A trial says the expensive way to aim TMS is worth it. That reshapes how the treatment scales.

A JAMA Psychiatry randomized trial found that aiming accelerated TMS with individual brain imaging beat the standard scalp-landmark method, an 80 percent response rate against 60. It is the strongest evidence yet that the personalized approach earns its cost, which pushes the fastest-growing depression neuromodulation toward precision, expense, and the imaging-capable centers that can deliver it.

A randomized trial in JAMA Psychiatry has answered a question that determines how the fastest-growing form of depression neuromodulation scales: is it worth aiming accelerated transcranial magnetic stimulation with an individual brain scan, or is the cheap scalp-landmark method good enough? The answer, from investigators at Mass General Brigham, is that personalized, connectivity-based targeting produced significantly better outcomes, an 80 percent response rate against 60 percent for scalp-based targeting. That tilts the field toward precision, and toward the cost and infrastructure that precision requires.

Why targeting is the question

TMS, cleared for depression since 2008, stimulates the dorsolateral prefrontal cortex with magnetic pulses. Accelerated TMS, which delivers several sessions a day and condenses a multi-week course into about one week, is the field’s frontier, the lineage of the Stanford-developed protocol that won FDA clearance. The unresolved problem has been where to aim. The conventional method uses scalp landmarks, cheap, fast, and requiring no imaging, but it frequently misses the intended cortical site and varies considerably from one patient to the next. The alternative uses an individual’s resting-state functional MRI to locate the prefrontal site most connected to a deeper mood-regulating region, which is personalized but requires a scan and specialized analysis. The whole debate over personalized TMS has turned on whether that extra cost buys better outcomes.

What the trial found

The trial randomized patients with treatment-resistant depression to connectivity-based or scalp-based targeting for high-dose accelerated TMS. One month after treatment, the connectivity-based group showed significantly greater improvement on the standard depression rating scale and a higher response rate, 80 percent versus 60 percent. It is among the first proper randomized comparisons to show that imaging-guided personalization actually improves outcomes, rather than simply being more precise in theory. Notably, it is a head-to-head of two active targeting methods rather than a comparison against sham, which sidesteps the placebo and blinding problems that have plagued so much of the neuromodulation literature, and that makes the result cleaner than most.

Why it matters beyond the result

TMS is the most widely deployed neuromodulation modality in psychiatry, and accelerated protocols are scaling quickly, which makes the targeting method the fork in how the whole field grows. Scalp-based targeting is cheap, fast, needs no imaging, and can be delivered almost anywhere. Connectivity-based targeting requires an individual functional MRI and expert analysis, which is expensive, specialized, and concentrated at imaging-capable centers. By giving the precision approach its first strong efficacy evidence, this trial pushes the field, and eventually payers and clinical guidelines, toward expecting imaging. That raises per-patient cost, narrows where accelerated TMS can be delivered well, and advantages the commercial players built around precision protocols. The trade is better outcomes for higher cost and less accessibility, and this result moves the field toward paying it.

The caveats

Several are substantial. The senior authors hold intellectual property on connectivity-based TMS targeting and have commercial ties to the companies advancing precision TMS, so a trial showing the personalized method wins, conducted by people with a stake in the personalized method, warrants scrutiny even with their statement that the specific intellectual property was not used in the study; independent replication matters more than usual here. The mechanism is also unresolved: related work from the same group found that symptom improvement did not correlate with how much the targeted circuit’s connectivity actually changed, so the personalized approach works without a clear account of why, which is a loose end beneath the headline. And this is a single trial, with the full effect size, blinding rigor, and durability still to be read in the complete paper and confirmed in replication. The trial demonstrates an efficacy edge, not cost-effectiveness, which is the question payers will ultimately ask: whether the scan is worth it per outcome gained.

The frame

Accelerated TMS is the neuromodulation modality genuinely scaling in depression, and this trial answers the question that governs how it scales. Personalization works, which means the field is heading toward imaging-guided precision and the cost and concentration that come with it. Unlike the sham-response problem that has dogged vagus nerve and deep brain stimulation, this is a mature, head-to-head comparison of two delivery methods, a sign that TMS sits further along the deployment curve than the rest of the neuromodulation category. The caveats, the author conflicts, the unexplained mechanism, and the single-trial status, are real and worth holding. But the direction is clear: the first strong randomized evidence on the question votes for paying for precision. The consequence is a version of TMS that needs an MRI and an expert to deliver well, which is better medicine and less accessible medicine at once, and the tension between those two is the thing to watch as accelerated TMS expands.