Improving image integrity in academic research

Dror Kolodkin-Gal explains how image integrity issues occur, their consequences and how to prevent them

Scientific images allow researchers to share their ideas in an engaging way while elaborating on complicated phenomena. Any images used in papers must be accurate to avoid misunderstanding, misinterpretation and allegations from readership. The scientific community understands the importance of image integrity, but issues such as duplication continue to be found in publishing. According to leading image data integrity analyst Jana Christopher, the percentage of manuscripts flagged for image related problems ranges from 20 to 35 per cent[1].

According to Christopher: “The focus on images has intensified over the last few years. These days, a growing number of journals perform regular checks, and we have a community of post-publication reviewers and whistle-blowers who will, often anonymously — flag image irregularities on PubPeer and on social media.” 

The impact of integrity issues

Most of the time, image integrity issues are unintentional mistakes of the researchers handling with plenty of similar images. However, unless they take the time and the proper software tools to detect and resolve any issues before publication and the editor does not flag it, these issues could negatively impact the scientists' and journal’s reputation.

Failing to detect image integrity issues before submission, either for grant requests or publication, can result in rejection. If a grant authority rejects a submission, it can delay access to funding, halting research. Alternatively, publishers do not need to disclose a reason for rejection during the peer review process, which makes it difficult for researchers to understand how they can improve the probability of publication elsewhere. 

If an issue remains undetected until reported post-publication, either to the journal or online, the publisher must investigate to determine if the allegation is true, how it occurred and how to resolve it. Investigations can take years, which is hard for researchers. During this time, they may find it difficult to generate further funding, conduct research or publish elsewhere. Consequently, regardless of the  investigation’s outcome, researchers must work hard to rebuild their reputation.

Image integrity issues can also damage future research elsewhere. Academics often base new research on existing papers — if the original submission contains inaccuracies, any data in new research will also be incorrect. Researchers may also find it difficult to replicate original results if they base their experimental procedures on an existing paper that contains errors, leading to more wasted time, materials and funding.

The cause

Jana Christopher’s research highlights the frequency of image integrity issues — to resolve them we must understand how they occur. It can be difficult to detect forms of image duplication such as overlaps or flips of complex images when checked manually — resulting in unintentional duplications.

Duplication means reusing the same image in different parts of the paper without outlining its usage. This can occur when an image is used the same way twice, or has been altered, for example, by changing the rotations, size or scale. The image may have also been flipped or cropped during duplication, or researchers may use two parts of the same specimen, but they overlap.

Such duplications are common because researchers will often collect thousands of images of specimens while conducting research, either for their own paper or for collaborative research. If these images are not properly managed, it becomes difficult to distinguish between the different files, meaning the risk of unintentional duplication will only increase.

Regular, proactive image checking     

If researchers and editors review their images manually, many of these issues, such as duplications, may go unnoticed, meaning there is no guarantee that they will detect them before submission and publication. Manually checking images is time consuming and is vulnerable to human error. On the other hand, automating image integrity checks gives researchers and editors the peace of mind that they are sharing credible data.

Advancements in artificial intelligence (AI) and computer vision has led to the development of valuable tools for scientists that researchers and publishers can use to check content for grammar, readability and plagiarism. Similarly, publishers and researchers can now use software to automate the image checking process.

For example, Proofig software uses computer vision to automatically scan every image in a research paper, producing a report in one to two minutes. The software checks each image against itself and the others in the paper, looking for any anomalies that might be caused by duplications or manipulations.

Researchers can use these tools to scan papers and check sub-images before submitting their work  to their journal of choice, enabling them to detect and resolve any issues before going public. During the review process, grant committees or editors can also use image integrity software to streamline reviews, enabling them to check more papers in less time without negatively affecting the impact factor of the journal.

In academic research, images do more than illustrate a point — in life sciences research, microscopy slides and western blots contain valuable data. Consequently, the academic community needs access to tools that ensure all data, whether in written or image form, remains credible. Software such as Proofig enables researchers and editors to detect and resolve duplications before publication, reducing the risk of printing mistakes or costly investigations while maintaining their reputation.

[1] https://ukrio.org/research-integrity-resources/expert-interviews/jana-christopher-image-integrity-analyst/