The field keeps treating neuromodulation as a last resort. In the clinic, that assumption costs patients years they don’t have. Every week I see treatment-resistant depression cases where we’ve cycled through multiple antidepressants, augmentation strategies, esketamine sessions, and still the amygdala keeps firing at the wrong amplitude, at the wrong time, in response to stimuli that shouldn’t register as threats. The pharmacological toolkit gives us ways to dampen that signal systemically. What it doesn’t give us is a way to reach in, with millimeter precision, and modulate the specific node driving the pathology. That gap is exactly what a first-in-human trial of amygdala-targeted low-intensity focused ultrasound just addressed, and the data, even from ten participants, deserves more attention than the field has given it.

Ten participants is not a power-calculation. But first-in-human trials earn their place in the literature on a different metric: do they demonstrate target engagement, tolerability, and a signal worth chasing? This one appears to. The trial used a single-blind randomized crossover design, imaging-guided FUS delivered in two 10-minute applications to the right amygdala, with the left primary somatosensory cortex as an active control. That control selection alone reflects careful thinking, it rules out nonspecific arousal effects while holding constant the acoustic experience of receiving ultrasound. The team stayed within diagnostic ultrasound safety limits throughout. For a modality that has drawn skepticism precisely because prior noninvasive approaches couldn’t reach deep structures with this kind of anatomical specificity, that design choice matters.

Why the Amygdala, Why Now

The amygdala has been a compelling theoretical target in affective neurocircuitry for decades. What’s changed is our ability to reach it without a burr hole. The FDA cleared MRI-guided focused ultrasound thalamotomy for medication-refractory essential tremor back in 2016, using InSightec’s Exablate Neuro platform, but that application uses ablative, high-intensity energy, a fundamentally different proposition from the low-intensity neuromodulatory approach tested here. Low-intensity FUS doesn’t destroy tissue; it transiently alters neuronal excitability through mechanical effects on ion channels and membrane dynamics. The distinction matters enormously for a psychiatric application, where you’re not treating a movement disorder with a localized, relatively stable lesion target, but a distributed affective circuit where you want modulation, not ablation.

Research published in Psychological Medicine in 2024 reinforced what fMRI work has been building toward for years: amygdala reactivity to emotional stimuli predicts treatment response in difficult-to-treat depression. That line of evidence gives the FUS targeting rationale a mechanistic backbone. In my experience treating many patients with ketamine and esketamine, rapid-acting interventions that work often produce a recognizable shift, patients report emotional stimuli feeling less adhesive, less capturing. Whether that reflects amygdala modulation, prefrontal disinhibition, or both simultaneously, we can argue. But the circuit is the same one low-intensity FUS is now being aimed at deliberately, with spatial precision that pharmacology can’t match.

Where this sits relative to esketamine, approved by the FDA in March 2019 as the first novel-mechanism antidepressant since fluoxetine in 1987, is worth thinking through carefully. Esketamine’s glutamatergic mechanism produces rapid symptom relief in hours, but the durability question remains genuinely open for many patients, and the REMS requirements around observed administration create real access friction. COMPASS Pathways’ COMP360 psilocybin data, published in the New England Journal of Medicine in 2022 across 233 patients with TRD, showed a dose-dependent signal at 25mg but also a durability ceiling that the field is still working to extend. Low-intensity FUS occupies a different mechanistic lane entirely, no psychoactive experience, no systemic drug exposure, no dissociation risk, which means it could reach patient populations that pharmacological rapid-acting approaches exclude or poorly serve.

The Design Questions That Come Next

Here’s where the trial raises as many questions as it answers, and where the field needs to be honest about what it’s measuring. Crossover designs in psychiatric populations carry a particular methodological tension: carryover effects are real, especially when the intervention produces measurable neurophysiological changes that may outlast a single session. A right amygdala that has been modulated by ultrasound on Day 1 may not return to baseline by the time the participant receives the somatosensory control on Day 14. The investigators likely accounted for this, but as trials scale, washout period selection for a modality whose duration of effect is itself a primary unknown will require explicit justification and probably dedicated characterization studies. The endpoint architecture also needs to evolve: subjective symptom scales capture what patients report, but target engagement in a structure as functionally specific as the amygdala calls for neuroimaging readouts, biomarker-anchored response definitions, and ideally a mechanism-matched endpoint that tracks what the amygdala is actually doing after treatment, not just what the patient’s PHQ-9 says a week later.

The DARPA SUBNETS program, launched in August 2014 with $70 million over five years to develop closed-loop neuromodulation for psychiatric illness, invested heavily in the conceptual architecture that precision neuromodulation requires: real-time biomarker feedback, circuit-level targeting, adaptive stimulation. Low-intensity FUS doesn’t yet close that loop, but the amygdala targeting rationale fits squarely within that framework. The next generation of trials in this space will need to decide whether they’re testing FUS as a standalone acute intervention, as a circuit primer before pharmacotherapy or psychotherapy, or as a maintenance modality, because those are three different trial designs, three different endpoints, and three different regulatory pathways.

For the populations I treat across multiple states, including patients with dual diagnosis and those who’ve failed multiple prior interventions, a noninvasive, non-psychoactive, anatomically precise neuromodulation option would fill a real gap. Patients with active SUD or unstable cardiovascular profiles are often excluded from esketamine protocols by label or by clinical judgment. FUS carries no pharmacokinetic interactions, no abuse potential, no dissociation. The exclusion criteria in future trials should reflect that advantage rather than default to the usual TRD enrollment profile that systematically filters out the most complex, highest-burden patients.

References

  1. medRxiv Psychiatry, “Low-Intensity Focused Ultrasound of the Amygdala in Depression and Anxiety: A First-in-Human Active-Controlled Trial”
  2. Focused Ultrasound and Deep Brain Stimulation Foundation, FDA-Cleared Focused Ultrasound Conditions Treated
  3. Psychiatric News / APA, “FDA Approves Esketamine Nasal Spray for Treatment-Resistant Depression” (2019)
  4. COMPASS Pathways, Phase 2b COMP360 Psilocybin Trial Results, New England Journal of Medicine (2022)
  5. Psychological Medicine, “Neural responses to facial emotions and subsequent clinical outcomes in difficult-to-treat depression” (2024)
  6. Neurology Today / AAN, DARPA SUBNETS Program, $70 Million Neuromodulation Initiative (2014)
+ posts

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