Study challenges reliability of EEG test for photosensitivity


The EEG strobe test produced astonishingly unreliable results in a recent study.

If the results of a recent study at SUNY at Buffalo are any indication, there are an awful lot of people vulnerable to visually induced seizures who are being told they aren’t at risk. The study showed that testing for photosensitivity using EEG with photic stimulation provides unreliable information. In the wake of these findings, ruling out photosensitive epilepsy–and ignoring the seizure risk from video games–simply on the basis of intermittent photic stimulation results would seem very unwise.

This study–not yet published–by the Buffalo research team found that the test is poorly correlated with vulnerability to visually induced seizures in everyday life. Just 6.2 percent of patients with a history of visually provoked seizures tested positive for photosensitivity.

The standard test for photosensitivity is intermittent photic stimulation—a strobe light flashing at specified frequencies—while hooked up to EEG. Abnormal waves provoked by the photic stimulation are known as the photoparoxysmal response (PPR).  (To avoid triggering a seizure during the procedure, the flashing is halted as soon as any of these abnormal waves appear.)

Researchers have known for a long time that many people who test positive never actually experience photic seizures. The Buffalo study confirms this:  of 86 patients whose EEG yielded a PPR, just 13 (15.11 percent) reported having experienced visually triggered seizures. What’s new here is the finding that many people who test negative do actually experience these seizures.

The investigators, led by Novreen Shahdad, MD,  initiated their study after a patient with a clear history of seizures provoked by electronic screens did not test positive for photosensitivity. “With the increasing popularity of video games and parental concern of their predisposition for seizures,” the authors wrote, “there is a need to identify individuals at risk for PIS [photic induced seizures].”


Find the above summary on page 14 of the 2012 SUNY at Buffalo School of Medicine and Biomedical Sciences Residents Research Day program schedule 

Shahdad and her colleagues examined a Buffalo EEG database and found 129 patients between 1999 and 2013 who reported seizures triggered by TV, computer use, and video games. Of those patients, a total of 8 tested positive for photosensitivity. Thirty of the 129 patients had reported a history of video game seizures. Of those 30, only one showed a photosensitive response on EEG!

Photosensitivity is defined by researchers as the appearance on EEG, in response to photic stimulation, of certain spike and wave patterns characteristic of epilepspy. Note that the criterion for photosensitivity is not the occurrence of visually induced seizures in everyday life—it refers only to the strobe test.

The study concluded:

“In contrast to the general impression, our study did not find a significant association of a positive response to photic stimulation in patients with photic induced seizures (PIS). This association was seen only in 6.2% of patients with PIS. In addition, PPR on EEG was not associated with statistically significant increase of PIS. Hence we conclude that PPR cannot be used as an isolated predictor for PIS.”

I’ve previously raised questions about diagnosing vulnerability to visually induced seizures soleley on the basis of EEG response to photic stimulation. More study is certainly warranted.

A continuum of visual sensitivity

Even readings outside the red zone may indicate the need for caution.

If a neurologist tells you that you don’t need to worry about seizures from electronic screen exposure, because you’re not photosensitive, what does that really mean?

It means that when you were tested for your response to a white strobe light, an EEG didn’t detect a particular abnormal electrical pattern in your brain. (I’ve noted some limitations of this procedure elsewhere.) Epileptology looks for yes or no, typically relying on EEG to rule out epilepsy. If yes, possibly medicate; if no, it’s not a case the clinician will pursue.

It does not indicate that bright flashing and/or patterns from electronic screens don’t adversely affect your brain function.

Researchers have gradually come to consensus on exactly what the EEG must look like to indicate photosensitive epilepsy (the photoparoxysmal response): certain spike/wave patterns that appear in both brain hemispheres. In arriving at these criteria, researchers excluded three other types of EEG abnormalities that in prior research “qualified” as a photoparoxysmal response. Epilepsy researchers aren’t certain what the significance of these other abnormalities is, but because the other patterns cannot conclusively be associated with epileptic seizures, there’s little interest in further research.

So these other EEG abnormalities from photic stimulation don’t count, in current neurology practice, and nobody would even tell you about them if they were found in your EEG. You’d be told the EEG was normal, period. But what if these other abnormalities were a sign that neurological function is in fact disturbed by visual stimuli, but not to the point of a seizure?

