The line between a bad habit and a clinical compulsion is not always clear, but the brain circuits driving them share surprising overlap. In a recent episode of Huberman Lab Essentials, neurosurgeon Dr. Casey Halpern, a professor at the Perelman School of Medicine at the University of Pennsylvania, walked through how deep brain stimulation and other neuromodulation techniques are being used to treat disorders rooted in impulsivity and craving — from Parkinson's disease and obsessive-compulsive disorder to binge eating disorder and depression.

The episode, drawn from a longer conversation with host Andrew Huberman, distills years of clinical and surgical research into a digestible look at what happens when the brain's reward and control systems go off balance, and what doctors can do about it.

What deep brain stimulation actually does

Deep brain stimulation (DBS) involves implanting electrodes in specific brain regions and delivering controlled electrical pulses to modulate neural activity. Halpern explained that DBS is not a new technology — it has been used for decades in movement disorders like Parkinson's — but its application to psychiatric conditions is a more recent frontier.

In Parkinson's, DBS typically targets the subthalamic nucleus or globus pallidus to reduce tremors and rigidity. The same basic approach is now being adapted for conditions where the problem is not movement but compulsive thought or behavior. The key is identifying which brain circuits are overactive or underactive in each disorder.

Halpern's research focuses on the nucleus accumbens, a region deep in the brain that processes reward, risk, and motivation. Dysfunction in this area has been linked to addiction, binge eating, and OCD. By adjusting the electrical stimulation delivered to the nucleus accumbens, surgeons can dampen the pathological signals that drive craving and compulsive urges.

OCD and the brain circuits that lock up

Obsessive-compulsive disorder affects roughly 1-2% of the population, and a significant portion of patients do not respond adequately to medication or therapy. Halpern described how OCD involves a loop between the orbitofrontal cortex, the anterior cingulate cortex, and the striatum — a circuit that gets stuck in a repetitive pattern. DBS can interrupt that loop.

The procedure is not a cure, but for patients with severe, treatment-resistant OCD, it can dramatically reduce symptom severity. Halpern noted that DBS for OCD is already approved under a Humanitarian Device Exemption by the FDA, meaning it is available for use in select cases.

What makes the approach powerful, according to Halpern, is that it does not sedate or blunt emotion. Patients remain fully aware and able to engage in therapy. The stimulation simply lowers the volume on the intrusive signals that drive compulsions.

Binge eating disorder and the craving circuit

One of the more striking applications Halpern discussed is binge eating disorder. Unlike occasional overeating, binge eating involves episodes of consuming large amounts of food in a short time, often accompanied by a sense of loss of control. Halpern and his team have been exploring whether DBS targeting the nucleus accumbens can reduce the intensity of those cravings.

In a small clinical trial, patients with severe binge eating disorder who received DBS reported fewer binge episodes and a reduction in the urge to eat compulsively. Halpern emphasized that the stimulation does not eliminate appetite or pleasure from food. It appears to act on the anticipatory craving — the moment when a person feels they cannot resist a bag of chips or a pint of ice cream, even when they know they should stop.

That distinction matters. Many people mistake craving for hunger, but they are driven by different neural pathways. DBS seems to give patients a window of control that they did not have before.

Non-invasive alternatives: transcranial magnetic stimulation

Not every patient wants or needs brain surgery. Halpern also covered transcranial magnetic stimulation (TMS), a non-invasive technique that uses magnetic fields to stimulate nerve cells in the brain. Unlike DBS, TMS does not require an implant. It is applied externally, typically over the prefrontal cortex.

TMS is already FDA-cleared for depression and OCD, and Halpern discussed its potential for treating compulsive behaviors. The advantage is obvious: no surgery, no risk of infection or hardware complications. The trade-off is that TMS cannot reach deep structures like the nucleus accumbens as precisely as DBS can. It affects broader cortical regions, which may be sufficient for some conditions but not others.

Halpern suggested that the field is moving toward a spectrum approach — starting with non-invasive stimulation for less severe cases and escalating to DBS when the condition is refractory and debilitating.

Using artificial intelligence to predict impulsive behavior

One of the more forward-looking segments of the conversation involved artificial intelligence and machine learning. Halpern and his collaborators are developing algorithms that can detect when a patient is about to have a compulsive episode — before the person themselves is fully aware of it.

By monitoring physiological signals such as heart rate, skin conductance, or even neural activity from implanted electrodes, machine learning models can identify patterns that precede a binge or a compulsion. In theory, a closed-loop system could then deliver a brief pulse of stimulation at the critical moment, preventing the behavior from unfolding.

Halpern described this as a kind of "neural fire alarm" — a way to intervene not after the fact, but just before the urge becomes overwhelming. The technology is still in early stages, but pilot studies have shown encouraging results in predicting impulsive behavior in real time.

The ethical questions around such systems are real. Who controls the stimulation? What happens if the algorithm is wrong? Halpern acknowledged these concerns but argued that for patients whose compulsions have wrecked their health and relationships, the potential benefit outweighs the risk.

What comes next

The episode makes clear that neuromodulation is not a one-size-fits-all solution. DBS and TMS target different parts of the brain with different levels of invasiveness. The choice depends on the condition, its severity, and the patient's tolerance for risk.

Halpern's work points toward a future where psychiatric treatment is not limited to pills and talk therapy but includes precise, circuit-based interventions. That shift will require more clinical trials, better imaging, and algorithms that doctors and patients trust.

For now, the takeaway is straightforward: the same brain circuits that make you reach for a second slice of cake or obsess over a locked door are not broken in people with severe compulsive disorders — they are stuck on high volume. DBS and its non-invasive cousins are turning the dial down.