What is image manipulation?
As career growth concerns, competition for funding, and the ever-growing “pressure to publish” continue to take the scholarly community by storm, researchers become tempted to commit image manipulation.
Image manipulation involves altering or falsifying visual data, such as images, graphs, or figures, to present misleading or fabricated results.
While certain image adjustments, like cropping, resizing, or color correction, are sometimes necessary for scientific purposes, new dangerous forms of image manipulation have emerged.
Researchers have begun to selectively edit images to omit crucial data points or exaggerate effects, or engage in duplicate publication by reusing the same image to represent different results. These practices distort visual data and warp readers’ understanding of scientific thought.
Recent advancements in image-editing software, including the rise of AI have made it easier for researchers to manipulate images subtly, making it difficult to identify alterations with the naked eye.
Integrity scandals at Stanford
Stanford University President Marc Tessier-Lavigne has recently found himself in a storm of controversy over concerns surrounding the integrity of his research on brain degeneration and diseases like Alzheimer’s.
The allegations of misconduct were first sparked on PubPeer, a website on which scholars throughout the academic community discuss their research and papers.
Once the investigation results were released, Tessier-Lavigne claimed he was cleared of any fraud, yet it was found that researchers under him engaged in inappropriate manipulation of research. He has since announced his resignation.
Research misconduct expert Elisabeth Bik first reported that Tessier-Lavigne’s papers contain “serious issues,” including altered imagery. “He should have supervised better. As a senior author in scientific papers, you're not the person standing in the lab, but you're ultimately responsible for the integrity of the work," she said.
What can we take away from this?
This news, coming from the head of a renowned American university has thrust image manipulation and the crisis of integrity within our community onto the national stage, underscoring the urgency for more effective fraud detection measures to curb such issues before they gain this kind of attention and cause such damage.
Confronting fraud in Alzheimer’s research
Just last year, the scholarly community was shaken when a major scandal surrounding Alzheimer’s research was revealed to the public.
Integrity sleuth, Matthew Schrag, reviewed images from an influential study on Alzheimer’s published in 2006 by first author Sylvain Lesné. After Schrag found evidence of fraud and tampering, including issues with hundreds of images, his findings were confirmed by several top Alzheimer’s researchers.
Elizabeth Bik weighed in on this issue again, stating that, “[T]he authors appeared to have composed figures by piecing together parts of photos from different experiments. The obtained experimental results might not have been the desired results, and that data might have been changed to … better fit a hypothesis.”
The cascade effect of this image manipulation was profound. Researchers had built their discoveries on these findings for the past 16 years. This incident revealed how easily image manipulation can misdirect and warp scientific thought for years, all while remaining undetected.
A Nobel Prize winner's dark secret
How can we trust the world’s latest breakthroughs when even the work of Nobel Laureates raises suspicion?
Last fall, Gregg Semenza, a Johns Hopkins researcher who shared the 2019 Nobel Prize in Medicine or Physiology, retracted four papers due to concerns surrounding alleged image manipulation. A research integrity sleuth, Clare Francis, took notice of several instances of image duplication across the researcher’s papers.
Before the retractions, the articles had already made a significant impact, being cited more than 750 times. This is another example of how pervasive image manipulation can be within the scientific community.
Conclusion
Do you remember that viral image of Pope Francis strolling down the street, donning a massive white puffer jacket? This AI-generated photo took the internet by storm, captivating audiences on platforms like Twitter, Facebook, and Reddit earlier this year. Astonishingly, many were convinced of its authenticity.
Web culture expert Ryan Broderick labeled this as "the first real mass-level AI misinformation case."
According to Retraction Watch, incidents of image manipulation retractions occur about once every other day. Despite its widespread prevalence, integrity sleuths often beat publishers to recognizing this issue. Journals must invest in robust pre-screening methods to get ahead of the curve and confront this problem at its core. Stay tuned for more updates on our progress in tackling image manipulation head-on as part of our growing set of integrity checks.
