The brain-training giant Lumosity is recalibrating its strategy and facing new challenges as it reels from a federal crackdown on bold health claims about its digital games.

The company behind the Lumosity brand, Lumos Labs, has dramatically cut back on TV advertising. It is facing sharp questions about its much-touted research, which found that users enjoyed a bump in IQ. And there are signs that the growth of Lumosity’s once impressive mobile app business may have stalled.

Meanwhile, the Federal Trade Commission is preparing to send out rebates to thousands of customers who may have been drawn to Lumosity by its misleading ads. The agency has verified 13,000 requests for refunds, which will be funded by $2 million from Lumosity, but has yet to disclose how much each customer will receive.

The FTC has moved against other digital health tools, too. In May, the marketer of the LearningRx suite of brain-training games agreed to a $200,000 settlement for making false claims about what the games could do to improve conditions like ADHD and Alzheimer’s disease.

But Lumos Labs is the biggest player, by far, targeted by federal regulators.

The privately held company, based in San Francisco, has drawn in 70 million people over the past decade to play games that challenge users to remember sequences of brightly colored animations, or to ignore visual distractions and click only on certain objects. The FTC charged that Lumosity oversold the benefits of the games. The company’s website, for instance, used to claim that the brain training could “ward off cognitive decline.” One ad featured a man who suffered from a stroke “and now uses Lumosity to regain lost mental function.”

TV advertising plunges

Since settling with the FTC in January, Lumosity has retrenched.

Lumosity has paid for only about $541,000 worth of TV ads so far this year, compared to about $11.8 million during the same period last year, according to the media research firm (The iSpot data reflect the list price of TV ad space, and don’t account for any discounts.)

It hasn’t used aggressive health claims for some time; the company’s biggest purchase of TV airspace this year has been for a commercial that urges viewers to “Take a 10-minute Lumosity fit test to challenge memory, attention, and more.”

A spokeswoman said Lumosity would be reevaluting its ad strategy in the coming months.

Meanwhile, Lumosity’s smartphone app may be struggling. Just 1.1 million users used the app in May, compared to about 1.8 million active users last November, according to an estimate from SurveyMonkey Intelligence.

And while Lumosity boasted of 2 million new downloads of its iPhone app in a single month in 2013, there were only 400,000 new downloads of the app across all Apple and Android devices last month, according to an estimate from app tracking firm Sensor Tower.

Lumosity spokeswoman Sara Colvin disputed the user number estimates, saying usage has been broadly stable all year, but she declined to disclose internal data that she first told STAT had been planned for public release around July.

To be sure, TV ads and mobile users offer just a snapshot of the overall performance of the Lumosity brand, which includes dozens of online and mobile games. So it’s hard to say how all the turmoil is affecting the company’s bottom line. Colvin said that “business is strong.”

Brain-boosting or placebo effect?

Lumosity’s promotions still emphasize its research—which may itself be on shaky ground.

Researchers at George Mason University reported in June that much of the benefit of brain-training games may be attributable to a placebo effect, something that many companies and university scientists fail to account for in their research. The study found that participants who were recruited for a study promising “brain training & cognitive enhancement” gained the equivalent of a 5- to 10-point IQ boost after an hour of playing the game, compared to no boost for participants who played the same game without the hype.

That calls into question a large study, conducted and funded by Lumosity and published earlier this year, that showed participants who played a cognitive training program performed better on measures of processing, memory, and reasoning than participants who had simply completed crossword puzzles.

The problem is that study participants were recruited from users who had created a free account on the brand’s website, a clearly self-selecting group whose expectations may have influenced the results of the trial, according to experts interviewed by STAT.

Colvin, the Lumos Labs spokeswoman, called the George Mason study “interesting” and said Lumosity welcomes “all research that serves to improve the quality of cognitive training research and, subsequently, the products available.”

Another recent study, funded by the federal government, did suggest that playing a specific kind of brain-training game could cut the risk of dementia years later, but the data have not yet been peer reviewed and some experts remain skeptical.

An industry barreling forward

These days, Lumosity is “focused on research and development,” Colvin said.

Lumosity just hired a new vice president with a digital health background, and the brand is testing products in the areas of mindfulness and sleep. This spring it launched a new suite of games designed to help players practice language skills like identifying synonyms and learning root words.

Despite Lumosity’s struggles—and the efficacy questions plaguing the industry—the type of games it pioneered are thriving. 

Peak, a two-year-old brain-training app, this spring surpassed Lumosity in active monthly app users and downloads, according to SurveyMonkey Intelligence. One of Peak’s slogans: “Meet the future of brain training.”

Republished with permission from STAT. This article originally appeared on September 6, 2016