A patient I’ve been treating for three years — severe, recurrent MDD, failed four antidepressants, failed esketamine, partial responder to ketamine infusions — finally agreed to a convulsive therapy referral last month. Her condition: not ECT. She’d watched her mother lose months of memory after a course of treatments in the 1990s, and no amount of updated evidence about right unilateral ultra-brief pulse techniques moved her. She’d rather stay depressed than risk her memory. That clinical standoff, repeated across thousands of practices, is exactly the treatment gap the CREST–MST trial was designed to close.

Published this month in Lancet Psychiatry, the CREST–MST randomized, double-blind, non-inferiority trial compared magnetic seizure therapy against right unilateral ultra-brief electroconvulsive therapy — the current cognitive-safety optimized standard — in patients with major depressive disorder across sites in Canada and the United States. The headline finding: MST achieved non-inferior remission rates relative to RUL-UB ECT, with a more favorable cognitive safety profile. For the field of convulsive therapy, this is not an incremental update.

Why Cognitive Safety Is the Enrollment Problem

The counterintuitive assumption baked into most ECT trial designs is that cognitive side effects are a tolerability issue — something to manage, disclose, and monitor. The CREST–MST data reframe it as a recruitment barrier and a refusal driver. ECT has long carried remission rates that any antidepressant sponsor would envy, routinely landing in the 50–70% range in adequately powered trials. The cognitive cost has been the ceiling on its adoption, not a secondary concern.

MST uses transcranial magnetic stimulation technology to induce seizures, but with significantly more focal cortical targeting than conventional ECT. Because the induced electrical field is concentrated at the cortical surface rather than distributed through deep brain structures — including the hippocampus, which is central to declarative memory consolidation — the mechanistic basis for reduced autobiographical memory disruption is real, not theoretical. The CREST–MST results appear to confirm that mechanistic hypothesis at scale.

That logic holds until you look at where the field has been stuck. Repetitive TMS, which operates on similar magnetic stimulation principles without seizure induction, demonstrated efficacy in treatment-resistant depression and gained FDA clearance, but its remission rates have consistently trailed ECT. The MST data suggest the seizure component remains therapeutically necessary — what MST changes is the precision of where that seizure is initiated and propagated.

What This Means for the Next Convulsive Therapy Trial

Sponsors designing trials in treatment-resistant depression have operated under a binary: pharmacological approaches that avoid cognitive risk but carry modest effect sizes, or neuromodulation approaches with larger effect sizes and real cognitive liability. The CREST–MST readout introduces a third position — a convulsive therapy with ECT-level efficacy and a cognitive safety profile that should materially change patient willingness to enroll.

Patient refusal is the enrollment problem that site feasibility assessments consistently underestimate.

In my practice, the patients who refuse ECT are not uninformed — they are often the most treatment-literate patients in any TRD population, having exhausted alternatives methodically. They have read the literature. They know RUL-UB ECT is gentler than bilateral sine-wave ECT. They still say no. That subpopulation, systematically excluded from every ECT efficacy trial by their own refusal, represents a substantial fraction of the severe TRD universe — and MST is specifically positioned to reach them. The CREST–MST investigators made this explicit, noting MST’s consideration as a first-line convulsive therapy particularly for patients who refuse RUL-UB ECT.

For any sponsor or CRO evaluating a neuromodulation trial in MDD, the protocol design implication is direct: MST-eligible populations will require enrollment criteria that explicitly capture prior ECT refusal as an inclusion pathway, not just prior ECT failure. Those are different patients, with different baseline cognitive profiles, different consent conversations, and — critically — different retention dynamics once enrolled. Retention in cognitively demanding longitudinal trials tends to track inversely with treatment-emergent cognitive complaints; if MST’s cognitive safety advantage holds in real-world site conditions the way it held in CREST–MST, dropout rates should follow.

Sites with established neuromodulation infrastructure and active TRD populations who have declined ECT referrals are positioned to move quickly on MST trials. The bilingual dimension matters here too — Spanish-speaking patients with severe MDD are systematically underrepresented in neuromodulation research, and cognitive side effect fears, amplified by cultural and familial stigma around memory loss, make ECT refusal rates in that population even higher than the general TRD average.

References

  1. Lancet Psychiatry — “Confirmatory efficacy and safety trial of magnetic seizure therapy versus right unilateral ultra-brief electroconvulsive therapy in depression (CREST–MST): a randomised, double-blind, non-inferiority trial in Canada and the USA” (April 2026)
  2. Journal of Affective Disorders — Vando et al., ketamine treatment outcomes in mood disorders (published prior to 2026)
+ posts

Dr. Leonardo Vando, MD is Principal Investigator of Psychiatry at Vando Medical Services.