Let’s say you had one of the three other EEG abnormalities (which you wouldn’t know about, because the EEG was deemed normal). Maybe these indicate that you’re vulnerable to symptoms of a visual-overload-not-to-the-point-of-seizures syndrome. Neurologists have been examining the overlap between epilepsy, photosensitive epilepsy, and migraines. More about this in a future post, but actually there are many overlapping symptoms and correct diagnosis can be difficult. So if video game exposure or photic stimulation produces headaches and visual disturbances, and an inconclusive EEG, it may be that the visual overload is triggering migraines. Or perhaps the exposure is triggering another form of hyperexcitability in the brain’s visual cortex, which has been termed visual stress. While research has been done on this, it’s not part of a conventional neurology practice.

What about patients with more subtle or mood-related symptoms of a visual-overload-not-to-the-point-of-seizures problem? Who is treating these patients? Could be psychiatrists and psychologists, who view altered behavior and cognitive function through the lens of their respective training. Because there’s such a dearth of research of the gray areas of brain dysfunction following exposure to electronic screens, mental health providers have no basis for treating these patients for anything but mental health disorders. It’s clear that more research is needed and that more effects on the brain will be uncovered. One intriguing paper explores the contribution of fluorescent lighting to agoraphobia. The SpongeBob study published last year showed diminished executive function in children who viewed the cartoon.

In her Psychology Today blog, psychiatrist Victoria Dunckley recently posted a compelling piece about the effects on her patients of electronic screen time. She recommends creating a diagnostic category called Electronic Screen Syndrome to identify a dysregulation of mood, attention, or arousal level due to overstimulation of the nervous system by electronic screen media. She has seen dramatic improvements in hundreds of patients’ mood, behavior, and cognition after they go on an “electronic fast.” (Some have underlying psychiatric diagnoses, some don’t.) Maybe these patients were having very subtle seizures from electronic screens. Maybe the effects on the nervous system weren’t quite what epileptology defines as seizures. Either way, many kids exposed to electronic screens are experiencing diminished quality of life (as are their families) for a problem that medicine has not yet acknowledged.

Pediatricians miss a big opportunity

Last month the American Academy of Pediatrics Council on Communications and Media issued a policy statement entitled “Media Use by Children Younger Than 2 Years.” The document reaffirms the AAP position of discouraging media use in children under age two, a position first articulated by the organization in 1999. The council based its recommendation on three factors, including the potential adverse health and developmental effects of media use.

Authors of the policy statement missed a big opportunity to mention to the public that visually induced seizures from electronic media are a possible health consequence worthy of concern. Instead, the AAP considered and listed the usual physiological and neurobehavioral outcomes that are typically measured in studies regarding media exposure:  behavior, attention, obesity, learning, mood, language development, and sleep. With the exception of obesity, all of these aspects of health and development can be indicators of seizure activity.

Why not even a mention of seizures, a well documented health risk?  Probably because visually induced seizures are considered so rare by the medical establishment that the vast majority of studies on media effects on children don’t even raise the issue. Interestingly, the policy statement didn’t mention the risk of media addiction, either. Too controversial perhaps—incredibly, there is still plenty of academic posturing about whether excessive time spent playing video games constitutes addiction.

In August 2004 the Epilepsy Foundation of America convened a task force on photosensitivity to review the considerable research on visually induced seizures and to make recommendations about prevention. In a paper published the following year in Epilepsia, the group concluded its findings by stating, “a seizure from visual stimulation represents a significant public health problem. No known method can eliminate all risk for a visually induced seizure in a highly susceptible person, but accumulation of knowledge about photosensitivity is now at a level sufficient to develop educational programs and procedures in the United States that substantially will reduce the risk for this type of seizure.”  

Who was supposed to take responsibility for this message and follow up? The nation’s pediatricians haven’t even received the message yet.

In its paper the task force laid out some of the methodological complexities and obstacles to ascertaining how common photosensitive seizures really are:  studies are generally done on selected populations that don’t represent an average cross-section, researchers have had varying definitions of what qualifies as an indicator of photosensitivity on an EEG, the inability of EEG to pick up all seizures, variation in specific stimuli that individuals respond to, the greater intensity of today’s computer graphics, and so on. As I’ve mentioned previously—and the Epilepsy Foundation photosensitivity task force states this—the conventional photosensitivity prevalence numbers are clearly an underestimation. Until researchers find ways to more accurately account for the number of those who are at risk, and recognize those whose symptoms are less easily noticed, doctors in the US aren’t going to take the problem seriously, and little progress will be made to protect the unsuspecting public against visual media-induced seizures.

What affects your odds of visual sensitivity?

Photo courtesy

Do video game seizures worry you in particular? Should they?

Nobody knows the percentage of people whose sensitivity to flash and patterns could cause seizures. To find out with any statistical accuracy, researchers would need to do EEG testing with photic stimulation and patterns on large numbers of people. A population screening would be difficult due to sheer logistics and cost – attaching and removing EEG electrodes is a labor-intensive process as is properly performing the test. Pattern testing is rarely offered in the US In addition there are ethical considerations, since the photic stimulation could provoke a seizure.

