Can deep TMS help with addiction cravings?
In pooled data from 12 studies, deep TMS was followed by a large reduction in self-reported craving (standardized mean change -1.26; 95% CI, -1.67 to -0.86; p < .001). This is a systematic review and meta-analysis, not a single trial, and most of the included studies were not sham-controlled, so the pooled figure describes change from before to after real stimulation rather than a benefit measured against placebo.
Deep transcranial magnetic stimulation (dTMS) uses coils built to reach deep and bilateral brain regions, which is why there is interest in it for addiction. Substance use disorders and gambling disorder run a chronic course, and conventional treatments have limited efficacy, so neuromodulation aimed at the altered circuits behind these disorders has drawn attention. No definite recommendation for its use in addiction treatment exists yet.
What the data show:
- Craving reduction: Standardized mean change of -1.26 (95% CI, -1.67 to -0.86; p < .001) from pre to post real stimulation, a large effect
- Conditions covered: Substance use disorders and gambling disorder
- Outcome measured: Self-reported craving scores, the outcome used in the main analysis
- Evidence base: Seventeen articles in the systematic review, 12 in the meta-analysis
- Consistency: High heterogeneity across studies at both quantitative and qualitative levels
This systematic review and meta-analysis, published in Biological Psychiatry: Cognitive Neuroscience and Neuroimaging in 2025, followed PRISMA guidelines and screened four electronic databases through February 2024 for original English-language research articles.
Dr. Kumar’s Take
The signal here is large, and the idea behind it is sound: deep TMS is designed to reach circuits that sit below what surface stimulation touches, and craving lives in those circuits. What I hold back on is the design. Because only a minority of the included studies used a sham condition, the main analysis measures how far craving scores fell from before treatment to after it, and craving falls for a lot of reasons in someone who has entered treatment: withdrawal resolving, structure, contact with clinicians, expectation. The authors call this preliminary evidence, and I read it the same way. It justifies more rigorous trials, not a change in what I recommend to patients today.
What the Research Shows
The authors screened four electronic databases up to February 2024 and selected original English-language research articles on deep TMS in substance use disorders and gambling disorder. Seventeen articles entered the systematic review and 12 contributed to the meta-analysis.
Because only a minority of studies used a sham-controlled design, the main analysis computed the standardized mean change from pre to post real stimulation as the effect size, with self-reported craving as the outcome measure.
Results in Real Numbers
Pooled across the included studies, real deep TMS was associated with a significant and large reduction in craving scores: standardized mean change of -1.26, with a 95% confidence interval of -1.67 to -0.86 and p < .001.
Heterogeneity across studies was high at both the quantitative and the qualitative level. In plain terms, the studies differed from one another enough that the single pooled number smooths over real variation in how the treatment was delivered and in whom.
Who Benefits Most
The evidence base covers people with substance use disorders and with gambling disorder. Both are chronic conditions in which conventional treatment has limited efficacy, which is the gap deep TMS is being tested to fill.
Beyond that, this review does not settle who the best candidate is. The qualitative heterogeneity the authors describe means the included studies are not interchangeable, and I would not use this analysis to select one patient over another.
Safety, Limits, and Caveats
The central limitation is the design of the underlying literature. With only a minority of studies sham-controlled, the pooled effect cannot separate the effect of stimulation from the effect of everything else that happens when someone starts treatment.
The outcome is also self-reported craving, which is a subjective measure and is sensitive to expectation. Craving is a reasonable target, but it is not the same as abstinence or reduced use.
High heterogeneity across studies limits how confidently the pooled estimate transfers to any one protocol or patient. The authors frame their conclusion as preliminary evidence and call for further rigorous research to refine the therapeutic potential and to develop consensus-based guidelines for clinical application.
Practical Takeaways
- Treat deep TMS for addiction as promising and preliminary, not established
- Understand that the pooled effect reflects change from before to after real stimulation, not a comparison against sham
- The measured outcome is self-reported craving, not abstinence or substance use
- Both substance use disorders and gambling disorder are represented in the literature
- No consensus-based guidelines for clinical use exist yet
- Expect variation between protocols given the high heterogeneity across studies
Related Studies and Research
Episode 31: Depression Explained, The Biology Behind the Darkness
Episode 32: Depression Recovery Roadmap: A Step-by-Step, Evidence-Based Plan
Psychiatric Applications of Repetitive Transcranial Magnetic Stimulation
Default Mode Network Mechanisms of Transcranial Magnetic Stimulation in Depression
FAQs
What makes deep TMS different from standard TMS?
Deep TMS is a neuromodulation technique built to reach deep and bilateral brain regions, which is the rationale for testing it against the altered neural circuits that underlie addictive disorders.
Which conditions did this analysis cover?
Substance use disorders and gambling disorder, both of which the authors describe as addictive diseases with a chronic course.
How large was the craving reduction?
Pooled across 12 studies, the standardized mean change from pre to post real stimulation was -1.26 (95% CI, -1.67 to -0.86; p < .001), which the authors classify as a large effect.
Is deep TMS ready for routine addiction treatment?
Not yet. The authors describe the findings as preliminary evidence and call for further rigorous research and consensus-based guidelines before clinical application is settled.
Bottom Line
Across seventeen articles reviewed and 12 pooled, deep TMS was followed by a large drop in self-reported craving in substance use disorders and gambling disorder. Most of the studies lacked a sham condition and heterogeneity was high, so this is preliminary evidence supporting effectiveness rather than proof of it. It makes a strong case for rigorous controlled trials, which is exactly what the authors ask for.
