A man in his forties comes into the office after his third residential program. He did the counseling, the meetings, and the 28 days, and he was using cocaine again within weeks of discharge. His family wants to know what he is failing to do. I have heard that question for thirty years, and the honest answer is usually that nobody looked at his brain before choosing his treatment.
A study in Nature Mental Health, led by Kanhao Zhao and Yu Zhang at Lehigh University, takes a step toward changing that. The team asked two questions. Could a resting-state functional MRI separate people with cocaine use disorder from healthy controls, and could the same scan predict who would get craving relief from repetitive transcranial magnetic stimulation (rTMS)?
What the scans showed in cocaine use disorder
The researchers built their model in a discovery group of 129 people in Mexico, 71 with cocaine use disorder and 58 healthy controls. The connectivity pattern they identified classified people correctly 83 percent of the time. They then tested it in an independent group of 82 patients and 81 controls from sites in the United States and Mexico, where accuracy fell to 70 percent. That drop is expected when a model meets new scanners and new patients, and it is the more believable number of the two.
The pattern itself had two parts. Networks for visual processing and attention, along with networks for goal-directed behavior, showed increased connectivity in patients. The default mode network, which is active when the mind wanders and turns inward, showed reduced connectivity with the limbic network that handles emotion and reward. These data are observational, so they describe how the brains differed and cannot tell us which differences came first. My own interpretation is that the pattern fits what we see clinically, a brain that is highly tuned to drug cues and poorly coordinated in the circuits that let a person step back from an urge.
Who got relief from magnetic stimulation
The second half of the study is the part that matters for treatment. The team looked at 45 patients from a randomized clinical trial, 25 who received active rTMS and 20 who received sham stimulation. Craving was self-reported and measured after two weeks. The connectivity model predicted which patients in the active group would have less craving, and it failed to predict outcomes in the sham group. That contrast is what you would expect if the scan were picking up something about how the brain responds to the stimulation itself.
Active treatment also appeared to shift connectivity between the networks involved in cognitive control and movement toward the pattern seen in healthy controls. The paper reports this as a change in brain function that accompanied the change in craving, and a trial of 45 people cannot show that one caused the other.
Why this belongs in the conversation about rehab
Most people who enter treatment for stimulant use are matched to a program by bed availability, insurance coverage, or reputation. Almost none are matched by what their brain is doing. When the program fails, the patient carries the label of someone who did not try hard enough. This is not about willpower, and it never was. Reward circuits, cue reactivity, and cognitive control are measurable, and they differ from person to person.
The limits are real and deserve a clear statement. The samples included considerably more men than women, so the signature may not hold in women. The replication data came from multiple sites with different scanners. A classification accuracy of 70 percent in new patients is a research finding and not a diagnostic test. And rTMS is not FDA-cleared for cocaine use disorder, so any use for this purpose is off-label and investigational, even though the device has clearance for depression and obsessive-compulsive disorder.
What this means for you or someone you love
If you or a family member has relapsed after rehab, ask what was measured before the treatment plan was chosen. Cognitive testing, imaging, sleep, nutritional status, and co-occurring depression or anxiety all change what a person needs. This is the reasoning behind Rescue From Rehab at The Neurogenesis Project, which starts with precision diagnostics of the brain and builds a plan around what we find, with neuromodulation as one option among several for patients who are appropriate candidates. Our Braincast episodes cover how that workup runs for readers who want to hear it explained.
For researchers, the next step is clear. A larger prospective trial with balanced enrollment of men and women, in which scan-guided selection for stimulation is compared against treatment assigned without a scan, would show whether this signature improves outcomes for patients. If it does, we move from asking who is willing to try again to asking which treatment fits this brain.
People in recovery deserve a plan that was built for their brain. A signature like this one is early, but it points toward a day when the first question is what the scan shows and the last question is about character.
Hero image: illustration created with AI image tools.
This article was drafted with the assistance of AI writing tools, then reviewed, edited, and approved by Dr. Drew W. Edwards, Ed.D., who holds full editorial responsibility for its content.
Sources:
Zhao, K., Zhang, Y., et al. "Discriminative functional connectivity signature of cocaine use disorder links to rTMS treatment response." Nature Mental Health. Reported by Petrova, K., PsyPost, November 7, 2026: Brain connectivity pattern predicts treatment response in cocaine use disorder.