George Lab

06/12/2026

Nine years of work, several labs, and a lot of people who refused to give up: our co***ne addiction study is finally out!
Co***ne addiction still has no approved medication, and one of the hardest questions is why some people are so much more vulnerable than others. It's hard to study that cleanly in people because you don't control the environment, so we turned to rats. We followed 836 of them over several weeks, letting each one decide how much co***ne to take. It is one of the largest studies of its kind ever done.
The behavior turned out to be inherited. Some of the genes we found are already known to matter for co***ne, alcohol, and to***co addiction in people, which tells us these rats point to the same biology we see in humans. We also found brand new genes for which we have little knowledge about, and those give us fresh targets to go after for future treatments.
This is a step, not a cure. But for a condition with no approved treatment, having real, specific leads is worth a lot.
Huge thanks to my lab, our collaborators, and the NIH for believing in this. Addiction is a medical condition with real biology behind it, not a failure of willpower.

The genetic basis of co***ne use behaviours is poorly understood. Here, the authors identify genome‑wide loci in an outbred rat population linking variants in co***ne‑metabolising enzymes and addiction‑related genes to self‑administration traits, providing insight into mechanisms of addictio...

06/05/2026

For more than thirty years, scientists have searched for a medication to help people addicted to co***ne or m**hamphetamine, and so far none has been approved. A new study looking back at 12 clinical trials and 2,000 patients points to one reason we keep falling short: the way we measure success.
Almost every trial asked the same yes-or-no question. Did the person stop using completely? Anything less counted as failure. But quitting cold is incredibly hard, especially early in treatment, and demanding it can hide real progress.
When the researchers instead looked at whether people meaningfully cut down their use, which on its own improves health and daily life, the medications started to look more promising. One older drug, cabergoline, stood out for co***ne: about 44 out of every 100 patients reduced their use, compared with 28 out of 100 on a placebo. That promise was invisible when the only thing that counted was quitting entirely.
We already made this shift with alcohol. Cutting back, even without quitting, is now accepted as real progress and a valid goal in treatment research. It's time we extend that same common sense, and compassion, to people struggling with co***ne and m**h, especially as overdose deaths keep rising.
Addiction is a brain condition, not a moral failing, and recovery rarely happens in a single perfect step. Recognizing that cutting back counts could change how we test treatments and finally help us find ones that work.

06/03/2026

You have probably heard of Ozempic and Wegovy, the new diabetes and weight-loss medicines. A growing body of research suggests this same family of drugs may also help people struggling with addiction.
In a new study, researchers tested one of these medicines, semaglutide, in rats that had been taking fentanyl, the powerful opioid behind most overdose deaths today. People relapse for three main reasons: they run into reminders of the drug (paraphernalia), they get a small taste of it again (single hit), or they hit a wave of stress. The researchers tested all three.
The medicine reduced drug seeking in every case. When relapse was triggered by a small dose of the drug or by stress, the seeking behavior dropped almost to nothing.
I study this kind of research for a living, so I will give you both sides. An effect this strong, across every type of trigger, is rare and exciting. It also means scientists need to confirm the medicine is acting on craving itself, and not just lowering motivation in general. That is the next question to answer.
If it holds up, it points to som**hing remarkable: that these medicines may quiet the brain's drive to seek a drug no matter what sets the craving off. For a crisis that takes tens of thousands of lives every year, that possibility is worth chasing.

https://doi.org/10.1097/FBP.0000000000000880

05/27/2026

I have to admit it. I get more excited by a careful behavioral study than by the latest fancy lab technology.

This is a great example. Researchers at the National Institute on Drug Abuse trained rats on he**in and asked a simple question: how long do drug-associated cues keep their power?

They tested at up to 385 days of abstinence. More than a year. That's over half the adult lifespan of a rat. In human terms, you can think of it like following someone in recovery for decades.

The cues still worked. The rats still responded fully to them after a year of being completely he**in-free. If you have a family member in recovery from opioids, this is the science backing up som**hing you already know. The pull of old places, old people, old routines does not go away just because the drug use stopped.

But here is the part I find most interesting for treatment. The researchers had also trained a different cue, one that told the rat "drug is not available right now." That cue kept suppressing drug-seeking just as durably across the whole year. It even worked when the rats were given an injection of he**in to try to push them back toward use. The brake held.

This matters for how we think about rehab. Most recovery programs focus on avoiding triggers. This study suggests we could also work on building stronger learned signals that actively tell the brain "no drug here, no drug now." That's a different tool than anything currently in standard treatment, and it might survive the moments when willpower fails.

Now here is what I want to talk about beyond the paper.

This kind of careful behavioral research is getting rare in the US. The funding agencies, the journals, the people who decide who gets hired and promoted, all of them have shifted heavily toward flashy new technologies and molecular biology tools. Those tools are powerful and I use them in my own lab. But the patient behavioral work, the kind that asks what a drug actually does to a living animal over a year of its life, is being quietly defunded. The senior behavioral scientists are retiring. The new generation is following the funding into other areas.

