Classification: Unverified account

Havana Syndrome: real symptoms, disputed origins

Since 2016, diplomats and relatives have reported sudden episodes of pressure, sounds, and vertigo. Studies, investigations, and official assessments diverged; the withdrawal of two intelligence reports in 2026 reopened the dispute without settling the cause.

Havana, Havana, Cuba · See stories from this place and nearby

The first reports in Havana

In 2016, members of the United States mission in Havana began reporting abrupt episodes: a sound perceived as directional, pressure in the head, pain, and dizziness. For some, the sensation passed; others described lasting problems with balance, concentration, sleep, and work. “Havana Syndrome” takes its name from this first cluster of reports, but official communications came to use “anomalous health incidents,” or AHIs. The distinction matters: AHI is a category for reporting and investigation, not a diagnosis of one disease or confirmation of an attack.

The authorities’ response

Officials responded before the cause was known. The U.S. State Department investigated the conditions reported by its diplomatic community, and in 2018 an independent board examined the institutional response, interviewed more than 116 people, and issued 30 recommendations. Its mandate was to assess protection, communication, and care, not to determine the mechanism behind the episodes. Similar reports later appeared at other posts, including Guangzhou, China; Canadian staff and family members in Cuba also reported symptoms. The phenomenon spread geographically without proving that every event shared one cause.

Symptoms, causes and the attack hypothesis

Accounts that call the case a “sonic attack” often collapse three questions into one: did people have symptoms, was there an external stimulus, and did someone create it deliberately? Clinical accounts and the need for care support the first. The other two remain contested. Perceiving a sound does not, by itself, identify its external source, and a normal scan does not make pain imaginary. Differences between episodes and the challenge of measuring a possible exposure after the fact are central problems for investigators.

The National Academies report

In 2020, a U.S. National Academies committee compared hypotheses including infections, chemicals, psychological factors, and pulsed radio-frequency energy. It judged the last to be the most plausible of the mechanisms it considered for certain abrupt episodes with distinctive sensory features. It did not identify a device, a source, an assailant, or the cause of every report. The report itself emphasized the variation among cases, limited access to individual data, and the possibility of multiple mechanisms. It also discussed persistent postural-perceptual dizziness, a functional balance disorder that can prolong symptoms after different triggers; “functional” does not mean feigned.

The Canadian investigation

Canada followed other leads in Havana. Global Affairs Canada documented a police investigation into possible radio-frequency exposure and a health study monitoring diplomats and family members. Following preliminary findings, it conducted an environmental review of pest-control practices. Pesticide exposure, raised as a hypothesis in this context, remains something investigated rather than a proven explanation. The several lines of inquiry show why a single, premature answer would mislead: the environment, individual health, and circumstances of each episode must be considered separately.

What the NIH examinations found

In March 2024, NIH published assessments of more than 80 people who had reported AHIs, comparing them with control groups. Clinical tests, biomarkers, and MRI scans found no consistent pattern of imaging-detectable brain injury or differences in most measures studied. The researchers were explicit: this does not rule out an event affecting the brain at the time symptoms began, nor deny how severe or lasting those symptoms can be. Some participants met criteria for a functional neurological disorder or had significant somatic symptoms; that observation cannot be generalized to every case. The study can measure only what its methods and participants allow, not every possible exposure in 2016.

Disputed intelligence assessments

The intelligence community also disagreed. An assessment published in January 2025 said five components judged a foreign actor’s involvement very unlikely and one judged it unlikely. Two components, however, had revised their views on whether a foreign technological capability was possible; one judged there was a roughly even chance that a device had harmed a small, undetermined subset of people, with low confidence. Possible capability, proven use, and identified perpetrator are different propositions. A House intelligence subcommittee separately challenged the assessments’ integrity and judged foreign involvement in some events increasingly likely. Much of the information it cited remained classified, limiting public verification.

The official shift in 2026

In June 2026, the Director of National Intelligence formally withdrew the 2023 and 2025 assessments and ordered a new review, citing failures in the analytic process. The 2025 document remains useful as a record of what agencies concluded at the time, but cannot be cited as a still-operative official position without noting its withdrawal. The decision identified no perpetrator and demonstrated no weapon. Joint reporting by The Insider, 60 Minutes, and Der Spiegel suggested a possible connection with members of a Russian intelligence unit; in 2026, 60 Minutes reported tests of a device capable of emitting pulsed energy. These investigations matter, but suggested proximity, a device’s existence, and a causal link to particular episodes require different kinds of proof. Reporting alone does not settle attribution.

People affected beyond the political dispute

The least visible part of the case lies away from spycraft. A U.S. Government Accountability Office audit found delays in care and gaps in the patient registry: of 334 people qualified for care in the military health system in January 2024, only 33 had been entered into the AHI registry by May that year. These numbers describe an administrative situation at that time, not the number of people affected or a current rate. They show, however, how consent, agreements between agencies, and medical coordination shape both care and the quality of data available to researchers.

The limit of what is known

Havana is thus the source of the name, not the boundary of the question. Symptom reports, institutional responses, clinical research, and official disagreements are documented. The public material reviewed does not establish one mechanism for every case or responsibility by any government. Progress requires comparable clinical data, transparent records of individual episodes, reproducible technical investigation, and care that does not depend on a single theory of cause. That double obligation, to investigate without inventing certainty and to provide care without demanding certainty first, makes the case larger than its mystery.

Open in Google Maps · Approximate location