It’s a rare thing for a scientist to stand up in front of a roomful of his peers and rip apart a study from his own lab. But that’s exactly what Vincent Walsh did in September at a symposium on brain stimulation at the UC Davis Center for Mind and Brain. Walsh is a cognitive neuroscientist at University College London, and his lab has done some of the studies that first made a splash in the media. One, published in Current Biology in 2010, found that brain stimulation enhanced people’s ability to learn a new number system based on made-up symbols. Only it didn’t really. “It doesn’t show what we said it shows; it doesn’t show what people think it shows,” Walsh said before launching into a dissection of his paper’s flaws. They ranged from the technical (guesswork about whether parts of the brain are being excited or inhibited) to the practical (a modest effect with questionable impact on any actual learning outside the lab). When he finished this devastating critique, he tore into two more studies from other high-profile labs. And the problems aren’t limited to these few papers, Walsh said, they’re endemic in this whole subfield of neuroscience.