MUPLAB
Bayesian Living Meta-Analysis · v1.0

The effect of auricular taVNS on learning

The effect of taVNS on accuracy and response time (RT) across declarative and working memory, cognitive control, and associative & skill acquisition. Bayesian random-effects (brms · bayesmeta); ACC and RT pooled separately, positive g favors taVNS.

Registration: PROSPEROData: OSFCode: brms · bayesmetaUpdated: 23 June 2026
v1.0 — registered. 24 studies, 36 primary effect sizes (22 ACC + 14 RT). Headline values are the registered brms model; in the interactive panel below you can include/exclude any study and run the analysis on your chosen subset (in-browser, not saved).
Model & priors
  • Random-effects normal-normal; priors: effect μ ~ N(0, 0.5), τ ~ Half-Normal(0.3).
  • ACC and RT analyzed separately; RT is ×(−1) so positive favors taVNS.
  • Backbone cluster-robust RVE; the Bayesian version adds robustness + P(g>0) + prediction interval.

Primary pooled (registered, brms)

Accuracy

+0.15
95% CrI [0.06, 0.25]
P(g>0) = 99.9% · τ = 0.11 · PI [-0.13, 0.44] · k = 22

Response time (RT)

+0.18
95% CrI [0.02, 0.37]
P(g>0) = 98.3% · τ = 0.24 · PI [-0.37, 0.76] · k = 14
Key finding. The strongest effect is on response time in associative & skill learning: RT × Associative & Skill g = 0.62 [0.36, 0.88] (fragile, k = 4). Overall ACC g = 0.15, RT g = 0.18. ρ(ACC,RT) = 0.14 [−0.73, 0.86].

Interactive explorer

Pick an outcome; from the Studies tab include/exclude any publication — the forest, posterior, and all statistics recompute instantly for your chosen subset.

Registered analysis (locked reference) — PROSPERO-registered corpus · brms · bayesmeta · metafor. The figures here are the registered R output and do not change; they are independent of the include/exclude choices in the tabs below. Switching ACC/RT shows the corresponding registered result.
—
Pooled g
—
95% CrI
—
P(g > 0)
—
Prediction interval
—
τ
—
k

Subgroup forest — registered

Registered subgroup forest

Contour-enhanced funnel — registered

Registered funnel

Normal Q–Q — registered

Registered Q–Q
Selected in this pool: —
—
Pooled g
—
95% CrI
—
P(g > 0)
—
Prediction interval
—
τ
—
k

Forest plot + Bayesian pooled

Each row is an effect size (g, 95% CI). Diamond = posterior pooled (95% CrI), dashed = prediction interval. Right favors taVNS, left favors sham.

Posterior distribution of the pooled effect (μ)

Shaded band is the 95% CrI; dashed line is g=0. P(g>0) = mass right of zero.

Move the priors → sensitivity.

Unchecking a study removes it from the pool; results update instantly. Extinction (S29c, S29d) is excluded by default.

Filter:
Selected (of total): —
IDStudyDomainOutcomeDesigngSE

Add a study

Enter g and SE, pick domain + outcome, Add. It joins the selected pool.

Selected pool — by domain (live)

DomainkPooled g95% CrIP(g>0)

Domain subgroups (registered, bayesmeta)

OutcomeDomainkPooled g95% CrI
AccuracyDeclarative Memory5+0.09[0.01, 0.17]
AccuracyWorking Memory6+0.27[0.15, 0.38]
AccuracyCognitive Control3+0.20[-0.05, 0.45]
AccuracyAssociative & Skill Acquisition8+0.11[-0.14, 0.37]
Response time (RT)Working Memory5-0.06[-0.21, 0.10]
Response time (RT)Cognitive Control5+0.22[0.03, 0.41]
Response time (RT)Associative & Skill Acquisition4+0.62[0.36, 0.88]

This tab shows live versions of our three registered figures (subgroup forest · contour funnel · Q–Q) for the selected subset. They update when you add/remove studies or press “Re-run”.

Subgroup forest
Live counterpart of the registered subgroup forest: studies grouped by domain (Declarative · Working · Cognitive Control · Associative & Skill), square size = weight (1/SE²), diamonds = subgroup and overall pools; dashed line = overall mean.
Funnel
Effect size (x) vs standard error (y). Dashed vertical line = pooled mean; shaded funnels = 95% and 99% pseudo-confidence regions. Marked asymmetry may indicate publication bias.
Q–Q (normality)
Random-effects standardized residuals vs normal quantiles. Points near the dashed line support the normality assumption.
Registered extra diagnostics — leave-one-out & meta-regression (fixed)

Registered R/bayesmeta outputs; ACC and RT shown separately.

Leave-one-out (caterpillar) — Accuracy
Leave-one-out (caterpillar) — Accuracy
Leave-one-out (caterpillar) — RT
Leave-one-out (caterpillar) — RT
Domain-level leave-one-out — Accuracy
Domain-level leave-one-out — Accuracy
Domain-level leave-one-out — RT
Domain-level leave-one-out — RT
Meta-regression (bubble) — Accuracy
Meta-regression (bubble) — Accuracy
Meta-regression (bubble) — RT
Meta-regression (bubble) — RT

Characteristics of included studies (paper level, 22 studies). Multi-experiment papers (Sun 2021, Ventura-Bort 2025) contribute two experiments/studies; the analysis counts 24 study units. Domain = 4-category; Outcome = ACC/RT.

StudyYearCountryDesignBlindingNAge (M)%FDomainOutcome
Jongkees 20182018Netherlands/Australia/GermanyBetweenSingle402280Cognitive ControlRT,ACC
Giraudier 20202020GermanyBetweenSingle60Declarative MemoryACC
Kühnel 20202020GermanyWithinSingle392659Associative & SkillACC
Mertens 20202020BelgiumWithinSingle412251Declarative MemoryACC
Thakkar 20202020USABetweenSingle372173Associative & SkillRT,ACC
D'Agostini 20212021BelgiumBetweenSingle712377Associative & SkillACC
Kaan 20212021USABetweenSingle622066Working MemoryRT,ACC
Sun 20212021ChinaWithinSingle46Working MemoryRT,ACC
Phillips 20222022USABetweenDouble452264Associative & SkillRT,ACC
Zhao 20222022ChinaWithin—632152Working MemoryRT,ACC
Konjusha 20232023GermanyWithinSingle3725Working MemoryACC
Sommer 20232023GermanyWithinSingle322659Cognitive ControlRT
Tian 20232023ChinaWithinSingle93Working MemoryRT,ACC
Bömmer 20242024GermanyWithinDouble272548Cognitive ControlRT,ACC
Chen 20242024ChinaWithinSingle222346Associative & SkillRT
Honda 20242024CanadaBetweenDouble452373Associative & SkillACC
Li 20252025ChinaWithinSingle61Cognitive ControlRT
Sönmez 20252025GermanyWithinSingle292551Cognitive ControlRT,ACC
Thakkar 20252025USAWithinSingle3520Associative & SkillACC
Ventura-Bort 2025 S12025GermanyWithinSingle302187Declarative MemoryACC
Çakır 20252025TürkiyeBetweenSingle80Associative & SkillRT,ACC
Mary 20262026BelgiumWithinSingle892452Declarative MemoryACC

Version history

v1.0June 2026

Registered corpus: 24 studies, 36 primary ES. brms + bayesmeta.

In production the model runs in R / brms · bayesmeta and is embedded as a Shiny app. The in-browser engine here matches the registered model within ~0.01.