In science, image transformation is rare

IIn 2004, two top editors of Cell Biology, Mike Rossner and Kenneth Yamada, wrote: editorial Alerting readers to an emerging problem in science: Thanks to Photoshop, researchers have tried to fraudulently embellish images in their manuscripts and eliminate peer review.

“Being accused of misconduct begins a painful process that can derail a person’s research and career,” they wrote, citing work published by the US Office of Research Integrity. two years agowhen the magazine and hoisted the flag. “To avoid this situation, it’s important to understand where the ethical line is between acceptable and unacceptable image editing.”

It is difficult to find wiser words.


In what might seem like an academic earthquake to many unfamiliar with the story, Stanford University this week investigating the president Concerns about the integrity of images in four published articles and for research misconduct.

But these cases are not so rare. According to the rollback control organization, there is one rollback per image change per day database Over 37,000 rejections and counting. Also investigative news about work Marc Tessier-Lavinwhich Stanford Daily Tuesday’s breakdown came as a surprise to those who had become accustomed to it PubPeer, a website where you can comment on scientific articles. There were critics flagging issues With images in Tessier-Lavigne’s article since 2015; The publications there have been published since 2001.


However, this revelation and other recent cases herald the public recognition of image manipulation as a serious scientific problem. Perhaps the most notable of these is the case Sylvain Lesne, an Alzheimer’s specialist at the University of Minnesota. Several of Lesne’s studies are now being examined after a.n the informant expressed concern about the pictures. “Some Alzheimer’s experts suspect that Lesne’s research has misdirected 16 years of Alzheimer’s research,” Science reported in July.

But the journal’s Editor-in-Chief Holden Thorpe’s commentary on the scientific article suggests that even some of the key players at the top of academic publishing seem to have slept through many of the alarms. “It would have been 2017,” Thorpe said [near] It was a time when we started to pay more attention, not only for us, but also for the scientific press.”

Such an explanation is thanks to Rossner, who shortly after accepting online submissions in 2001 began using “digital imaging experts” to display images of submitted manuscripts. What we were doing, and to convince them to make the same effort,” he told STAT of his time at JCB, which he left in 2013. “There was an effort to screen images before they were published that included dozens of people, including a lot of big players. . .”

“My disappointment is that I don’t know of any other publisher in the last 10 years that has started prescreening images,” Rossner said. Some even seem to have stopped, he added.

Instead, publishers entrusted image screening to unpaid professionals after publication. Among those who highlighted Tessier-Lavin’s work was the microbiologist Elizabeth Bick, who became one of the world’s most influential data scientists. In 2015, when questions were raised about Tessier-Lavin’s research, Bick was working in a Stanford lab unrelated to Tessier-Lavin, and was merely an image buff.

Not anymore. Bic moved into the mainstream, especially during the pandemic intended to do high-level work. He has more than 134,000 people Followers on TwitterThe New Yorker his presentation last yearand published by the New York Times written by him Last month, the headline read: “Science Has a Bad Photoshopping Problem.” Thank him unusually sharp eyes Journals have retracted nearly a thousand articles after discovering blatant misconduct.

Although Bick may be the best known, more collaborate with him on similar projects or specialize in finding and plagiarizing other literary issues statistical red flags It is careless, or outright fraud.

Meanwhile, researchers from various institutions have tried automation of image manipulation and copy detection, sometimes using artificial intelligence – that is, finding ways to do what Bick and others do, but on a much larger scale. The magazine’s publishers say they are joining forces to share the tools, but have yet to provide details.

The tools are currently missing validation. “If the technology really exists to do this well and at scale, it would be good to encourage more publishers to screen images before printing,” Rossner said. “They just need to know that it works.”

Just celebrating its 10th birthday, PubPeer has become a community of citizens with scientific concerns who can’t ignore criticism from journals, publishers, and institutions about not only images, but methodological, statistical details, and other issues of concern.

Of course, they often try. Consider the long timeline of the Stanford case. So far, the limited comments from the university suggest it’s following a familiar playbook for politicians: Ignore the problem until they can’t, and say it’s about the target before launching an investigation. A secret will be kept until one day everyone hopes to forget about the matter.

To date, none of Tessier-Lavin’s four articles have been corrected or retracted. Science, which published two articles on the issue, received a correction from Tessier-Lavin in 2015 but was unable to publish it “due to an error on our part,” Thorpe said in a statement.

According to Bick, the photoplay in the four articles under investigation is intended to “beautify” the data rather than to falsify or fabricate it. However, he believes fifth article Tessie-Lavin published in Cell in 1999 while at the University of California, San Francisco, there are signs of intentional and improper fraud, degree STAT reported. Bik, who now consults magazines and others on image ethicsHe told STAT that he “will testify in court” that the photo was “digitally altered.”

Since leaving Rockefeller University Press, which publishes JCB magazine, Rossner has been a full-time career consulting on image replacement and related issues. He’s disappointed with journals that don’t encourage him to screen images before publication, but says advances like PubPeer offer hope. They said, “Hopefully these issues will be resolved after publication.”

The private sector is taking notice. A company called Confirmation Researchers, journals and others involved in the publication stream will screen articles for signs of image doctoring to reduce “the risk of costly post-publication research and retraction”. A recent email from the company urged authors to “avoid scientific controversy and use Proofig.”

Of course, what is left unmentioned is not manipulating the image in the first place. Maybe it should be a more frequent message.

Adam Marcus, editorial director of Medscape, editor-in-chief of Spectrum, and Ivan Oransky, a distinguished writer at New York University’s Arthur Carter Institute of Journalism, are co-founders of Retraction Watch. Oransky is a volunteer board member of the PubPeer Foundation.

Source link

Leave a Comment