As the integrity of the scientific research community becomes increasingly suspect — with MIT professor Luk Van Parijs losing his job over fraudulent data last year and Dr. Hwang Woo Suk’s human embryonic stem cell scandal in South Korea — a new test adopted by The Journal of Cell Biology has found a surprising amount of researchers have fraudulently doctored research images in Adobe Photoshop.
Dr. Suk shocked the scientific community when he was charged with heavily distorting images he used in his research on human embryonic stem cells, a study that claimed his team had successfully obtained stem cells from cloned human embryos
At The Journal of Cell Biology, the test has found a significant amount of manipulation in research photos. Since 2002, when the test was implemented, 25 percent of all accepted studies have had one or more images that were distorted in ways that breach the journal’s rules, said Michael Rossner, the executive editor.
Ira Mellman, the editor of the journal, said that most cases did not result in fraud when the authors provided the original copies of the images. “In 1 percent of the cases we find authors have engaged in fraud,” he said.
Boston College biology professor Daniel Kirschner said altering a scientific image for aesthetic purposes takes away from the credibility of the image.
“Getting rid of blemishes and changing individual spots instead of darkening or lightening the whole research image would be totally unethical – it would be cheating,” said Daniel Kirschner, a Boston College biology professor.
Kirschner, who uses Photoshop to analyze cell structure, said images are an important part of his research projects and he prepares all the figures for his research with various computer programs, including excel.
“I, in fact, do change contrast, but I don’t do it to hide the data as I know some other people do,” he said. “I’m always in favor of showing the rawest data possible in publications.”
He said some researches opt to cut out cells in the images they study and leaving out other blemishes in unclear images, to “clean up” their data.
“I had a colleague who was using Photoshop to get rid of problems and get rid of background blemish,” he said. “I said you can’t do that – if the blemish is there you can’t delete it.”
Kirschner said there are no written regulations in his lab as to what researchers can and cannot touch up, but he supports scientific journals that verify the authenticity of their research images.
“It’s too bad that they have to do it,” he said. “Unfortunately some people are using these tools inappropriately and the journals are right to make sure the data hasn’t been modified.”
Mike Rossner, an editor at the Journal of Cell Biology, works to detect these scientific photo manipulations.
“This can reveal inconsistencies in background that are clues to manipulation,” he said in an email. “For black and white images we adjust brightness and contrast and for color images.”
He said Dartmouth researcher Hany Farid has developed mathematical analysis methods that involve algorithms to detect manipulation in digital images.
According to Rossner, some examples of “fraudulent manipulations” include adding or deleting cells to or from an image and adjusting the intensity of a specific area of an image.
“I think the problem has probably existed since applications like Photoshop became universally used by scientists for preparation of figures for publication,” he said.
Rossner said the Journal began to check figures when their system became fully electronic and many researchers welcome the surveillance of their data.
“Authors seem to appreciate our vigilance in ensuring that image data are presented accurately,” he said.
John Celenza, a Boston University biology professor, works with plants and has published a number of photos of plant tissue with digital and film cameras.
“If I wanted to misrepresent my imaging data, I could have done that before Photoshop,” Celenza said. “In any generation, it really comes down to the integrity of the researcher.”
Although many of his research photos are cropped and resized, Celenza said he tries to stay consistent with his camera settings and he works to convey a valid image.
“When composites are made, borders are left around individual samples so that it’s clear that different images are being used,” he said. “These adjustments are not so extreme as to mask details but are done to accentuate the details of the experiment or are for aesthetic reasons.”
Celenza said he thinks some cropping is permissible to remove background that is not directly related to the subject being photographed but not if it removes or hides data.
“This is not new to the digital age,” he said.
