A Decade in the Making

Each year in the United States, more than 300,000 people die from sudden cardiac arrest — a condition where the heart's electrical system malfunctions without warning. The medical emergency can kill both high-risk older adults and seemingly healthy young athletes with no history of heart issues. While internal defibrillators that shock the heart back into rhythm can save lives, figuring out who actually needs one remains a high-stakes guessing game. For more context on this story, see our ongoing breaking AI news.

Now, a UC Berkeley-led research team has discovered a previously unrecognized signal in electrocardiograms (EKGs) that can detect high-risk patients before their heart stops. The study, published June 24, 2026, in the journal Nature, could transform how doctors identify who needs a life-saving implant.

"Medical decisions are really hard, and I think that's why AI is so exciting for me," said Ziad Obermeyer, an associate professor at UC Berkeley's School of Public Health and the study's lead author, in a university press release. "We can not only make better decisions, but also start to understand what's actually going on with these patients before their heart stops."

Training on 440,000 EKGs

Using more than 440,000 EKGs from Sweden paired with information from death certificates, researchers trained an artificial intelligence model to analyze the spikes and waveforms produced by the heart's electrical currents. They fed the model scans from healthy people, at-risk patients, and those who later suffered cardiac death until it recognized waveform patterns unique to people who went on to die suddenly.

The team then validated the algorithm over multiple years on thousands of additional patient files from both the United States and Taiwan. According to the university, the data compilation alone took roughly a decade, drawing on six years of scans from Sweden's unified health system, two years of de-identified EKGs from a hospital system in San Diego, and a separate dataset from Taipei.

"Good AI starts with good data," said Obermeyer, who is also part of the joint UCSF-UC Berkeley Computational Precision Health program. "Unfortunately, data like the ones we used for this study are incredibly hard to access. It's a big part of why there's so little clinical AI in use today."

Outperforming Standard Tests

The results represent a meaningful leap over current screening methods. The most commonly used technique to identify at-risk patients measures how much blood the heart ejects with each beat — known as the ejection fraction. Those tests identify a high-risk group with a 4.6% annual rate of sudden cardiac death.

The AI system, by contrast, isolates a high-risk group with a 7% annual rate — flagging a larger pool of patients who are at genuine risk but who look low-risk by every current standard. The vast majority of these patients would never have been referred for further evaluation under existing protocols.

Whereas a heart attack stems from restricted blood flow, cardiac arrest occurs when the heart's electrical current suddenly stops firing. CPR and a shock from an automated external defibrillator can save lives, but approximately 90% of those who suffer sudden cardiac arrest outside a hospital die within minutes. The speed and unpredictability of the event make prevention extraordinarily difficult.

The Defibrillator Dilemma

The irony of sudden cardiac death, Obermeyer noted, is that the treatment already exists — doctors just cannot reliably identify who needs it in time. Implantable cardioverter-defibrillators (ICDs) are small devices placed in the chest that monitor heart rhythms and deliver shocks when dangerous patterns are detected.

The problem is twofold. First, the standard ejection-fraction test requires patients to undergo a more involved medical evaluation — something the vast majority of eventual victims never knew they needed. Second, two-thirds of implants placed in patients flagged by current tests never actually fire, meaning patients undergo invasive, costly procedures to prevent an emergency they may never face.

"In some fraction of those people, we could have prevented those deaths if we had just known it in time," Obermeyer said. "There are a lot of lives being lost from people who are dropping dead of sudden cardiac death that are preventable if we just had better AI tools to find these things."

What Comes Next

The next phase of the project has already begun. Obermeyer is working with health systems in Sweden, Taiwan, and the United States to deploy the algorithm on hospital EKG databases. For scans the system flags as high-risk, doctors would notify patients and offer them the option of wearing a patch that continuously monitors their heart. That monitoring data could help researchers better understand the physiological mechanism within the heart that generates the telltale signals — and could ultimately lead to the placement of a potentially life-saving internal defibrillator.

The study also opens the door for new research into what the AI tool actually detected. The waveform patterns it homed in on appear related to the heart suddenly and fatally misfiring, but the precise physiological mechanism remains unclear — a question the team hopes continuous monitoring will help answer.

Obermeyer also built a website where individuals interested in assessing their own risk can submit basic information and an email address, allowing the research team to contact them for EKG analysis once the tool becomes more widely available. The project draws on the work of two organizations Obermeyer co-founded — Dandelion Health and Nightingale Open Science — that operate at the intersection of AI and medical research.

Beyond its immediate clinical implications, the study underscores a broader point about AI in medicine: the technology's greatest value may lie not in replacing doctors, but in revealing patterns invisible to human eyes — patterns that, once understood, could save thousands of lives every year.

"There is also going to be a new way of doing science that comes out of these tools," Obermeyer said, "and it's fun to think about how that starts happening."

---

Stay Ahead of AI

Get the latest AI news, analysis, and breakthroughs — all in one place.

Read more AI news →