If this continues, we lose the scientists who know how to do this work, and we lose it in one generation.

So I want to celebrate this paper, and I want to ask you to care about som**hing most people never hear about: behavioral addiction research needs public support. Please tell your representatives that NIH funding for basic behavioral research matters.

Credit to the journal Neuropsychopharmacology for publishing it.

https://doi.org/10.1038/s41386-026-02447-6

05/24/2026

A new scientific review just came out that I think is one of the most important stories in addiction medicine right now.

The drugs you may have heard about for weight loss and diabetes, the ones in the same family as Ozempic and Wegovy, also reduce the urge to use co***ne, m**hamphetamine, ni****ne, alcohol, and opioids. The same medication. Across all of these.

If that holds up in larger human trials, and the evidence is pointing that way, it would be the first medication ever to treat addiction across multiple substances at once. Right now we have one drug for one drug. This would change that.

Here is the part that means a lot to me. Preclinical scientists have shown for over a decade that peptides, including GLP1A, may decrease the motivation for alcohol or drugs (a group in Sweden pioneered it). However, we were often told that animal research cannot predict useful treatments for mental health conditions, and that this kind of slow careful laboratory work is not worth the investment. Researchers kept doing it anyway, in labs that were often underfunded and overlooked.

Now, 15 years later, that work is paying off in a way that could help millions of people who have not had real options. People with addiction, who are loved by people who care about them, finally have a reason to be hopeful that better medications are coming.

This is what science looks like when it works. Slow, quiet, patient research in a laboratory eventually changing what doctors can offer a person sitting in their office a couple of decades later.

Worth supporting. Worth funding. Worth the wait.

https://doi.org/10.1016/j.biopsych.2026.05.002

05/22/2026

I smoked for more than 30 years. I tried ni****ne electronic ci******es to quit, and they never worked for me. What finally worked was varenicline (a prescription medication). I share that upfront because everyone is different, and quitting smoking is one of the hardest things a person can do.

That said, this is a clean clinical trial worth knowing about, especially if you have a smoker in your family.

Researchers at Penn State took 104 adults who smoke ci******es and wanted to quit. Half got a real ni****ne v**e. The other half got a v**e that looked, felt, and tasted identical but contained zero ni****ne. Neither the participants nor the researchers knew who got which. Six weeks later, they checked who had actually managed to quit.

The real ni****ne v**e: 36.5% of people quit ci******es completely.
The dummy v**e: only 11.5%.

That is roughly three times the quit rate. The people who switched to the ni****ne v**e also had significantly lower levels of two of the most dangerous chemicals in cigarette smoke (acrolein and acrylonitrile, both linked to cancer and lung disease).

The takeaway is simple. The ni****ne is what helps people quit. The v**e device on its own does almost nothing. For a public conversation that often treats all v**es as either miracle cures or pure villains, this kind of carefully controlled trial gives us som**hing rare: a real answer. And it adds one more tool to the list of things that can help people quit, alongside medications like varenicline and ni****ne patches.

Smoking still kills about 480,000 Americans every year. Tools that triple the odds of quitting matter.

https://doi.org/10.1001/jamanetworkopen.2026.13292

05/22/2026

Go big or go home.

Bigger reward leads to better learning is not exactly news. What's striking is that much of brain research has stuck for decades with very small rewards, based on old findings suggesting the brain's reward signal maxes out at modest sizes, so going bigger was assumed pointless (more trials is better for recordings though).

Worth flagging that this wasn't universal. The addiction field, including our lab, has long used larger rewards (water, alcohol, sweet drinks) when training animals. Looks like we were onto som**hing.

Mice learn complex tasks (navigation, motor skills, decision making) about 10 times faster when given a few big rewards instead of many small ones. Same total reward per session. The brain's dopamine signal scales with reward size, keeps animals engaged in the task, and speeds up learning.

https://doi.org/10.1126/science.aeb0813

05/20/2026

Interesting line from the Robin paper:

"All hypotheses, experimental directions, data analyses, and data figures in the main text of this report were produced by Robin."

Curious what NIH thinks of this. They currently prohibit grant proposals that are "substantially developed by AI" from being considered the original work of applicants.

From X: https://twitter.com/i/web/status/2056898746961846460

05/20/2026

I'm not buying the doomsday scenario.

AI multiplies and accelerates research, making it more robust and replicable, which leads to more success and growth, and inevitably more scientist employment, not less.

Yes, the job will change. But we'll need scientists to guide the process, curate findings, validate results, translate them into real-world outcomes, and be accountable for them.

From X: https://twitter.com/i/web/status/2056898746961846460

05/20/2026

One of many reasons we'll still need scientists for many years to come in the AI world.

Because sometimes AI produces garbage outputs, and while the tech bros and VCs will gladly accept those uncritically, a trained scientist will not.

From X: https://twitter.com/i/web/status/2056956645331410993