Stanford’s five-day protocol once posted remission rates near 80%, so why did it fail here?
STUDY: Appelbaum LG et al, Journal of Affective Disorders 2026;415:122515
STUDY TYPE: Randomized, sham-controlled trial
FUNDING: Wellcome Leap. One author advises Brainsway and has received in-kind support from MagVenture.
Background
Stanford Neuromodulation Therapy, FDA-cleared as SNT (formerly SAINT) in 2022, brought an unprecedented 79% of depressed patients to remission even with high levels of treatment resistance (an average of five failed antidepressants).
To improve accuracy, SNT uses fMRI to guide the placement of the magnet, but accuracy might be further enhanced by adding EEG-guidance to the fMRI procedure. This trial aimed to test that approach against traditional SNT, but arrived at a surprise. Neither worked better than placebo (sham).
Negative studies are common, and often due to an exaggerated placebo response (a failed study). But this setback comes alongside news of a larger trial (n=130) that found a dampened and delayed response to SNT, due to be published soon by Meiling Li and colleagues in Lancet.
The Study
- 51 adults with treatment-resistant major depressive disorder or bipolar II depression were randomized at one site, and 43 completed the protocol.
- Three arms: 1) fMRI-guided iTBS, 2) fMRI-guided iTBS with EEG-tuned frequency, 3) sham (fake iTBS).
- All received 50 sessions over five days, with follow-up at 1 and 4 weeks.
- The primary outcome was change on the Depression scale (MADRS) at 1 week.
Results
All groups improved by about 7 points. Sham fell 27%, compared with 22% for fMRI alone and 34% with EEG tuning. Differences between groups were small (Cohen’s d 0.08–0.47) and could have arisen by chance.
- Response: 24% fMRI, 33% fMRI plus EEG, 25% sham
- Remission: 24%, 22%, and 13%
- Anhedonia scale (DARS): no difference
Results were similar at 4 weeks.
Limitations
- With only 14 patients per arm, the trial may have been underpowered to detect a difference, but the original SNT trial was similar in size
- There was a mid-study change in the fMRI mapping method
Practice Implications
- These new trials don’t mean that SNT doesn’t work, but suggest its benefits are not as profound as expected. Expect a response rate closer to 1 in 4 in a typical caseload.
- Better results are likely for depressions that had a clear onset, separated by past periods of full recovery, resembling those in the positive trials.
- Find SNT providers through Magnus Medical.
—Chris Aiken, MD
Director, Psych Partners
Editor in Chief, Carlat Psychiatry Report
