Watch Out for Brain Game Claims

TORONTO -- It's time to hit the reset button on recent headlines about 'brain games' potentially reducing dementia risk.

Many media reports missed the mark on a study presented here at the Alzheimer's Association International Conference that suggested a computer game focused on brain processing speed may be able to diminish dementia risk 10 years down the road.

The New Yorker wrote that "no brain game ... had ever been shown in a large, randomized trial to prevent dementia. That was the case until today."

TIME says the study provides "the strongest evidence yet for the power of brain training to reduce the risk of dementia."

Surely the 'brain games' industry is hopeful for some positive news after recent negative press. In 2014, about 70 researchers from Stanford and the Max Planck Institute signed on to a letter saying there wasn't enough evidence that any of these games worked. A year later, Lumosity paid a federal fine of $2 million to settle false marketing claims.

Those companies will get no redemption from these latest findings.

The data presented here come from a secondary analysis of a trial that wasn't powered to look at dementia outcomes. And results presented at the meeting haven't been through a rigorous peer review process.

The ACTIVE study, which was supported by the National Institute on Aging and the National Institute on Nursing Research, reported its 10-year outcomes in 2014.

ACTIVE was indeed a huge trial, randomizing 2,785 healthy adults ages 65 to 70 who didn't have cognitive impairment to one of three cognitive training interventions or to an inactive control.

Training programs targeted either memory, reasoning, or processing speed, and participants attended 10 sessions once a week for 5 weeks; though some could do more than that.

The study showed that early improvements in memory wore off after 10 years – but effects on reasoning and processing speed remained, with the latter being "remarkably strong," Jonathan King, PhD, of the National Institute on Aging who was an author of that study, told MedPage Today.

Speed of processing training involves strengthening the "unified field of vision," or being able to process at least two objects in one image, one of which is focal and the other in the periphery. The goal is to be able to divide your attention between the two.

"It's like the world's most boring video game," King said. "It can be difficult, and if you're the kind of person who wants to get faster and better you'll pay attention, but it's not enticing. There are no Pokemon running around."

The 10-year study didn't report dementia outcomes, because, again, it wasn't powered to detect them.

But King and colleagues retroactively developed a methodology to look at dementia risk over 5 years. In 2012, they reported in a secondary analysis that cognitive training didn't affect rates of incident dementia.

King also noted that 5 years may not have been long enough to capture conversion to dementia. So Jerri Edwards, PhD, of the University of Southern Florida, who had been involved in previous work on the ACTIVE study, applied a similar methodology to assess 10-year outcomes. (King did not participate in this study.)

Edwards found that processing speed was associated with a 33% reduced risk of developing dementia or cognitive impairment over 10 years (HR 0.67, 95% CI 0.49 to 0.91, P=0.012).

She said during a press briefing here that the findings for memory training and reasoning training were not significant, but she did not provide these data.

Dorene Rentz, PsyD, of Harvard, who moderated the oral session where the findings were presented, said Edwards didn't provide hazard ratios at all. King said this information came out during the press briefing.

"She just made very broad claims that I didn't think she backed up," Rentz told MedPage Today. "The whole presentation was like that. You were left walking away with your hands in the air not knowing what to make of it."

Edwards did not return calls for comment.

King said it's difficult to dismiss the data completely out of hand. Speed of processing training "was the largest training effect, and it did last 10 years," he told MedPage Today.

Although this latest analysis compared 700 controls and 700 in the processing speed group, it was not designed to be a dementia prevention trial in the first place.

"If it were, it would have to have had many more people," King said.

Announcing these data could have a big impact on uptake of speed processing training, which is currently available online to subscribers. Now known as Brain HQ Double Decision from Posit Sciences, it was originally developed by Edwards' mentor Karlene Ball, PhD, of the University of Alabama.

Ball sold certain rights to the program to Posit, which developed it commercially.

On its website, Posit puts an emphasis on the science behind its software – yet the program used in the ACTIVE study is different from the current version. The original technology was developed 14 years ago and the training was completed on a PC in a research center. The commercially available version looks different and can be done on any consumer device.

King said the technology should translate well, but technically the evidence doesn't pertain to this exact product.

Edwards was a consultant to Posit in 2008 and worked for another company that owned the intellectual property surrounding speed-of-processing training software. She also serves on the data safety and monitoring board of NIH grants awarded to employees of Posit.

Co-author Frederick Unverzagt, PhD, received a site license for cognitive training programs from Posit for investigator-initiated research projects.

King said peer review will have the final say on this secondary analysis of the ACTIVE study.

Slicing the data different ways could make the effect shrink, or disappear completely – but King said being left with even a 10% reduction in dementia would be something other researchers would want to follow up on.

Those follow-up trials should include a larger numbers of patients, as well as an official diagnosis of dementia (this was not recorded in ACTIVE; it was determined later).

"This was a very creative secondary use of the data using techniques that most would accept under the conditions," King said. "But in these intervention trials, the devil is in the details, and we have to make sure there's no hidden confounding influence that would be a threat to the validity of the findings."

Comment