Studies have found that 3 to 5 percent of epilepsy patients test positive for photosensitivity (whether or not they experience visually induced seizures). As I’ve noted previously, it’s really not known what percentage of the population without epilepsy (spontaneous seizures) is at risk for experiencing seizures induced by visual stimuli.

So, what known biological factors place you at higher risk for photosensitivity? These are things you can’t, in general, do much about:

  • Being female
  • Age 7 – 25
  • Parent or sibling with photosensitivity
  • Parent or sibling with febrile seizures
  • A specific form of epilepsy, juvenile myoclonic epilepsy
  • History of concussion
  • History of frequent headaches
  • Need for corrective eyeglasses
  • In those with epilepsy, a history of myoclonic, tonic-clonic, or absence seizures
  • Learning, behavioral, or psychiatric difficulties

These factors were ascertained in studies primarily by Graham Harding and Peter Jeavons in the UK and Dorothée Kasteleijn-Nolst Trenité in the Netherlands, leading researchers in photosensitivity and visually-induced seizures.

Note that studies measure the presence of a well-defined “photoparoxysmal” EEG pattern during exposure to photic stimulation (and sometimes, striped patterns). Its presence is a laboratory finding that does not invariably mean the test subject will experience seizures when exposed to flashing light and other visual provocation in everyday life.

Sensitivity in the same individual is affected by additional variables over which you have some control, such as fatigue, alcohol, distance away from the screen, etc.

How reliable is photic stimulation testing?

Most people who experience video game seizures have a condition called photosensitivity, whereby flashing or flickering light disrupts the brain’s normal electrical patterns and produces epileptic discharges. For these people a bright light flashing at certain frequencies (the number of flashes per second) can lead to the firing patterns of spikes and waves that occur during seizures. In the 1940s scientists discovered that a flashing strobe light can cause the brain to generate abnormal discharges that can be detected on EEG.

So as part of a routine EEG, a strobe light flashing at a range of frequencies is placed in front of the patient. The procedure is called intermittent photic stimulation, or IPS. IPS is the most widely used method of assessing susceptibility to visually induced seizures. The EEG and patient are watched closely, and if abnormal discharges or unusual symptoms appear, testing at that frequency stops right away, before a seizure develops.

Unfortunately, as with so much of the data used in seizure diagnostics, there is a lot of uncertainty about the meaning of photic stimulation test results. A negative result when testing for photosensitivity means simply that no seizure-like discharges were detected on EEG scalp electrodes and/or judged significant under the IPS testing conditions that were used. A positive finding may indeed point to vulnerability to seizures from flashing visual stimuli. However, in as many as 8 percent of the non-epilepsy population, particularly in children and adolescent girls, IPS produces a positive result that may have no clinical significance. A positive test for photosensitivity suggests (but does not prove) the following:

  • Because photosensitivity is characteristic of certain types of epilepsy, the patient may have one of those types. However, some people who never have seizures of any kind produce these discharges in the presence of a strobe light.
  • Since these abnormal discharges appear in response to a flashing light, the patient may be more likely to experience seizures when exposed to other sources of flash and flicker. The patient should be cautious about exposure to these visual stimuli

Compounding the weaknesses of EEG as a diagnostic tool for seizures, many procedural issues affect the accuracy of EEG results for determining photosensitivity. Results can be influenced by many things, including:

  • The type of strobe light used, including its shape and how diffuse the light is
  • The distance of the strobe light from the patient’s eyes
  • Whether the patient’s eyes were open or closed when the flashes and discharges occurred
  • Which flash frequencies were used (some people have an abnormal response only to a narrow range of frequencies)
  • The brightness of the flash
  • The wavelengths of the strobe light—the presence of certain colors in it
  • The assessing physician’s interpretation of the EEG—how abnormal are the altered wave patterns?
  • The assessing physician and the technician’s observations of the patient’s symptoms/behavior during the procedure
  • Whether the patient’s symptoms/behavior during the procedure are considered along with the EEG record

Some people who have seizures from video games do not show an abnormal response to photic stimulation. After all, the experience of playing a video game isn’t the same as viewing a pulsating strobe light. Other diagnostic tests for sensitivity to the visual experience of playing a video game — the movement of patterns and rapid fluctuations of colors — have been developed but are not widely available.

Update, 7/23/2014:  Please read here about a new study that found only 6.2% of patients with a history of visually induced seizures showed a positive result on EEG during photic stimulation.

