Symptom Provocation and Clinical Response to Transcranial Magnetic Stimulation: A Systematic Review and Meta-Analysis

Abstract

Importance: Transcranial magnetic stimulation (TMS), a form of noninvasive brain stimulation used to treat major depressive disorder, obsessive-compulsive disorder (OCD), and nicotine dependence, has well-established state-dependent effects on brain circuitry. However, a major question for TMS remains: does brain state affect clinical response?

Objective: To quantify the association between symptom provocation and clinical response to TMS for OCD and nicotine dependence, the only Food and Drug Administration-cleared TMS indications with symptom provocation.

Data sources: PubMed, CINAHL, Embase, PsycInfo until August 30, 2024.

Study selection: Randomized clinical trials of TMS for OCD or nicotine dependence with a clinical outcome. Of 600 studies identified, 71 met inclusion criteria.

Data extraction and synthesis: Data extraction was completed independently by 2 extractors and cross-checked by a third. Standardized mean difference (SMD) and SE were estimated via Hedges g and synthesized data in a 3-level random-effects meta-analysis. Study data were analyzed from August 2023 to March 2025.

Main outcomes and measures: Primary outcomes were clinical response measures.

Results: A total of 71 studies met inclusion criteria and included 3246 participants (mean [SD] age; 37.8 [8.0] years; mean [SD] percentage female, 44.1% [17.2%]). Included in the meta-analysis were 63 studies with 2998 participants. For OCD studies, active TMS was associated with better clinical response than sham both with (SMD = -0.51; 95% CI, -0.96 to -0.07; P = 0.04) and without (SMD = -0.29; 95% CI, -0.40 to -0.17; P < .001) symptom provocation. For nicotine use, active TMS was associated with better clinical response than sham when used with (SMD = -0.56; 95% CI, -1.12 to 0; P = .05) but not without (SMD = -0.35; 95% CI, -0.74 to 0.04; P = .08) symptom provocation. For OCD studies, the estimated expected added effect of provocation was SMD of -0.22 (95% CI, -0.65 to 0.20; P = .22). In nicotine studies, the estimated expected added effect of provocation was SMD of -0.21 (95% CI, -1.00 to 0.58; P = .57).

Conclusions and relevance: Results of this systematic review and meta-analysis suggest that symptom provocation may enhance clinical response to TMS for OCD and nicotine dependence. Studies comparing TMS with and without provocation are critical to establish the causal effect of provocation.

Figure 1.. Model of Hypothesized Effects of Symptom Provocation on Clinical Response to Transcranial Magnetic Stimulation (TMS)

Repetitive TMS (rTMS) has known state-dependent effects. The current US Food and Drug Administration–cleared TMS protocols for nicotine dependence and obsessive-compulsive disorder (OCD) both include provocation of a symptomatic state (craving, obsessive-compulsive thoughts, respectively) in an effort to improve TMS response. However, these experiments have been performed in small numbers, and the overall effect of symptom provocation has not been quantified. Figure created with BioRender.com.

Figure 2.. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Flow Diagram

Shown here is the process of identification, review, and selection of articles included in the systematic review.

Figure 3.. Systematic Review Results for Obsessive-Compulsive Disorder (OCD) and Nicotine Studies (N = 71)

Displayed here are study characteristics (use of provocation, transcranial magnetic stimulation [TMS] coil, stimulation protocol) for studies of OCD (n = 44) (A) and nicotine dependence (n = 27) (B) included in the systematic review. The “other” category for TMS protocol includes 1 study with 6 Hz then 1 Hz and 1 study with alpha TMS: 8 to 12 Hz.

Figure 4.. Forest Plot of Meta-Analysis Results for Obsessive-Compulsive Disorder (OCD) and Nicotine Studies

Shown here are standardized mean difference and 95% CI for all studies included in meta-analysis (N = 63). Overall estimated effect size indicates reduction in symptoms, with more negative values indicating more effective treatment (pooled estimate shown with 95% CI). The effect size for each study is a weighted average of the effect sizes within each study assuming a compound symmetric correlation structure (eMethods in Supplement 1). The size of points indicates their relative weight in the meta-analysis model.