Penn Phase 1 Trial Tests Focused Ultrasound in Cocaine Use Disorder
The University of Pennsylvania is preparing a Phase 1 trial of low-intensity focused ultrasound for adults with cocaine use disorder, a study designed to answer a narrow but important question: can a noninvasive brain stimulation device be delivered safely and feasibly to the ventral striatum, and does it produce measurable changes in neural activity?
It is an early-stage experiment, not a treatment claim. The protocol is explicit about that boundary.
Twenty participants will be enrolled in Philadelphia, with half randomized to active stimulation and half to sham. The work sits inside the HEAL Initiative and will run as a double-blind randomized study in which participants and the outcome assessors remain masked. Only the team member operating the stimulation session will know which condition a participant receives.
Why the Ventral Striatum Matters in Cocaine Use Disorder
The ventral striatum sits at the center of reward processing, cue reactivity, and reinforcement learning. In cocaine use disorder, that circuitry becomes clinically relevant every time drug-related cues provoke craving, urgency, or compulsive seeking. A device aimed at this region is trying to influence the neural machinery that helps drive relapse, not merely suppress symptoms at the surface.
The investigators are not asking whether the device reduces cocaine use. This trial does not test abstinence, relapse prevention, or long-term substance use outcomes. Instead, it is built around three immediate questions: whether the stimulation can be delivered safely, whether participants can complete the protocol, and whether brain activity shifts after sonication.
A Small, Carefully Controlled First Look
The study uses a parallel assignment design with 10 participants in the active arm and 10 participants in the sham arm. A single-element ultrasound transducer is positioned over the head and aligned to the target region with a neuronavigation system linked to reflective head markers. Participants wear headphones playing white noise throughout the session so that auditory clues do not reveal whether they are receiving active treatment or sham stimulation.
That masking strategy matters because procedural blinding is especially difficult in neuromodulation studies. If participants can guess their assignment from sound, sensation, or operator behavior, the comparison becomes harder to trust. Here, the protocol attempts to reduce that risk by equalizing the external experience across both groups.
The intervention itself remains investigational. Regulatory reviewers judged the device to pose non-significant risk, but it is not FDA-approved for any indication. The protocol also specifies bilateral application to the same brain target in both active and sham conditions.
Safety Is the First Endpoint for a Reason
The primary outcome measures begin with adverse events. Investigators will track the number of participants experiencing any adverse event, graded by CTCAE severity and by relationship to the intervention, as well as the number with serious adverse events over a three-week period. For a first-in-time neuromodulation study in this population, those are the right questions to ask first.
Craving circuitry may be a promising target, but the ventral striatum is not a casual place to apply a novel device. The protocol reflects that reality by building in MRI-based screening, individualized acoustic modeling, and a series of eligibility filters intended to reduce neurological and psychiatric risk.
Participants must be 18 to 65 years old and meet DSM-5 criteria for moderate or severe cocaine use disorder. The study requires at least $100 worth of cocaine use in the past 30 days, written informed consent in English, willingness to attend an overnight Day 3 clinical research unit admission, and ability to comply with study visits and telehealth follow-up. Those criteria point toward a group with active, clinically meaningful disease burden rather than occasional use.
Eligibility Filters Aim to Protect a Vulnerable Cohort
Several exclusions reveal how cautiously the protocol is being built. Individuals with psychotic disorders, active suicidal ideation with intent and plan, a seizure history, significant head injury, structural brain abnormalities on baseline MRI, pregnancy, breastfeeding, or a contraindication to MRI are excluded. Cognitive impairment is also a barrier, with a Montreal Cognitive Assessment score below 24 disqualifying enrollment.
The study also excludes participants with substance use disorders other than cocaine, alcohol, marijuana, or nicotine, as well as those using medications that could substantially interact with the central nervous system effects under study, such as benzodiazepines or opioid substitution therapy. Those restrictions are not incidental. They are necessary if the trial is to interpret any change in brain activity as plausibly related to the stimulation rather than to confounding medication effects, acute psychiatric instability, or uncontrolled neurological risk.
A separate exclusion addresses the skull itself. If skull anatomy or scalp characteristics preclude safe delivery, individualized acoustic modeling may rule out participation. In practice, this means the biology of the head can determine whether the brain target is accessible at all.
Feasibility Endpoints Will Shape the Next Step
Beyond safety, the protocol measures whether participants complete the full sonication session, whether they complete all scheduled study visits, and whether they finish daily brief craving assessments. Those may sound operational, but in a phase 1 study they are essential indicators of whether a larger efficacy program is even possible.
Adherence is a major issue in substance use research, especially when the intervention is unfamiliar and the follow-up schedule extends across repeated visits. A trial can be scientifically elegant and still fail if participants do not return for imaging, craving assessment, or end-of-study evaluation. Here, completion of the full sonication session and the Day 4 post-sonication MRI are built into the primary readout because tolerability and workflow are part of the intervention’s viability.
Neural Signal Change Will Be the Most Closely Watched Readout
The protocol includes a Day 4 MRI endpoint measuring change from baseline in cue-reactive BOLD signal within craving-related brain regions on fMRI. That readout is the most mechanistically important part of the study because it asks whether stimulation can shift the brain’s response to cocaine cues even before any behavioral benefit is tested.
For a disorder driven by cue-induced reward learning, a measurable change in BOLD signal would suggest the device is reaching the circuitry it was designed to modulate. A lack of signal change would not close the door on the approach, but it would make the path to clinical efficacy less direct. Either way, the imaging endpoint is the bridge between device physics and addiction biology.
The trial is estimated to begin on 2026-11-01, with primary completion expected on 2028-08-31 and final completion on 2028-09-15. Those dates frame a long runway for a small, early-stage study whose main job is to show whether this form of neuromodulation can be delivered, tolerated, and measured with enough precision to justify what comes next.