The Pokemon Shock

Summary

On the evening of 16 December 1997, a children's cartoon was broadcast across Japan. Twenty minutes in, the screen filled with red and blue flashes, alternating about twelve times a second, for a few seconds.

Six hundred and eighty-five children were taken to hospital by ambulance that night.

It is one of the most-cited incidents in the history of broadcasting, and almost every account of it stops at that number. The number is real. What it means is not what people think, and the reason is the most important thing in this record.

Not all of those children had seizures, and how many did has never been settled. And many of the children who fell ill that week, as Radford and Bartholomew have argued, did so after watching the news report about the children who had.

What is documented

The broadcast. Episode thirty-eight of the Pokemon animated series, titled Dennou Senshi Porygon, was broadcast at 6:30 pm on 16 December 1997 on TV Tokyo and its affiliates.

The sequence. Approximately twenty minutes into the episode, a scene depicted an explosion using a technique in which the full screen alternates rapidly between two saturated colours, red and blue. The alternation ran at roughly twelve cycles per second and lasted several seconds.

The immediate effect. Viewers, overwhelmingly children, reported dizziness, nausea, headache, altered vision, and in some cases convulsions.

The hospitalisations. Six hundred and eighty-five children were taken to hospital by ambulance. A survey by Japan's Fire and Disaster Management Agency found that 208 were admitted.

Photosensitive epilepsy. A minority of the population, particularly children and adolescents, have a susceptibility in which flashing light at certain frequencies can trigger a seizure. The frequency range that is most provocative is roughly fifteen to twenty-five flashes per second, and susceptibility extends below that.

The Porygon sequence sat squarely in a provocative range.

And here is the difficulty. How many of the roughly seven hundred children taken to hospital had genuine photosensitive seizures has never been settled; one follow-up study tracked 103 who did. The five to ten per cent figure often attached to this case measures something else: the share of all child viewers who reported any symptom, such as headache or dizziness, in later Japanese surveys.

Many had milder symptoms that were not seizures.

The second wave, which is the part that matters. According to Radford and Bartholomew (2001), a substantial number of the affected children became symptomatic after the incident was reported in the news.

Japanese television covered the story that night and the following morning. Some coverage described the sequence. Some replayed it, or replayed portions of it. Children who had not seen the original broadcast, or who had seen it without effect, watched the news and then fell ill.

The response. The episode has never been rebroadcast, anywhere in the world, in any territory. Pokemon was suspended for four months. Japanese broadcasters adopted formal guidelines restricting the rate and intensity of flashing imagery, similar to guidance the United Kingdom had introduced after a 1993 television advertisement triggered seizures.

Leading explanations

A real physical trigger, and, as Radford and Bartholomew have argued, a psychogenic wave on top of it. The first is established; the size of the second is disputed. The two must be kept apart.

The physical part. Photosensitive epilepsy is real, well characterised, and was genuinely triggered in a number of children by a sequence that would today be prohibited by broadcasting rules that exist because of this episode. Those children had seizures. The seizures were neurological events. Nothing about them is disputed.

The psychogenic part. Many of those who fell ill did not have seizures. They had headaches, nausea, dizziness and distress, in a population of frightened children whose parents were frightened, in a country where the story was on every channel.

And the vector. On this reading, the news coverage propagated it: children who had been unaffected by the cartoon were affected by the reporting on it. Some replays also triggered further genuine seizures, which is one more reason the two parts must be kept apart.

On that reading, it is one of the cleanest demonstrations of media-mediated mass psychogenic illness ever recorded, precisely because there is a hard, physical, undeniable seed event underneath it.

Current status

Later explained. A physical trigger that caused real seizures, followed, as Radford and Bartholomew have argued, by a wider wave of psychogenic illness spread by the news coverage. The size of that wave is disputed.

The registry files this with AR-0037, AR-0092 and AR-0094, and it is the most useful of the four, because it is the only one with a hard physical event at the bottom of it.

In Mattoon, nothing was in the air. In Seattle, nothing had touched the glass. In the War of the Worlds, nobody panicked.

Here, something really did happen. Children really did have seizures. And even so, and precisely because something real had happened, the reporting on it appears to have added cases of its own.

That is the mechanism this whole archive is about, and this is the case where it comes closest to being measured.

Last reviewed: September 2026. Records are provisional. Where the evidence changes, the entry changes. Found an error? Tell us.