Changing Standards for Manipulation of Digital Images in Biomedical Articles

Addeane S. Caelleigh* and Kirsten D. Miles+

*UVa School of Medicine; +P. I. Outcomes; formerly Research Computing Lab, Brown Science and Engineering Library, UVa

Introduction

ResultsWe identified 162 PIs who published in 475 different journals during 1995-2010. Table 1 gives the basic breakdown of these journals’ guidelines for digital images.

Journals’ standards. For each journal, we examined the instructions to authors, ethical standards, policies and procedures, and other materials to identify its guidelines for digital images.

Categorization system for standards. Based on the first 25 journals identified, we created a categorization system to describe the range of guidelines. These categories were then applied to the entire sample.

Discussion This research revealed a trend among biomedical journals to adopt standards for the digital images that authors submit for publication.

Increasing trend. Over the past 10 years, biomedical journals have introduced new standards for appropriate manipulation of the digital images they publish.

In 2003, when the Journal of Cell Biology became the first major journal to set guidelines for digital images and to screen submitted images, very few journals were prepared to follow its lead. In less than 10 years, however, over half in our sample had adopted at least minimal guidelines, with 11% of the sample now having detailed standards or guidelines for digital images. Importantly, many of those with detailed guidelines are among the most prestigious journals.

It is important that authors know where their journals fit in the spectrum of guidelines for manipulation of published images. For example, Science, Blood, and all 34 Nature Group journals now have detailed, explicit guidelines, and all journals published by Elsevier have basic guidelines as well as strictures against inappropriate image manipulation.

© Kristen D. Miles

Data-management requirements. The National Institutes of Health and the National Science Foundation, recognizing the need for data to be shared and repurposed, now require plans for data-management and data-sharing. These plans apply to digital-image data just as to all other data.

Standards for appropriate manipulation of digital data have developed more slowly than has the software for manipulating the images. Half of all cases now investigated by the federal Office of Research Integrity involve questions about digital images. Editors of biomedical journals have begun setting standards for published images, but many researchers are not aware of these changing standards.

It is crucial that biomedical researchers and trainees understand journal requirements, assure the integrity of their digital images, and make the best decisions about the collection and management of data images.

Table 1 Categorization of 475 biomedical journals’ standards for digital images

Category No. %

1. no instructions or guidelines for images or illustrations 9 1

2. instructions or guidelines refer to “art” (illustrations) * 224 47

3. instructions or guidelines refer to digital manipulation 192 40

4. detailed instructions and guidelines for digital images 50 11

Totals 475 100

*Traditionally, journals and publishing houses have used the term “art” to refer to anything that was not text. Therefore, instructions for “art” can cover all types of illustrations, although very few mention digital images.

Half of the journals in the sample had at least minimal guidelines for digital manipulation of images (242 journals, 51%) . Of these journals, 192 (40%) informed authors of general expectations for digital images, and another 50 (11%) gave detailed information about these expectations.

Excerpts from journals in categories two through four illustrate the range of guidelines found in the sample.

• Guidelines for “art”/illustrations only. Journals whose instructions dealt only with “art” (illustrations) had statements about graphical formats such as “Art should be created/scanned and saved and submitted as either a TIFF …” (Optometry and Visual Science) or “Symbols, letters, and numbers must be legible after reduction …” (Magnetic Resonance in Medicine).

• General guidelines for digital manipulation. Journals that gave general guidelines for digital images had statements such as, “For Graphical images, journals published by Elsevier apply the following policy: no specific feature within an image may be enhanced, obscured, moved, removed or introduced” (Gene) and occasionally described policies for screening submitted images.

• Detailed instructions or guidelines. Journals that gave detailed guidelines had requirements such as “Authors should retain their unprocessed data and metadata files …” and “The use of touch-up tools … or any feature that deliberately obscures manipulations, is to be avoided” (Nature). They often described their screening policies, such as “All images in Figures and Supplemental information from manuscripts accepted for publication are examined for any indication of improper manipulation or editing” (Blood) .

Resources on manipulation of digital imagesA list of resources and other information on image-integrity standards is available on a handout.

RecommendationsAll biomedical researchers need to be prepared to meet journals’

expectationsand submit images that meet their standards, and they need support

to do so.1). The University should develop training for researchers at all

levels about meeting the guidelines and standards for published image data. 2). Researchers need to have procedures that meet the evolving

standards for capture of image data, retention of unaltered originals, and

maintenance of audit data for manipulations of all data images.

Spectrum of journal image guidelines

Research DesignThis research was designed to a. develop a categorization system to identify

the range of standards among biomedical journals

b. categorize the standards of a large sample of biomedical journals

c. determine the proportion of journals along this spectrum of standards

d. determine the trend in standards of journals over the past 10 years.

Study sample. The study sample was the journals in which a defined group of biomedical faculty had published 1995-2010. We created a list of the principal investigators (PIs) on all research grants awarded between January 2005 and June 2010 to faculty of 10 biomedical departments and centers of the University of Virginia School of Medicine. (These departments and centers were Biochemistry, Biomedical Engineering, Biomedical Ethics, Cell Biology, Cell Signaling, Microbiology, Molecular Design, Molecular Physiology and Biological Physics, and Neuroscience.) Using PubMed searches, we compiled a list of the journals in which these PIs had published during 1995-2010. The resulting list of journals was the sample for the study.

Results We identified 162 PIs who published in 475 different journals during 1995-2010. Table 1 gives the basic breakdown of these journals’ guidelines for digital images.