Evaluating semantic control with transcranial magnetic stimulation: a systematic review with meta-analysis

Abstract

Background: This meta-analysis investigates the role of specific brain regions in semantic control processes using Transcranial Magnetic Stimulation (TMS). According to the Controlled Semantic Cognition framework, control processes help manage the contextually appropriate retrieval of semantic information by activating a distributed neural network, including the inferior frontal gyrus, the posterior middle temporal gyrus, and inferior parietal lobule. Lesions in these areas can lead to difficulties in manipulating weakly activated or competing semantic information. Researchers have used TMS to simulate such deficits in healthy individuals.

Method: By synthesizing results from TMS studies that targeted these regions, we aimed to evaluate whether neurostimulation over these areas can effectively impair participants' performance under high semantic control demands.

Results: Results from different meta-analytical approaches consistently showed no significant effects of TMS, especially after correcting for publication bias. Nevertheless, variability in experimental methodologies was evident.

Conclusion: These findings raise questions about the effectiveness of TMS in simulating deficits in semantic control and highlight the need for methodological improvements in future studies to enhance reliability and interpretability.

Keywords: controlled semantic cognition; semantic aphasia; semantic control; semantic representation; transcranial magnetic stimulation.

Figure 1

Flow chart of the selection process for the included studies. See the main text for the description of the exclusion criteria.

Figure 2

Risk of bias and effect sizes for the included effects. The traffic light plot on the left shows the risk of bias summary based on the authors’ judgments about each risk of bias item for each included study. Green, yellow, and red indicate low, unclear, and high risk of bias, respectively. See the main text for the description of the risk of bias dimensions. The forest plot on the right summarizes the meta-analysis results for all the included effects. Effects indicated in orange, purple, and green are derived from TMS stimulation of IFG, pMTG, and IPL, respectively (see inset); suffixes a and b indicate multiple effects from the same study and stimulated region. The bottom part of the forest plot shows the combined effect sizes (CES) derived from different meta-analytical approaches, as described in detail in the main text (3-level + RVE, three-level random effects model with a cluster-robust variance estimation method; T&F, trim-and-fill; RoBMA, robust Bayesian meta-analysis).

Figure 3

Trim-and-fill funnel plot. The funnel plot shows the effect sizes of the individual effects included in the meta-analysis as black dots. The empty dots represent the imputed and added effects after the trim-and-fill analysis.

Figure 4

Z-curve analysis. The figure shows the results of the Z-curve analysis of the included effects converted into z-scores. The histogram displays the distribution of observed z-scores from the 35 effects included, with 13 being statistically significant (z > 1.96, indicated by the red vertical line). The grey line represents the observed density of the significant z-scores, while the solid blue line depicts the fitted z-curve model.