On January 14, faced with a deadlocked jury, a federal judge in California declared a mistrial in a consumer class action involving the marketing of Prevagen, a popular dietary supplement based on jellyfish-derived proteins that claims to improve brain functioning and memory. This outcome runs counter to the conventional, but mostly untested, viewpoint that juries tend to favor the plaintiffs in consumer class actions. The Prevagen trial also underscores that scientific uncertainty about the truth of an advertising claim may present challenges for the defense in the earlier stages of a class action, but become an advantage for a defendant who chooses to fight all the way to trial.

Prevagen’s sales—more than $165 million over the last decade—have made it a target for the plaintiffs’ bar and regulators alike. The California lawsuit, Racies v. Quincy Bioscience, LLC, is one of several around the country challenging Wisconsin-based Quincy’s representations that Prevagen “improves memory within 90 days” and “supports healthy brain function, sharper mind, and clearer thinking.” Others include an enforcement action jointly brought by the Federal Trade Commission and New York Attorney General, which was reinstated by the Second Circuit Court of Appeals last year after the district court had granted dismissal.

But Racies itself is, in a sense, a unicorn: It is one of the few false advertising class actions to have made it all the way through trial, and, at least in recent memory, one of the only in federal court to have gotten this far. (Another, Lewert v. Boiron, Inc., was also in California and resulted in a verdict for defendants.)

The week-long trial in Racies followed extensive motion practice over the measure of Quincy’s potential liability and the nature of the proof. In 2015, District Judge Haywood S. Gilliam dismissed plaintiffs’ claims insofar as they challenged Prevagen’s “clinically tested” representation, reasoning that private plaintiffs are barred from bringing lack-of-substantiation claims under California’s consumer protection statutes. The Second Circuit allowed similar claims to go forward in the related FTC/New York AG action based on allegations that the single study cited by Quincy showed no statistically significant improvement from taking Prevagen.

Principally at issue in Racies is the efficacy of Prevagen’s sole active ingredient, a protein called apoaequorin that was originally discovered in jellyfish. Plaintiffs and their expert, Dr. Richard Bazinet of the University of Toronto, argue it is biochemically impossible for apoaequorin to directly impact the brain because it is digested into amino acids before it enters the bloodstream, like any other protein. Moreover, plaintiffs contend that even if some apoaequorin were to survive the digestive process, it would be incapable of crossing the blood-brain barrier because it is too large. Quincy and its experts have challenged Dr. Bazinet’s failure to specifically test how apoaequorin is digested and his reliance on general scientific principles of protein digestion. They also claim that apoaequorin may be digested into small peptides that are capable of passing through the blood-brain barrier.

In 2016, the district court rejected Quincy’s attempt to exclude Dr. Bazinet’s testimony, finding that he was qualified and that the general principles he intended to offer were supported by extensive scientific literature. The court also denied the parties’ cross motions for summary judgment, because it was not necessary for Dr. Bazinet to test apoaequorin but a reasonable jury could still refuse to follow his deductions.

In a somewhat unusual twist—and contrary to conventional wisdom that the plaintiffs’ bar will fight tooth and nail for a jury—plaintiffs then moved to drop their damages claim and only seek equitable restitution under the UCL in the hopes of having a bench trial, over Quincy’s objection. Late last year, the district court denied that motion in a short order, citing the fact that both sides had previously demanded a jury trial, Quincy continued to do so, and plaintiffs had not requested leave to formally amend their complaint.

Most recently, the court denied plaintiffs’ attempt to reopen discovery on the eve of trial to obtain further testimony about Quincy’s sales records. Quincy had argued that plaintiffs’ damages model was premised on wholesale sales data, not retail sales, and thus could not show how much consumers buying at retail supposedly paid (or overpaid) for Prevagen. While plaintiffs claimed not to realize the records were for wholesale sales, the court found they were clearly on notice and ordered them to try to prove damages using the information they had.

After three days of deliberation, the eight-member jury informed the judge that they were “hopelessly deadlocked” and disagreed on “fundamental issues” in the case. Both sides have moved for judgment as a matter of law. Plaintiffs argue that Quincy failed to challenge Dr. Bazinet’s opinions, which found support in letters from Quincy and its counsel to the Food and Drug Administration (FDA). On damages, plaintiffs argue that Prevagen is “worthless,” entitling the class to restitution of the full retail price, and the wholesale sales data is a conservative measure. For its part, Quincy focuses on the absence of retail sales evidence and class-wide reliance.

That plaintiffs’ claims reached a trial in the first place is unusual. But it is even more unusual—and telling—that the result was a mistrial. Racies followed on the heels of the related FTC/New York AG action, which alleged that Quincy’s representations were contradicted by the only study it cited in support. Plaintiffs also had considerable evidence in the form of letters from Quincy’s regulatory counsel to the FDA stating that apoaequorin would likely be digested into amino acids “that will be absorbed in a process similar to other dietary proteins”—effectively the same as Dr. Bazinet’s testimony at trial. Indeed, in denying summary judgment to Quincy, the district court noted plaintiffs had “more than enough evidence” and “a substantial argument.” And, of course, plaintiffs had the good fortune of being in front of a jury in California, a forum widely understood as friendly to consumers.

Even so, plaintiffs were unable to claim victory in front of the jury. Part of that is likely the result of suing based on the unknown. Dr. Bazinet did not specifically test apoaequorin, and different jurors may have approached scientific uncertainty with different assumptions. While we might never know what “fundamental issues” divided the jurors, it may well have been the thorny question of determining when a statement is “false” when the science is arguably unresolved, or the dubious proposition that a product is totally “worthless” unless the science supporting its marketing is clearer. Still, it is hard to explain the result in Racies without rethinking the traditional notion that reaching a jury in these sorts of cases means game over for defendants.