Music exposure reduces anxiety- and depression-like behavior in rodents: a systematic review and multilevel meta-analysis

Reproducible workflow, full outputs, and extended methods

What this site contains

This online supplement provides the complete computational workflow for the systematic review and multilevel meta-analysis of music exposure effects on anxiety- and depression-like behaviour in laboratory rodents.

It includes:

  • Full data processing and effect-size derivations (lnRR, lnVR, and lnCVR), including safeguards for boundary values and audit checks

  • Multilevel model specifications (VCV construction, random-effects structure) and overall-effect models for each effect size

  • Uni-moderator meta-regressions across study and design characteristics, with post-hoc contrasts where applicable

  • Publication-bias diagnostics and leave-one-out sensitivity analyses

  • Figures and downloadable tables generated directly from the analysis pipeline

How to navigate

Use the left sidebar to browse the workflow chapters:

  • Home (index.qmd)

  • Effect size calculation (effect_size_calculation.qmd) — data cleaning, continuity corrections, and computation of lnRR, lnVR, and lnCVR

  • Overall effects (overall_effect.qmd) — intercept-only multilevel models for each effect size

  • Uni-moderators (uni_moderators.qmd) — uni-moderator meta-regressions and post-hoc contrasts

  • Publication bias (publication_bias.qmd) — small-study effects / bias diagnostics

  • Leave-one-out (leave_one_out.qmd) — sensitivity analyses excluding one study at a time

  • Main figures (main_figures.qmd) — reproduction of figures from the manuscript

  • Alluvial (Alluvial.qmd) — alluvial visualization of study characteristics and outcome structure

Reproducibility

All code is executed in Quarto and frozen for stable rendering. Where relevant, we provide intermediate outputs (effect-size tables, continuity-correction audits, VCV matrices, model summaries, and contrasts) alongside the exact R code used to generate them.