Focused ultrasound for treatment-resistant depression may be moving beyond the idea of stimulating one brain region at a time. A new case study tested whether personalized ultrasound could sequentially reach several connected targets and temporarily change communication across a depression-related network.
The experiment did not improve the participant’s symptoms. Its more immediate contribution was technical: demonstrating that multiple deep and superficial brain targets could be selected through individual functional mapping, reached noninvasively, and evaluated for changes in network connectivity.
Depression Treatment Still Faces A Network Problem
Treatment-resistant depression is increasingly viewed as a disorder involving distributed circuits rather than dysfunction within one isolated brain region. Evidence from brain stimulation, lesion mapping, and neuroimaging has repeatedly implicated connections involving the subcallosal cingulate cortex, ventromedial prefrontal cortex, anterior cingulate cortex, and frontostriatal pathways.
Established interventions can influence these circuits, but each has limitations. Transcranial magnetic stimulation primarily reaches cortical targets, while deep brain stimulation requires surgery. Electroconvulsive therapy can be effective but produces broader brain effects and may carry cognitive side effects.
Most stimulation strategies also use a single anatomical entry point, even when the relevant illness involves several interacting regions.
How Focused Ultrasound For Treatment-Resistant Depression Was Personalized
Researchers studied a 39-year-old man who had previously received multiple medications, psychotherapy, repetitive TMS, electroconvulsive therapy, and ketamine without adequate improvement.
Two baseline resting-state functional MRI scans were used to create an individualized connectivity map. Starting with the subcallosal cingulate as the primary seed, the team identified six treatment volumes across three bilateral regions: the subcallosal cingulate cortex, ventromedial prefrontal cortex, and rostral anterior cingulate cortex.
These targets were selected because of their functional relationships within the participant’s own brain, not solely from standard anatomical coordinates.
A Controlled Test Of Multi-Target Neuromodulation
The participant received active and sham treatments four weeks apart under blinded conditions. During active treatment, a stereotactically fixed hemispherical array delivered low-intensity focused ultrasound sequentially to each personalized target.
The system incorporated MRI and CT coregistration, skull-correction modeling, electronic beam steering, and acoustic monitoring. These features were designed to improve targeting accuracy despite the way the skull can distort ultrasound energy.
The sham condition used randomized transmission phases to create an unfocused acoustic field while maintaining comparable energy passing through the scalp and skull.
Network Connectivity Changed Within One Hour
Approximately one hour after active treatment, average connectivity among the targeted regions decreased more than it did following sham treatment. The change from baseline was reported as Δr = −0.21, with a blockwise permutation value below 0.001.
Connectivity moved back toward baseline by 24 hours. This temporary pattern supports the interpretation that the intervention produced a reversible physiological effect rather than progressive scanner drift or a nonspecific session effect.
Supporting analyses also found reduced connectivity between the targeted network and distributed cortical areas, particularly regions overlapping the default mode network. Control regions in the visual cortex and postcentral gyrus did not show the same pattern.
Temporary Decoupling Without Clinical Improvement
The largest whole-brain connectivity change was associated with the ventromedial prefrontal cortex, followed by the rostral anterior cingulate cortex. Changes linked specifically to the subcallosal cingulate were more modest.
The participant experienced no clinical improvement across mood, anxiety, or quality-of-life measures. However, the study involved one active exposure and was designed to test feasibility and network engagement, not therapeutic efficacy.
No adverse events, structural MRI changes, sustained mood worsening, or hypomanic symptoms were reported. The participant also could not distinguish active treatment from sham.
A Platform For Testing Personalized Circuits
What separates this study from earlier ultrasound work is the combination of personalized functional mapping and sequential stimulation of multiple bilateral targets during one session.
Still, a single case cannot establish whether multi-target treatment is more effective than stimulating one region. The design also cannot determine whether the observed network changes resulted from cumulative multi-target stimulation or downstream effects initiated at one particularly influential node.
Larger studies will need to test repeated dosing, durability, clinical outcomes, and different target combinations. For now, the findings position focused ultrasound as a potentially precise tool for temporarily interrogating personalized depression networks and studying how individual circuit nodes interact.
Citations
Davidson B, et al. “Multi-target focused ultrasound neuromodulation of a subcallosal cingulate mediated depression network: A case study using personalized functional mapping.” Brain Stimulation. 2026. https://doi.org/10.1016/j.brs.2026.103198
Lynch CJ, et al. “Frontostriatal salience network expansion in individuals in depression.” Nature. 2024;633:624–633. https://doi.org/10.1038/s41586-024-07805-2
Explore more at https://www.interventionalpsychiatry.org/
