How to… avoid Misleading images

A quick overview of what are misleading images in publications
images
how-to
Author

Helena Jambor

Published

September 22, 2023

I care about science and images that is widely accessible. At present, that is not always the case. Image figures are confusing and unclear. I have written and published about communicating understandable image data, and last week our community-developed guidelines for publishing understandable images got published.

For scientists it is worth investing time in producing clear papers: incomprehensible data is demonstrably less read, less often cited, and unlikely to be included in the scientific canon. Don’t believe me? Letchford et al have data!

Figure 1: From Letchford et al. 2015, doi: 10.1098/rsos.150266.

Accidentally misleading images

Published images at times are not just unclear but misleading. This may be because image data was incorrectly processed. I previously wrote about one example published by Miura and Norrelykke, which illustrated that the sequence of contrast adjustment and image cropping matter. I regularly witness the education gap in image handling in my DataViz courses. Graduate students ask: Is it ok to turn images? Can I crop away a neighboring cell? Am I allowed to show green fluorescent protein in red? When images are processed without in-depth knowledge of the software, accidentally misleading images may be the result.

Figure 2: Sequence matters when adjusting brightness contrast. Adapted from Miura & Norrelykke

Figure 3: How turning/transforming images can falsify intensity distributions in image data, example from Christopher Schmied.

Actually misleading images

Less common than accidentally misleading images, are instances of intentional image manipulation. These are cases when authors fabricate or forge data to mislead audiences. Elisabeth Bik has brought attention to this problem and also quantified instances of image manipulations in publications.

Figure 4: Example of “composed” image, that contains cloned/spliced image parts, from Elisabeth Bik.

Guidelines to avoid misleading images

Already before Biks quantification and activism, early community initiatives of scientists and editors clarified what counts as image manipulation and which image processing is acceptable for publication. The CSE guidelines are long and update annually! Here is the relevant section on digital image:

Figure 5: CSE 3.4.1 Guidelines for Handling Image Data

Read more

Bik, E.M., Casadevall, A., Fang, F.C., 2016.The Prevalence of Inappropriate Image Duplication in Biomedical Research Publications. mBio 7.

Figure 6: Example in publication.

Accurate and reproducible image quality

Recommendations to avoid misleading images

Detection of manipulated image data

Appropriate image handling and analysis

Guidelines for writing materials and methods sections for images