Seizures from 2017’s best video games

What Remains of Edith Finch, Gorogoa, Everything, Night in the Woods, Divinity: Original Sin 2, Legend of Zelda: Breath of the Wild, PlayerUnknown's Battlegrounds, NieR:Automata, Persona 5, Super Mario Odyssey

The New Yorker’s Top 10 video games of 2017. Five contain scenes that could trigger seizures.

When you play a game ranked in the GameSpot or New Yorker Top 10 video games of 2017, the chances are about even that you will be exposed to images that could trigger photosensitive seizures. These images, which violate established guidelines for reducing the risk of photosensitive seizures, appear in 6 of the games in a combined Top 10 list. In 7 games these images were not found (some games are on both Top 10 lists).

What are photosensitive seizures?

Resident Evil 7, Horizon Zero Dawn, Wolfenstein 2: The New Colossus, Hellblade: Senua's Sacrifice

GameSpot’s 2017 Top 10 list includes 6 games that contain images that can trigger seizures.

Photosensitive seizures can occur when people with photosensitive epilepsy are exposed to intense visual stimuli: bright, rapid flashing light and bold patterns with strong contrasts. An unknown segment of the population has photosensitive epilepsy, including people with no history of seizures. It is under-reported and under-diagnosed.

In those who develop the condition, photosensitive epilepsy typically is hidden until the first noticeable seizure occurs in the presence of bright flashing or patterns. Most people with other types of epilepsy are not photosensitive. In other types of epilepsy, seizures are much more unpredictable.

Seizures can be of any type, from tonic-clonic episodes with loss of consciousness to brief absence seizures that can be as subtle as a brief hesitation or stare. Most people do not have photosensitive epilepsy, but many who do are unaware that they have the condition until a they experience a seizure during or after exposure to flashing or patterned images. Some individuals may have seizures that are too subtle to notice.

Screen grab from a sequence in The Legend of Zelda that could provoke a seizure in individuals with photosensitive epilepsy.

The seizure reduction guidelines test 

Guidelines for seizure reduction originated in 1994, when the UK adopted technical guidelines to accommodate TV viewers with photosensitive epilepsy. These guidelines, based on studies by photosensitive epilepsy experts, outline the characteristics of flash rates and spatial patterns that typically trigger seizures in people with photosensitive epilepsy. They were later updated and some have been adopted by the World Wide Web Consortium (WC3), the international group that produces website standards for all types of applications, and the International Telecommunication Union.

The guidelines define criteria for photosensitive seizure risk involving:

  • flash rate greater than 3 per second and less than 60
  • stripes and geometric patterns with high contrast
  • large areas of very bright (“saturated”) red
  • any of the above problem images taking up more than one quarter of the total screen area

Visuals adhering to these guidelines are unlikely to provoke seizures in 97% of people with photosensitive epilepsy.

Testing video games for compliance with guidelines

Although most games carry seizure warnings, not all games contain the types of images that can bring on seizures. The warnings are not specific to the content of a given game, so consumers who pay attention to the warning don’t know whether it pertains to the game they are about to use. So I test them.

I tested the games using downloaded clips of gameplay that I loaded into an application called the Harding Flash and Pattern Analyzer. The FPA is widely used by TV producers and networks in the UK—including by the BBC—to reduce the risk of seizures from material on broadcast TV, and is used by some game studios. It examines video sequences frame by frame for very specific and measurable image qualities that research shows can trigger seizures.

For more specifics on how to interpret the test results, please see this prior post. For more on my testing process, see this one.

Safe to play a game that passes the seizure test?

I've been using charts like this one to show results of testing for adherence to photosensitive epilepsy guidelines.

Until now I’ve posted charts in this format to show whether a game meets photosensitive epilepsy guidelines…

According to a Reddit post, a game that I’ve said “passed the seizure test” triggered a seizure. Recently the same game—Hearthstone—could have been implicated in a professional gamer’s seizure that happened during a live stream. What’s going on?