People without epilepsy at higher risk for visually triggered seizures?

Pokemon photosensitive seizures and video gamesFor reasons that are not understood, people without epilepsy probably have a much higher risk for video game seizures than those with epilepsy. This is why awareness of the problem needs to spread from epilepsy community websites to reach people who don’t typically think about seizures.

When hundreds of children across Japan experienced seizures during a Pokémon cartoon broadcast in 1997, researchers became aware that the prevalence of seizures precipitated by visual sensitivity–particularly in those without classic epilepsy–is considerably higher than previously thought. The commonly cited statistics surrounding this event include the arrival of 685 children at hospitals for treatment of seizures. A more significant figure is that of the children treated in hospital emergency rooms for the Pokémon-induced seizures, 76 percent had no seizure history. While the study that provided this figure can’t be accessed online in English, you can see the abstract for another study with a very similar outcome.

Far more than those 685 children affected by the Pokémon broadcast were not taken to hospitals. Japanese researchers surveyed thousands of families to determine the number of children who experienced symptoms during the broadcast who did not require emergency care. During the program 1.4 percent of all the children who watched it had symptoms clearly indicating epileptic seizures and noted that this compares with 0.5 percent of the general population with active epilepsy. In addition, many other children–perhaps 10 percent of viewers–had experienced symptoms consistent with partial seizures that were probably not identified as seizures and were therefore not brought to the attention of doctors.

Based on these findings, one study suggests that the prevalence of visually induced seizures in individuals who do not have seizures at any other time is twenty times the rate of  those with epilepsy. This means that although having a relative with epilepsy may not make any difference in your vulnerability to video game seizures, if you have epilepsy yourself, your chances of also having visually induced seizures may be much lower than average! Another study looking at the Pokémon data suggests that visual sensitivity may be present in everyone, on a continuum. Anyone can be induced to have seizures with the administration of electric shock, for example, so perhaps that may be the case with visual stimulation as well. If a stimulus is provocative enough, it may produce seizures in people who ordinarily are not affected.

Like classic epilepsy, the tendency to have seizures from visual stimuli is an inherited trait. In people with both photosensitive seizures as well as epilepsy (spontaneous seizures), the two conditions are inherited separately. This means that while having a relative with epilepsy does not affect your chances of having a seizure from flashing lights and video games unless that relative also has the light sensitivity, having a close relative with seizures triggered by flashing light does make it more likely. Of course, it is hard to know for sure that anyone is not sensitive to flashing images–the underlying sensitivity may be there but may never have been activated.

Is the risk of video game seizures really one in 4,000?

genetic risk of photosensitivity and epileptic seizuresFor a lot of reasons researchers just don’t know how often video game seizures occur. The commonly used statistic of 1 in 4,000 people (and about five times that number in young people age 7-19) was carefully arrived at in rigorous clinical studies showing the prevalence of an abnormal EEG response to a flashing strobe light. It does not represent the prevalence of actual seizures triggered by today’s video game graphics.

Studies of photosensitivity began well before video games were invented. Researchers in the 1940s found that when exposed to a strobe light flashing at certain intervals, some individuals have a sudden change in their brain waves corresponding with the timing of the flash. They called this the photoparoxysmal response (PPR). The PPR does not indicate an actual seizure. The strobe is stopped as soon as the EEG changes so that the test subject doesn’t go on to have a seizure. Different researchers over time have used different criteria to define exactly what the altered brain wave needs to look like to qualify as a PPR. If your EEG shows one of these patterns in response to the strobe light, it doesn’t necessarily mean you will have light-induced seizures. It just means you’re more likely to. Playing a video game is a pretty different neurological experience than responding to a strobe light.

The obvious question is then, what is the value of the data if there is so little certainty about the meaning of the results?  I think that the EEG response to a strobe is simply the best tool that researchers have in that it provides measurable data that can be replicated in other EEG studies, and it allows relevant information to be captured without subjecting participants to seizures. It would be reasonable to be cautious around flashing lights if you’ve had an EEG that shows sensitivity to the strobe, but there’s no certainty you’d have a problem with video games.  And an EEG that does not show sensitivity to the strobe does not mean you’ve been given the green light, as it were, to play video games endlessly without needing to be concerned about seizures. Unfortunately the only way to find out if you’ve got a video game seizure problem is to encounter a game that triggers a recognizable seizure.

The point of all of the above is to give you some idea—without getting into lots of nitty gritty—of the squishiness and complexities of the data surrounding the prevalence of photosensitive seizures. In future postings I’ll talk about some other issues with the prevalence data. Next time I’ll discuss population sectors that are at higher risk.