Hundreds of animal studies flagged for problematic images

Researchers from Radboud University Medical Center have identified over 240 scientific publications on animal models of haemorrhagic stroke that contain potentially problematic images, raising concerns about the trustworthiness of the body of literature this field. The team’s findings have been published in the journal PLOS Biology.

Researchers often use images in their publications to provide evidence of whether a treatment works; for example, by showing the presence or absence of specific proteins or cells or changes in brain tissue. But the researchers found many instances of image duplication within publications, as well as across different publications, even when authors claimed the images came from separate experiments with different experimental conditions. This kind of duplication can cast doubt on the validity of the study’s conclusions and its scientific integrity.

Originally, the team set out to systematically review animal studies on early brain injury following haemorrhagic stroke to find promising treatments suitable for clinical testing. But after noticing suspicious images in some papers in mid-2023, they set out to perform a systematic investigation of image-related problems in all 608 potentially relevant publications. The researchers found that 243 of these papers were problematic, most often for containing problematic images.

“We found a high prevalence (40%) of papers with image-related issues among preclinical literature on early brain injury after haemorrhagic stroke,” the authors said.

“These findings shocked us to our core and might explain why, in spite of hundreds of animal studies published in this field, we still do not have any effective treatments for early brain injury in haemorrhagic stroke patients.

“During the summer break of 2023 we were texting each other figures from papers to collectively find image overlaps. Every time a new message came in, we knew somebody had found yet another inappropriate image overlap. It was at this moment that we knew that we had a serious problem on our hands that we needed to investigate in a systematic way.”

With their findings raising concerns about the overall trustworthiness of animal-based research in this field, the authors have stressed the need for journals and publishers to investigate these issues carefully and take corrective editorial action where necessary. They said the scientific record needs to be as error-free as possible in order for scientists, clinicians and policymakers to make well-informed, evidence-based decisions that benefit patients’ health.

Image credit: iStock.com/fotosipsak