I write about games I’ve tested to alert readers to the games that don’t meet internationally recognized image safety guidelines. But I don’t want to create undeserved confidence that a game that passed the seizure test will be safe for anyone with photosensitive epilepsy.

Ian Hamilton, a user interface designer who specializes in and advocates for game accessibility, clarifies the role of testing this way:

“Passing the Harding test doesn’t mean that a game is safe. It means ‘reasonably safe’ because common triggers have been avoided. Something that gets a ‘pass’ can still absolutely cause seizures.”

I regularly write that your experience may differ, that I’m not trained in quality assurance, that I test excerpts of game play, and that health and lifestyle variables affect every individual’s vulnerability to seizures at any given time. Still, the meaning of my findings could be misleading without an understanding of the limitations of the seizure test itself:

  1. the pass/fail guidelines aren’t expected to prevent seizures in all individuals
  2. the test was designed for TV images, not video games

What the Pass/Fail guidelines mean

The guidelines originated in 1994, when the UK’s agency for regulating TV broadcasting (now known as Ofcom) inserted into its code of standards some technical guidelines to accommodate viewers with photosensitive epilepsy. These guidelines, based on studies of photosensitive epilepsy and consultation with Prof. Graham Harding and other photosensitive epilepsy experts, detail flash rates and spatial patterns that typically trigger seizures in people with photosensitive epilepsy. Specifications regarding saturated red images were added later, after the 1997 Pokémon incident in Japan.

Some compromises in the guidelines were made for the sake of practicality. Criteria for acceptable images (commonly referred to as the Harding test) were developed with the understanding that they would realistically protect most individuals with photosensitive epilepsy, but not all. For example, the guidelines permit images that flash at a rate of up to 3 times per second because flash at that frequency affects only 3 percent of photosensitive individuals. UK regulators decided that was “an acceptably small risk.”

I'm taking the word "safety" out of the chart since passing the test doesn't guarantee seizure safety

…but I’m updating all the charts by removing the word “safety” since passing the test doesn’t guarantee seizure safety.

The introduction to the guidelines states that their purpose is “reducing the risk of exposure to potentially harmful stimuli.” It also concedes that even when broadcasting images that comply with the guidelines,

“it is…impossible to eliminate the risk of television causing convulsions in viewers with photosensitive epilepsy.”

Applying TV guidelines to video games

There are no formal guidelines for reducing the seizure risk from video games. A 2005 consensus paper by experts on photosensitive seizures acknowledges that additional work would be required first on the existing guidelines for TV. In the meantime, it is reasonable to use the television guidelines since the impact of screen images on the visual system is the same.

The biggest challenge in applying TV specifications to video games is explained in the consensus paper:

“These principles are easier to apply in the case of fixed media (for example, a prerecorded TV show), which can be analyzed frame-by-frame. Interactive media, such as video games, may afford essentially limitless pathways through the game, depending on user actions. Therefore …in the case of video games, the consensus recommendations apply to typical pathways of play but cannot cover every eventuality of play.”**
Also,when players set their own viewing perspectives and preferences in newer video games, they may create unanticipated seizure risks.

Reducing risk going forward

In sum, a game that fails the Harding test is best avoided by those with photosensitive epilepsy. A game that passes is less likely to act as a trigger. Despite all the qualifiers, I believe there’s value in reminding people that seizures can happen to anybody, that certain video games can trigger them, and that you can lessen the risk by selecting games without lots of flash and patterns. Other strategies to lessen the risk of photosensitive seizures can be found here and here.

Tip of the hat to Ian, who suggested that I avoid the word “safe” when describing games that have passed the test. I also will be revising my prior posts to do some rewording.

