Graphical Abstracts and Video Abstracts: What Medical Journal Authors Need to Create in 2026

The submission checklist for most biomedical journals has grown longer in the past few years, and somewhere near the bottom of many of them now sits a line that catches authors off guard: graphical abstract required. Elsevier has listed graphical abstracts as a submission component on hundreds of its journals for years. MDPI has been actively encouraging them across its catalog of more than 400 clinical and life science titles. Oxford University Press published dedicated creation guidance for journal authors in July 2026. If you have been ignoring this field because the word "optional" appeared on your target journal three years ago, it is worth checking again.

This matters for practical reasons that go beyond editorial compliance. A graphical abstract is what shows up in search result previews on ScienceDirect, in journal table-of-contents emails, and on social media shares posted by the journal itself. If it is absent, the slot is filled by the first text line of your abstract, which is rarely compelling. If it exists but was assembled in ten minutes using a cropped version of Figure 2, it typically signals something about how much care went into the surrounding manuscript. Editors notice this, and reviewers sometimes do too.

This guide covers what graphical abstracts and video abstracts are, what the major publishers actually require in 2026, which tools are available for researchers without design training, what Elsevier's prohibition on AI-generated images means in practice, and how to build a submission-ready visual summary without spending a week on it.

Why Graphical Abstracts Have Moved From Optional to Expected

Elsevier introduced graphical abstracts in 2012 as a feature on ScienceDirect, framing them as a visual table of contents that would help readers quickly evaluate papers before reading them. At the time, most journals listed them as optional. Adoption was patchy. Many researchers submitted without them and heard nothing from editorial staff.

That pattern has shifted, for two reasons. The first is discoverability. As research databases compete with preprint servers and social media for the attention of working clinicians, publishers have invested in making their content easier to browse. A visual element that can be shown in an email preview or a Twitter card is genuinely more shareable than a paragraph of statistical methodology. The business case for pushing authors to create these has only strengthened.

The second reason is evidence. The BMJ conducted an experiment in 2016 in which some papers were promoted with visual abstracts on social media and others were not. Papers with visual summaries reached a measurably wider audience. That study was small, but similar findings have since been replicated in ophthalmology, emergency medicine, and surgical specialties. Journals that saw citation gains after introducing visual abstracts have had little reason to walk that back.

MDPI's position is perhaps the clearest of any major publisher. In announcements to authors of several high-volume journals including the Journal of Clinical Medicine, MDPI stated explicitly that graphical abstracts "help visualize your study" and encouraged authors to submit them with every original research article. A July 2026 guidance page from Oxford University Press addressed both graphical and video abstracts together, giving specific format and design recommendations for the first time. That guidance is a useful benchmark even if you are not submitting to an OUP title.

What Major Publishers Actually Require

Publisher-level guidance and journal-level requirements are not always the same thing. Elsevier sets minimum dimensions of 1,328 x 531 pixels at 300 dpi for its graphical abstracts, with accepted formats including TIFF, EPS, PDF, and JPEG. The image as displayed on ScienceDirect is scaled to a 500 x 200 pixel window in most article listings, which means anything with fine text or dense detail becomes illegible at the display size. Elsevier's recommendation to design for the small display size first is genuinely useful advice, not just a disclaimer.

Springer Nature does not impose a universal graphical abstract policy across its portfolio, so requirements differ by journal. Scientific Reports encourages authors to include a graphical abstract but does not require one. Nature Medicine and Nature Communications treat them as optional. If a Springer Nature journal has specific requirements, they appear in the submission instructions for that title rather than in the group-level author guide. Checking the specific journal page, not the publisher homepage, is the only reliable approach.

Wiley uses the term "graphical abstract" or "visual abstract" depending on the journal, and requirements are similarly distributed at the journal level. The American Diabetes Association's journals, published under a society model rather than a large commercial house, have partnered directly with BioRender and list specific image export settings on their submissions page.

The Lancet uses a specific format it calls a "visual abstract" rather than a graphical abstract, and it differs structurally from the Elsevier standard. A Lancet visual abstract is a formatted one-panel image that presents the study's population, design, and key finding in a structured layout with specific typography. If you are submitting to a Lancet family journal, consult the most recent author instructions specifically for the visual abstract section. It is not interchangeable with the general-purpose image you would prepare for an Elsevier title.

The Anatomy of an Effective Graphical Abstract

The purpose of a graphical abstract is not to reproduce your paper in icon form. It is to answer, in one image, the question a reader asks in the first five seconds: what did these authors study, how, and what did they find? That question has four parts, and a graphical abstract that answers all four clearly will outperform one that is more visually sophisticated but harder to read quickly.

