Repetitive transcranial magnetic stimulation shows promise for improving cognition and executive function in patients with mild cognitive impairment

BackgroundTranscranial magnetic stimulation (TMS) has been explored as a non-invasive intervention for cognitive impairment in mild cognitive impairment (MCI), but evidence regarding its overall efficacy remains inconsistent. This meta-analysis evaluated the pooled effect of TMS on cognitive outcomes in patients with MCI.MethodsRandomized controlled trials (RCTs) enrolling participants with MCI only were identified from PubMed, CENTRAL, and Embase through November 10, 2025, in accordance with PRISMA and a PROSPERO-registered protocol (CRD420251274531). Cognitive outcomes were pre-specified and synthesized by domain, including global cognition and executive function (primary focus), as well as memory, language/naming, and neuropsychiatric measures. Pooled effects were calculated using fixed- or random-effects models as appropriate and reported as mean differences (MD) or standardized mean differences (SMDs) calculated as Hedges'g with 95% confidence intervals. Risk of bias, methodological quality, and certainty of evidence were assessed using RoB 1, PEDro, and GRADE.ResultsEleven RCTs (532 participants with MCI) were included. rTMS improved global cognition on MMSE (MD = 2.35, 95% CI 1.74–2.96; k = 5, n = 120) and MoCA (MD = 1.31, 95% CI 0.80–1.82; k = 6, n = 337). Verbal memory also improved in AVLT-based outcomes (delayed recall: k = 5, n = 125; recognition: k = 4, n = 103). Attention/executive measures favored rTMS, including Digit Span (MD = 0.62, 95% CI 0.37–0.87) and shorter completion times on Trail Making Test-A (MD = −3.33, 95% CI −3.76 to −2.90) and TMT-B (MD = −4.08, 95% CI −7.92 to −2.23). Naming ability and neuropsychiatric outcomes also improved, whereas language fluency showed no significant between-group difference. Overall risk of bias was low; however, due to limited sample sizes, GRADE certainty was mostly moderate.ConclusionIn MCI-only RCTs, rTMS improved global cognition (MMSE/MoCA) and AVLT-based verbal memory (delayed recall/recognition). Selected executive/attention, naming, and neuropsychiatric outcomes also improved, whereas language fluency was unchanged. Evidence certainty was mostly moderate, and larger high-quality RCTs are needed.Systematic review registrationhttps://www.crd.york.ac.uk/PROSPERO/view/CRD420251274531, identifier: CRD420251274531.