Gamer’s seizure on live stream

Here’s a reminder that seizures can happen to anyone. A professional gamer known as Lothar had a seizure recently during his live feed while playing Hearthstone on Twitch. Lothar apparently has no history of seizures and the seizure may or may not have any connection to Hearthstone. In updates about his condition and hospital stay, Lothar didn’t mention photosensitive epilepsy nor has he said he’s been advised to limit his gameplay. 

For the record, Lothar is also a body builder—he’s obviously a guy who has enjoyed good health and takes good care of himself. Lothar has a large and caring following and has been receiving lots of well wishes as he recovers. Why do I mention this incident here? Viewing the incident (you can find it on YouTube) and seeing how it affected so many fans who care about him reinforced for me the seriousness of seizures and the importance of preventing those that are preventable.

** Graham Harding et al. (2005), Photic- and Pattern-induced Seizures: Expert Consensus of the Epilepsy Foundation of America Working Group. Epilepsia, 46: 1423–1425.

The royals in New York, as viewed from the UK

Arrival of the royal couple in New York, as shown on the CBS Evening News.

William and Kate arriving in New York, as seen on the CBS Evening News.

When the Duke and Duchess of Cambridge arrived at their New York hotel this week, it was big news.  Did you catch the squeals of delight from the crowd and the flicker of so many cameras flashing? The arrival scene was featured on a number of American news programs.


How it was seen in the UK

This same New York arrival scene was shown in the UK as well—but with an important difference.

This same scene was shown on the UK’s Sky News. Notice anything different?

Rapid-fire flash photography from the swarm of media creates a strobe-like effect on the image, which is known to trigger seizures in some people. In the US this material is broadcast despite the seizure risk.

But for more than 20 years, TV programs and commercials flashing at frequencies that can provoke seizures have been barred from broadcast in the UK. An exception to the regulation is made for breaking news coverage, where potentially seizure-inducing material is allowed to air if accompanied by a warning. These regulations were put in place after a noodle commercial triggered seizures in several British viewers.

Out of curiosity I tested a video clip of that New York sidewalk scene to see if the flash freqency falls into the category of a seizure hazard. I used an application called the Harding Flash and Pattern Analyzer (below) that assesses video sequences for seizure safety. The clip did in fact exceed flash safety limits for people vulnerable to visually induced seizures. Folks in the UK at least got an on-screen warning, which perhaps gives susceptible viewers a chance to avert their eyes.

The flash photography did indeed create a hzard for viewers with photosensitive epilepsy. A clip of the royals arriving in New York failed the seizure safety test.

News video of the royals arriving in New York did indeed fail the seizure safety test, as indicated by the green line above the pass/fail boundary on the graph. Screen is taken from the Harding Flash and Pattern Analyzer.

What happens when American TV programs are shown in the UK?

TV programs and music videos imported to Great Britain from the US (and other countries) need to be tested first, then modified to comply with the British seizure safety mandate. An engineer in a London post-production studio who does this compliance work shared with me that a popular US crime series had nearly 150 violations of seizure safety guidelines in a single episode. Where is the consumer lobby for seizure-safe TV programming in this country? Anybody?

Disenfranchisement of reflex seizures ending?

The new definition of epilepsy makes it clear that reflex seizures "count" as evidence of epilepsy. In previous definitions, this was not explicity stated.

The new definition of epilepsy makes it clear that reflex seizures “count” as evidence of epilepsy. In previous definitions, this was not explicity stated.

A new definition of epilepsy published this week affirms that photosensitive and other reflex seizures qualify as “real” epilepsy. This clarification may eventually help increase awareness of seizures from video games and other electronic media.

Although reflex seizures have long been included in official classification schemes of epileptic seizures, they don’t fit cleanly into established categories of seizure types and epilepsy syndromes. In neurology training they are typically mentioned only briefly. And typically they are taken too lightly by doctors using the prevailing diagnostic criteria for epilepsy: at least two unprovoked seizures at least 24 hours apart.     

