Something strange was happening in a German clinic. And it pointed to a much larger problem.
In the spring of 2019, psychiatrist Kirsten Müller-Vahl began noticing an unsettling pattern at the Tourette’s syndrome clinic at Hannover Medical School. Teenagers and young adults were arriving with sudden-onset, elaborate tic-like behaviours that looked almost nothing like the syndrome she had spent decades studying and treating. Classic Tourette’s typically begins with simple, involuntary movements — a blink, a twitch, a shoulder roll. These new patients were different. Some made complex, choreographed motions. Others shouted obscenities or strange, specific phrases, including, bewilderingly, the words “flying sharks.”
Stranger still, their symptoms were sometimes nearly identical to one another’s.
That last detail is the key. In organic neurological conditions, symptoms vary. They emerge from individual neurology, individual history, individual bodies. When patients start presenting with the same unusual phrases, the same gestures, the same dramatic flourishes, a clinician trained to notice patterns begins asking a different kind of question — not “what is wrong with this person?” but “where did this person learn these behaviours?” The answer, researchers would eventually determine, pointed directly at social media, and at a constellation of popular online creators documenting their own Tourette’s experiences in vivid, widely shared videos.
Rhode Island psychiatrist Nebyou Belay, writing from clinical experience, uses this episode to illuminate something that extends far beyond Tourette’s syndrome. We are living through a moment when the traditional sequence of mental health care — symptoms appear, a person seeks help, a trained clinician evaluates and diagnoses — has been quietly inverted. On platforms designed to maximize engagement and reward relatability, the diagnosis frequently arrives first. A teenager watches enough content about ADHD, borderline personality disorder, or autism spectrum disorder to begin organizing their own experience around those frameworks, long before they have sat across from anyone qualified to assess them.
This is not a simple story about misinformation. The pain is real. That deserves to be said plainly and without condescension. People who turn to social media for answers about their mental health are often doing so because accessible, affordable, timely psychiatric care is genuinely difficult to obtain. In Canada, as in the United States, wait times for psychiatric assessment can stretch into months or years. General practitioners, however well-intentioned, are frequently under-resourced when it comes to mental health. The impulse to seek community and explanation online is not foolish — it is, in many cases, a rational response to a system that has failed to show up.
But the mechanism by which social media delivers those answers is fundamentally misaligned with good diagnostic practice. Algorithms surface content that generates emotional resonance and shares, not content that is clinically accurate or appropriately cautious. A video describing the “hidden signs” of a particular condition, framed around personal testimony and emotional authenticity, can accumulate millions of views precisely because it is compelling — because it makes viewers feel seen. Feeling seen is valuable. It is not the same as being accurately diagnosed.
The Hannover cases illustrate the clinical stakes with unusual clarity. Functional neurological disorder, which is what many of those patients were ultimately assessed as having, is itself a real and serious condition — one rooted in genuine distress and deserving of genuine care. The danger is not that these young people were suffering. They were. The danger is that a misidentified condition leads to mismatched treatment, and that years can pass before someone receives the help that would actually address what is happening to them.
There is also a subtler harm. When communities form online around shared diagnoses, those communities develop their own internal logic, their own hierarchies of suffering, their own tests of authenticity. To question a self-diagnosis within such a community can feel like an act of cruelty. The social pressure to maintain the diagnosis becomes entangled with the social pressure to belong. A person who might otherwise be open to reassessment finds themselves defending a label because abandoning it feels like abandoning their people.
None of this is an argument for dismissing patients who arrive at a clinical encounter with ideas about their own condition. Good psychiatric practice involves listening carefully to how a person understands their own experience — that self-understanding is itself clinically meaningful data. The problem is not that patients have opinions. It is that a platform optimized for engagement is a poor substitute for a trained human being who can sit with complexity, tolerate uncertainty, and resist the pull toward a tidy, shareable answer.
The solution is not to shame people off the internet. It is to build mental health systems robust enough that people do not have to go to TikTok for a diagnosis because no psychiatrist is available for six months. It is to fund public mental health infrastructure seriously, to train more practitioners, to reduce the financial and logistical barriers that make professional care inaccessible to so many. It is, in other words, a policy problem as much as a media literacy problem. Your pain is real. It deserves a real response — one that a recommendation algorithm, however sophisticated, is simply not equipped to provide.
