Retest reliability of repetitive transcranial magnetic stimulation over the healthy human motor cortex: a systematic review and meta-analysis

Abstract

Introduction: Repetitive transcranial magnetic stimulation (rTMS) is used to induce long-lasting changes (aftereffects) in cortical excitability, which are often measured via single-pulse TMS (spTMS) over the motor cortex eliciting motor-evoked potentials (MEPs). rTMS includes various protocols, such as theta-burst stimulation (TBS), paired associative stimulation (PAS), and continuous rTMS with a fixed frequency. Nevertheless, subsequent aftereffects of rTMS are variable and seem to fail repeatability. We aimed to summarize standard rTMS procedures regarding their test–retest reliability. Hereby, we considered influencing factors such as the methodological quality of experiments and publication bias. Methods: We conducted a literature search via PubMed in March 2023. The inclusion criteria were the application of rTMS, TBS, or PAS at least twice over the motor cortex of healthy subjects with measurements of MEPs via spTMS as a dependent variable. The exclusion criteria were measurements derived from the non-stimulated hemisphere, of non-hand muscles, and by electroencephalography only. We extracted test–retest reliability measures and aftereffects from the eligible studies. With the Rosenthal fail-safe N, funnel plot, and asymmetry test, we examined the publication bias and accounted for influential factors such as the methodological quality of experiments measured with a standardized checklist. Results: A total of 15 studies that investigated test–retest reliability of rTMS protocols in a total of 291 subjects were identified. Reliability measures, i.e., Pearson's r and intraclass correlation coefficient (ICC) applicable from nine studies, were mainly in the small to moderate range with two experiments indicating good reliability of 20 Hz rTMS (r = 0.543) and iTBS (r = 0.55). The aftereffects of rTMS procedures seem to follow the heuristics of respective inhibition or facilitation, depending on the protocols' frequency, and application pattern. There was no indication of publication bias and the influence of methodological quality or other factors on the reliability of rTMS. Conclusion: The reliability of rTMS appears to be in the small to moderate range overall. Due to a limited number of studies reporting test–retest reliability values and heterogeneity of dependent measures, we could not provide generalizable results. We could not identify any protocol as superior to the others.