Photosensitive Seizures Induced by Animated TV Program called “Pocket Monsters” and Long-Wavelength Red Light from CRT

Purpose: In December 1997, an animated TV program called “Pocket Monsters” induced photosensitive seizures simultaneously over widespread regions of Japan, and several hundreds of viewers were rushed to emergency hospitals. We investigated factors involved in determining why only some photosensitive individuals had seizures while watching the program.

Methods: Cathode ray tube (CRT) emissions were evaluated for 7 patients with induced seizures in response to the program (induced patients) and 4 photosensitive patients without seizures in response to the program (photosensitive controls) using a spectroradiometer. The color bars televised by TV-AICHI stations that aired the program in the Chukyou district were recorded on videotape via receiving systems in the patients houses. The recorded color bar was played on each CRT in a darkened room of the patients houses, and in blue, green, red, white, and black images of the color bar, luminance energy was measured by spectroradiometer. Testing was performed l hour after turning on the CRTs. Luminance energy at 4 subdivided ranges: 380-480 nm (blue range), 480–590 nm (green range), 590–720 nm (red range), and 720–780 nm (long-wavelength red range) were integrated from the spectroradiometer data.

Results: Black color bar emissions showed characteristic longwavelength red light patterns. All CRTs of induced patients emitted light containing long-wavelength red light at a higher rate than the CRTs of photosensitive controls. CRTs of patients 1,5, and 7 presented long-wavelength red light at higher rate (patients 1, 5 and 7), and the characteristic emission extended to the red range. CRTs of photosensitive controls lacked the characteristic emissive pattern, and only the CRT of patient 9 emitted a few long-wavelength red light. Each CRT of induced patients and photosensitive controls seemed to have almost consistent emission level of long-wavelength red light independent of color of images. The mean amount of long-wavelength red light emitted from CRTs of induced patients was significantly more than that in CRTs of photosensitive controls in blue (p = 0.02, Mann-Whitney U), green (p = 0.04), and black images (p = O.Ol), but was not significantly more in red and white images. The mean amount of total visible rays emitted from CRTs was not significantly different between CRTs of induced patient and CRTs of photosensitive controls. The mean ratio of long-wavelength red range to total visible range was significantly higher in CRTs of induced patients than that in CRTs of photosensitive controls in every color of images (P= 0.008–0.023, Mann-Whitney U). The mean contrast of luminance energy between red and blue images in CRTs of induced patients was not significantly different from that in CRTs of photosensitive controls, at blue, green, red, and long-wavelength red ranges.

Conclusions: The emission of long-wavelength red light from CRTs played a decisive role in induction of seizures in photosensitive patients. Long-wavelength red light components from CRTs 011 flashing images (redlblue) of the program correlated strongly with subsequently induced seizures. CRTs with high long-wavelength red light emission stimulated wavelength-dependent pathophysiological mechanism and might have increased the level of photosensitivity of photosensitive viewers by everyday use of such CRTs. We must consider limiting long-wavelength red emission from CRTs in order to prevent photosensitive seizures produced by watching TV.