Because reflex seizures, by definition, are provoked by specific triggers, there’s confusion and most doctors have been reluctant to diagnose epilepsy in people whose epileptic seizures require an environmental provocation.  The authors of the new definition paper acknowledge this:

Under limits of the current definition, [a] patient might have photosensitive epilepsy, yet not be considered to have epilepsy because the seizures are provoked by lights…People with reflex epilepsies previously have been disenfranchised by the requirement that seizures be unprovoked. The inclusion of reflex epilepsy syndromes in a practical clinical definition of epilepsy now brings these individuals into the epilepsy community.” 

Diagnostic criteria under the new definition now include at least two unprovoked or reflex seizures at least 24 hours apart. The new definition also allows an epilepsy diagnosis after a single seizure–either unprovoked or reflex–if there is a high probability of recurrence.

I’ve written previously about the inconsistency inherent in using the criterion of “unprovoked” to diagnose epilepsy in people whose seizures happen only in response to sensory triggers such as flashing light. This thinking (along with the assumption that photosensitive epilepsy is very rare) has led to marginalization of reflex seizures in the research community and among clinicians as well. Marginalization means doctors have been underdiagnosing reflex epilepsy, researchers seeking funding pursue other topics to study, and the public and public policy makers are largely unaware of the public health issue of photosensitive seizures.

The practical clinical definition was developed by a 19-member multinational task force of the International League Against Epilepsy (ILAE), incorporating input from hundreds of other clinicians, researchers, patients, and other interested parties. I’m more than pleased that the ILAE is choosing to make it clear that reflex epilepsy deserves the same respect as other forms of the disease (the new definition paper characterizes epilepsy as a disease rather than a disorder). It’s fortunate that the chair of the ILAE task force that produced the new definition is Robert Fisher, MD, PhD of Stanford, lead author of the 2005 consensus paper describing seizures from visual stimulation as “a serious public health problem.” No doubt Dr. Fisher’s appreciation of the magnitude of the problem was instrumental in ensuring that the task force addressed it.

Not all epileptologists agree with all aspects of the new epilepsy definition–and Epilepsia has given them a voice as well, publishing half a dozen commentaries, all of which are available free online. I contributed a piece as well, providing a patient/family perspective.

Of course, it remains to be seen how long it will take for neurologists to shift their attitudes and diagnostic practices regarding reflex epilepsy. Perhaps the inclusion of reflex seizures in the epilepsy definition will help dispell the idea that reflex seizures are rare.

Removing seizure triggers doesn’t spoil the fun

A short video clip on the website of Vanderquest Limited simultaneously shows a music video both before and after changes were made for photosensitive epilepsy safety. Read on for a link to it.

Gaming fans who object to the concept of seizure-safe games have some major misconceptions about how their favorite games would look after seizure-inducing visuals are removed.

I’ve seen people in online forums threaten to stop their game subscription if the game’s images are made safe. These folks are assuming that making safety modifications to the graphics will mean “neutering” the game experience in a way that ruins their enjoyment. In fact, very minor changes to images, which can be pretty hard to detect, are often all that’s needed.

Thanks to extensive research that defined the characteristics of images that induce seizures, it’s possible to make small changes that interrupt any such seizure-generating sequences. For example, after changes are made to some video frames, the flash interval–more than 3 flashes per second–which can lead to a seizure in people with photosensitive epilepsy, no longer induces seizures.

Want some proof?

NOTE: For those with visual sensitivity, before you click on the upcoming link to side-by-side videos, know that you should be able to view the page safely because the problem images aren’t very large. If you’d prefer not to risk it, you might want to step back farther from your screen or cover the left side of your screen, because that’s where the problem images are.

