Wiley & Sons.

Cited by

Sorted by citations

Related

Authors
Syed, M., Lum, J. A. G., Byrne, L. K., & Skvarc, D.
Affiliations
Deakin University (AU)
Corresponding author
Syed, Maria
ORCID
Venue
Journal of Intelligence
Publisher
MDPI AG
Published
2024-11-12
Volume
12
Issue
11
Pages
114
ISSN
2079-3200 (ISSN-L)
DOI
10.3390/jintelligence12110114
PMC
PMC11595675
OpenAlex
W4404276653
Cited by
5
References
75
Citations by year
2026: 4 · 2025: 1
FWCI
2.14
Type
journal article
Language
en
Open access
Licence
Locations
  • Journal of Intelligence (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • PubMed Central (submittedVersion, cc-by)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
  • Figshare (submittedVersion, cc-by)
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Cognitive Abilities and Testing
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Cognitive Functions and Memory
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neural and Behavioral Psychology Studies
Keywords
working memory training, working memory capacity, healthy adults, mindfulness training, video game training, physical activity training, adaptive training, cognitive training intervention

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Claims read by deepseek-v4.1-flash under definition version 3 (extract-v3) on 2026-10-08.

Classified by jev-1.13-20260917 (gold-guideline-v0.2) on 2026-10-08.

Cited sources are checked by gpt-6-sol.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review