Most important take away

Compulsive disorders such as OCD, binge eating, and addiction share a common neural signature: pursuing a reward despite obvious risk, gated largely by the nucleus accumbens within the ventral striatum. Dr. Halpern’s work suggests that identifying disease-specific “craving” or “obsession” signals in this circuit can enable closed-loop deep brain stimulation that interrupts the urge before the behavior occurs, opening a path to more precise, eventually non-invasive therapies.

Summary

Dr. Casey Halpern, a stereotactic functional neurosurgeon at Penn Medicine, joins Andrew Huberman to discuss how deep brain stimulation (DBS) and related tools can treat severe, treatment-resistant compulsive disorders. The conversation centers on a unifying theme — the urge despite the risk — that links OCD, binge eating disorder, addiction, and other compulsive behaviors, and explores how mapping and modulating specific brain circuits could transform care for patients who fail medication and therapy.

Key themes:

• Compulsion as a common circuit problem. OCD, binge eating, addiction, and eating disorders all involve pursuing a reward despite punishment or harm. This pattern points to dysfunction in cortico-striatal circuits, particularly the prefrontal/orbitofrontal cortex and the ventral striatum/nucleus accumbens, which gates reward-seeking and compulsive behavior.
• Layered standard of care. First-line OCD treatments include SSRIs, tricyclics (serotonin system, with downstream noradrenergic and dopaminergic effects), and exposure-and-response-prevention therapy. Roughly 30% of patients remain symptomatic and become candidates for surgical options.
• DBS and ablation for severe cases. DBS implants a thin electrode to deliver electrical stimulation to a precise target; capsulotomy ablates a small (3–4 mm) region. Both produce a roughly 50% responder rate, and responders often retain symptoms — motivating more disease-specific targets.
• Finding “craving cells.” Borrowing from how tremor cells are localized intraoperatively in Parkinson’s surgery, Halpern’s team listens for electrical signatures of craving (binge eating) and obsession (OCD) in the nucleus accumbens. A single-case OCD study showed proof of concept for eliciting disease-specific signals during awake surgery.
• Closed-loop, responsive devices. Implanted devices can both record and stimulate. In lab “mood provocation” studies — patients wear eye trackers, are videoed, and are deliberately triggered toward a binge — researchers validate that the device can detect the craving signal so stimulation can be triggered only when needed at home.
• Awareness has limits in severe disease. CBT and improved self-awareness help many, but the most refractory patients binge even under one-way-mirror surveillance because they cannot control the urge. For them, restoring control via neuromodulation is the goal.
• Non-invasive frontiers. TMS is FDA-approved for depression, OCD, and nicotine addiction but lacks spatial precision. MR-guided focused ultrasound is FDA-approved for tremor and offers non-invasive ablation; the bottleneck for psychiatric use is identifying the right target. Halpern argues we must “get in the brain before we get out of it” — invasive recordings in humans (leveraging epilepsy SEEG techniques) will define targets for future non-invasive therapies.
• AI and wearables. Voice, sleep, and behavioral pattern analysis (e.g., work at the University of Washington on suicidal ideation) may detect impending impulsive or depressive episodes before patients consciously notice — a potentially scalable complement to surgery, which has only been performed ~200,000 times globally against tens of millions of affected patients.

Actionable insights:

• For people with mild-to-moderate compulsive tendencies, building self-awareness of triggers, paired with CBT/exposure-response prevention and SSRI evaluation, remains the highest-yield path.
• For patients with severe, treatment-resistant OCD, binge eating disorder, or addiction, ask specialty centers about DBS trials, capsulotomy, MR-guided focused ultrasound, or SEEG-based research studies — therapies are imperfect but improving rapidly.
• Recognize that “urge despite risk” is the operative warning sign across compulsive conditions; it justifies seeking treatment escalation rather than relying on willpower alone.
• Watch the convergence of invasive human neuroscience, focused ultrasound, TMS, and AI-driven behavioral monitoring — these are the most promising scalable solutions for what Halpern calls epidemic-scale public health problems (obesity, opioid use, depression, suicide).

Chapter Summaries

• What a neurosurgeon does and what DBS is. Halpern explains the breadth of neurosurgery and his subspecialty in stereotactic functional neurosurgery, where he performs DBS and MR-guided focused ultrasound. DBS delivers “electricity as medication” through a thin implanted electrode, and side effects from nearby circuits (e.g., laughter, mood lifts) have inspired new psychiatric applications.
• Understanding OCD. OCD spans a spectrum; first-line treatments are SSRIs/tricyclics and exposure-response prevention therapy, but ~30% of patients remain symptomatic. For severe cases, DBS or capsulotomy ablation can help, with roughly 50% responder rates that still leave symptoms — motivating more targeted approaches.
• The compulsion circuit and the nucleus accumbens. Cortical hyperactivity (prefrontal/orbitofrontal) projects to the basal ganglia and ventral striatum, where the nucleus accumbens gates reward-seeking and compulsive behavior. The “urge despite risk” pattern unifies OCD, addiction, eating disorders, and gambling.
• Listening for craving and obsession cells. Drawing on the Parkinson’s playbook of identifying tremor cells, Halpern’s team records single-cell activity intraoperatively to find craving-related signals in binge eating disorder and obsession-related signals in OCD, optimizing electrode placement.
• Non-invasive stimulation: TMS and focused ultrasound. TMS is FDA-approved for depression, OCD, and nicotine addiction but is spatially imprecise. MR-guided focused ultrasound is FDA-approved for tremor and is being explored for modulation, blood-brain-barrier opening for drug delivery, and eventually psychiatric targets — but first researchers must identify those targets via invasive human studies.
• SEEG, epilepsy techniques, and the path forward. Stereo-EEG (originally for epilepsy) is being adapted to study OCD, depression, and binge eating disorder. Pooling invasive data could yield validated targets that are eventually treated non-invasively with ultrasound or TMS.
• Lab provocation studies and the limits of awareness. In mood-provocation experiments with implanted recording/stimulating devices, severe binge eating patients still binge under direct observation, demonstrating that awareness alone is insufficient for the most refractory cases — neuromodulation is needed to restore control.
• AI, wearables, and scaling solutions. Voice analysis, sleep patterns, and machine learning could flag impending impulsive, depressive, or suicidal episodes before patients consciously notice. Surgery alone cannot address tens of millions of patients, so scalable AI-assisted detection paired with rigorous neuroscience is essential.
• Closing reflections. Halpern emphasizes that compulsive behaviors are among the largest public health problems globally, and progress will come from combining invasive human neuroscience, non-invasive devices, and AI — with surgery reserved for the most severe patients.