One Sound, Seven Agencies, No Agreement

What We Know

In November 2016, two American intelligence officers stationed in Havana started hearing something outside their homes at night: a sharp, high-pitched sound with no obvious source. On December 30, one of them walked into the embassy clinic with a headache, ringing ears, and a description that would follow the case for years: a sound like a beam, aimed, he said, straight at his head. Within weeks, colleagues at the same small CIA station were reporting the same thing.

It didn’t stay in Cuba. In December 2017, in a Moscow hotel room, CIA officer Marc Polymeropoulos was hit by vertigo so severe he couldn’t stand, along with ringing ears, nausea, and a headache that wouldn’t quit. He was nearly three decades into a career running operations against Russian targets across Europe and Asia. The agency’s own medical office told him his case didn’t match what was by then called Havana Syndrome. It took him three years, and a diagnosis of traumatic brain injury at Walter Reed, before anyone agreed with him. He retired on medical grounds still fighting to be believed.

By then reports were arriving from personnel and family members in dozens of countries: China, Austria, Germany, Colombia, even Washington. In 2019, JAMA published brain scans of forty affected Havana personnel against forty-eight healthy controls and found measurable differences: less white matter, altered connectivity in the brain’s sound- and space-processing networks. The authors were careful. They didn’t know what the differences meant, and the pattern matched no known injury. Five years later, a larger NIH study scanned eighty-one more people the same way and found nothing distinguishing them from controls at all. Two of the most careful studies ever run on this question, five years apart, disagree with each other about what a brain scan should show.

The Roadblock

The shortage here was never investigation. After nearly a decade, seven separate intelligence agencies have reviewed largely the same case file and produced two different verdicts from it. Havana Syndrome resists proof from every direction at once. There’s no debris field, no black box, no injury that survives long enough to test twice. The government’s own bioacoustic analysis of the recordings personnel made of “the sound” matched one to a Caribbean cricket with high confidence — and stated plainly that no single energy source, sonic or electromagnetic, explains both those recordings and the reported medical effects. Symptom overlap compounds it: vertigo, migraine, and cognitive fog are also the symptoms of ordinary vestibular disorders and chronic stress, so any single case is genuinely hard to sort from background noise. And the file has become politically load-bearing on every side of it: an intelligence community facing diplomatic incentives to close the case, congressional investigators facing incentives to keep it open, and, underneath both, hundreds of people whose pain was never actually in dispute.

Best Guesses

Where you land depends on which fact you weight heaviest: the panel that called a weapon “most plausible,” the study that found nothing, or the government machine that keeps splitting 5-2 instead of landing on one answer.

Directed pulsed microwave weapon

The National Academies’ 2020 panel called “directed, pulsed radio frequency energy” the most plausible mechanism for the core cases it examined. The physics isn’t fringe: such energy can, in principle, produce auditory sensations and vestibular disturbance through a wall. The 2019 JAMA scans found real anatomical differences between exposed personnel and controls, not imagined ones, and two of the seven intelligence agencies that reviewed the case still keep some version of a foreign-capability finding alive, at low confidence. Still unresolved: no device has been recovered from an actual incident, one of the recordings personnel made matches a cricket, and a larger, later NIH study couldn’t reproduce the 2019 imaging.

Mass psychogenic illness

Neurologist Robert Baloh and sociologist Robert Bartholomew argue the pattern fits mass psychogenic illness: real, often disabling symptoms overlapping with known stress and vestibular disorders, spreading through a small, tightly connected community anxious about a real adversary. It strains hardest against the original 2016 Havana cluster, reported before public coverage existed to seed expectation, and it asks hundreds of people, including trained intelligence officers, to accept that an injury they experienced as external came from somewhere else. In its favor: the 2024 NIH study, the most rigorous one run so far, found nothing distinguishing patients from controls.

Many causes, no one culprit

This has been the U.S. intelligence community’s own position since March 2023: five of seven agencies concluded it’s very unlikely a single foreign adversary is behind these incidents, and that most cases trace to ordinary things, undiagnosed illness, environmental factors, preexisting conditions, that only look connected because they share a name. It doesn’t require a weapon nobody can find, and the IC’s own panel still carves out a subset of cases it won’t fold into this explanation. A House Intelligence subcommittee spent a year on the same evidence in 2024 and reached the opposite conclusion, calling it “stonewalling, slow-walking, and cherry-picking.”

An unidentified external attack

Marc Polymeropoulos and other affected officers separate two questions the government keeps merging: whether something external happened, and what it was or who did it. Their case rests on consistency — hundreds of people, many with no contact between them, describing overlapping symptoms after a specific trigger — and on the career cost several paid rather than accept a diagnosis nobody could explain. Polymeropoulos has gone further in his own public statements: “my contention is that it’s the Russians,” he’s said — his belief, not a government finding, and that gap is what this theory can’t close. Without a recovered device or an official conclusion behind the stronger claim, it rests on testimony, however genuine, rather than evidence anyone outside it can check.

The Verdict

Federal labs have been testing a device since early 2025 toward the only thing that would actually end this: not more testimony, not another agency memo, but a recovered device, a documented signal, and a reproducible injury, landing in the same file with an unbroken chain of custody. Reporting since has moved closer: animal tests reportedly produced injuries resembling victims’ symptoms, and a researcher who tested a similar device on himself reported the same effects. Real data, still short of a documented case in an actual human incident. Until one arrives, the honest picture holds two things at once without collapsing either one into the other. A National Academies panel found real signs of injury in some patients that standard medicine still can’t explain. The government’s own intelligence agencies, with more access to classified information than anyone outside them, mostly can’t find a foreign hand behind it. Marc Polymeropoulos is still waiting for the version of this file that matches what happened to him in that hotel room.