Now click to see the side-by-side images of a music video running in its original form (with safety violations) and after modifications were made to comply with seizure safety guidelines (in the UK, they’re requirements, not guidelines). The changes are very subtle! Can you find them? As you’ll see in the modified version, even visual sequences that contain some flashes and quick cuts can therefore pass the safety test. Basically, portions of some flashing sequences (9 instances) have been altered or removed, which does little to change the overall viewing experience. Edits to the video that allowed it to be broadcast were done within a single day by London-based post-production firm Vanderquest Limited.

Readers familiar with the Harding Flash and Pattern Analyzer that I routinely use to test video material for seizure safety will recognize the HFPA results shown beneath both versions of the music video. Although the original, uncorrected official version could not be broadcast in the UK due to regulations barring seizure-inducing material from being shown there, it lives on on YouTube.

Imagine these types of edits to the user interface of a video game. Exactly how do they ruin the gaming experience?! Or artistic expression?

Grand Theft Auto V looks seizure-safe

Grand Theft Auto V passed the seizure safety assessment test.

Grand Theft Auto V passed the seizure safety assessment test. This scene got close to the safety limit but stayed within it.

Say what you like about the just-released Grand Theft Auto V, but it’s not likely that the game will trigger seizures. You’d think an action-adventure game like this would be full of bright screen flashes that pose a seizure risk. I tested a lot of trailers and game play clips and didn’t find any scenes with unsafe-for-seizures visuals. Your results could vary, of course.

But I’ve got a hunch the folks at developer Rockstar Games are making the effort to make the game safe. In a few instances the flash level got close to the safety limit, but didn’t go past it, which makes me think the developers know where the official limit is (3 or more flashes per second) and have decided to respect it. While it’s possible that someone could, under the right circumstances, experience a seizure even when the flash level is near but below the danger zone, the chances of that happening aren’t high.

Unlike many immersive games with missions and quests, GTA isn’t set in an exotic fantasy world, where art directors might claim the need for artistic freedom (to use whatever visual effects they deem necessary). It’s not full of bright, flashing explosions. The appeal of the game is in the very down-to-earth characters, story lines, and realistic settings–not the quality of the visual experience.

Even a scene like this in GTA IV passed the safety test (in the caution zone).

Scenes like this in Grand Theft Auto IV (2008) also passed the safety test because they did not employ quick sequences of alternating light and dark images.

Grand Theft Auto has evolved quite a bit since its beginnings as an arcade shooting game in 1997. The visual style of earlier versions, from 2001 GTA III and prior, is noticeably more jarring: jumpier, more flash, and indeed, a GTA III trailer failed the safety test.

Guidelines for seizure-free video sequences were developed more than 20 years ago in the UK in response to seizures provoked by TV. In 2005 the International Telecommunication Union, an agency of the United Nations, published recommended guidelines for reducing photosensitive seizures from television. These guidelines for safe flash rates and pattern movement could be applied to any screen-based media. I tested downloaded trailers for compliance with the safety guidelines using an application designed for this, the Harding Flash and Pattern Analyzer.

GTA III (2001) contained flickering scenes such as this one that failed the seizure safety guidelines because of rapidly alternating lighting levels.

GTA III (2001) contained flickering scenes such as this one that failed the seizure safety guidelines. Rapidly alternating light, from dark (screen above) to bright (screen below) and back, create a strobe-like effect.

In its first three days after its launch last week, the game grossed a billion dollars in sales, setting a record for any kind of entertainment release. So much for the argument that making games safe will spoil all the appeal and fun.

A well lit screen alternating with the darker version above can trigger seizures if the flicker occurs 3 times per second or more.

This brightly lit screen alternating with the darker version can trigger seizures if the flicker occurs 3 times per second or more.

Actually, maybe revealing that the game won’t cause seizures is not good PR for the company. Somehow a safe game doesn’t seem in keeping with GTA’s guys-will-be-guys attitude. Maybe Rockstar would prefer that customers don’t realize the game seems to lack seizure-provoking images. The average GTA player may not want to feel that the game has been “softened” in any way. In GTA culture, playing a seizure-safe game is probably for wimps.