The population element usually appears at the left side of the image: who the study involved, expressed as a simple icon or schematic with a brief label (adults aged 18-65 with type 2 diabetes; patients admitted to the ICU within 24 hours of surgery; children under 5 in rural sub-Saharan Africa). The intervention or exposure comes next. For a clinical trial, this is the treatment versus the comparator. For an observational study, it is the exposure. For a diagnostic study, it is the index test and the reference standard. The outcome sits at the right side of the image, typically with a single key number that tells readers whether the study found an effect.

Text should be minimal. A graphical abstract that contains a paragraph of fine-print methodology is not a graphical abstract; it is a methods section that has been formatted into a rectangle. Most effective graphical abstracts contain fewer than 50 words of text, often considerably fewer. Numbers, arrows, and simple icons carry most of the meaning.

Color discipline matters more in a small image than in a full figure. Pick two or three colors, apply them consistently (one color for the intervention group, one for the comparison group, one for outcomes), and avoid gradients or highly saturated combinations. Red and green together are effectively invisible to about 8 percent of male readers, who are color blind in that range. Tools like Coblis or the Colblindor simulator let you check your image before submitting.

The question of whether to include the paper title is genuinely contested. Some journals want the title on the image; others find it makes the design cluttered and unnecessary because the title already appears adjacent to the image in the journal layout. If the instructions say nothing specific, leave the title off the image and include it in the file name instead.

Tools: What Works for Biomedical Authors Without Design Training

BioRender is the tool most biomedical researchers now use for graphical abstracts, and for good reason. It was built specifically for life science and clinical content, which means its icon library contains biologically accurate representations of cells, organs, laboratory equipment, and clinical scenarios that general-purpose design tools do not have. The American Diabetes Association's journals link to BioRender directly from their submission instructions. BioRender images exported with a license are explicitly permitted by major publishers, including those that restrict AI-generated content.

The limitation of BioRender is cost. Individual subscriptions run roughly $35 to $99 per month depending on plan. Many academic libraries and medical schools maintain institutional licenses, and it is worth checking whether your institution offers access before paying out of pocket. BioRender's template library for graphical abstracts is a reasonable starting point even for researchers with no prior design experience.

Canva is the more accessible alternative for clinical and public health papers that do not require biological diagrams. It is free at the basic level, has a wide range of medical and healthcare icons, and allows export at sufficient resolution for most journal requirements. Its strength is flexibility; its weakness is that it does not carry BioRender's institutional credibility for anatomy-specific content. For a clinical outcomes paper where the key graphic is a patient cohort diagram and an odds ratio bar, Canva is perfectly adequate.

Microsoft PowerPoint produces graphical abstracts that are acceptable for most journals if you export at high DPI. The technique is to set up a custom slide dimension matching the journal's required pixel dimensions, build the image in that space, then export as PNG at 300 dpi. The result lacks the polish of BioRender but is readable and submission-compliant. For researchers at institutions without BioRender access who are under a submission deadline, this remains a workable option.

Adobe Illustrator produces the highest quality output and the most flexible vector graphics, but it requires training that most clinical researchers have not had. If your institution employs a medical illustrator or graphic designer, Illustrator is worth requesting for high-profile journals. For most researchers, the time investment in learning Illustrator from scratch is not proportionate to the result.

Elsevier's Ban on AI-Generated Graphical Abstract Images

Elsevier's current policy on generative AI states explicitly that general-purpose AI image tools may not be used to create graphical abstracts. The named category includes tools like Midjourney, DALL-E, Stable Diffusion, and Adobe Firefly when used to generate novel images. The prohibition is separate from Elsevier's stance on AI writing assistance, which it permits with disclosure.

The reasoning is not arbitrary. General-purpose image generators are trained on photographic and artistic content, not on validated anatomical or clinical illustration. An AI-generated image of a heart valve or a histological slide can look plausible while depicting structures that do not exist or that have features inconsistent with the described pathology. For a scientific figure, that kind of plausible inaccuracy is worse than a simple error because readers may not think to question an image that looks convincingly real.

The practical implication is that BioRender is not prohibited by this policy because it does not generate novel anatomical content. It assembles images from a curated, validated library of scientific illustrations. The policy targets tools that generate pixel-level images from text prompts, not tools that use licensed scientific assets. If you are unsure whether a specific tool falls inside or outside this restriction, the safest test is to ask whether the tool generates or assembles. Generation from a prompt is prohibited; assembly from validated library components is not.

Video Abstracts: Format, Length, and What Journals Expect

Video abstracts are shorter research summaries in video format, typically running three to five minutes. MDPI has been the most active in encouraging them across its journal portfolio, and several society journals in surgical specialties and ophthalmology now include a video abstract alongside the paper as a standard supplementary component.

Most journals that accept video abstracts specify MP4 format, a 16:9 aspect ratio, and a maximum duration of three to five minutes. Resolution requirements vary from 720p to 1080p depending on the platform. Subtitle or caption files (typically in SRT format) are increasingly requested alongside the video file, and in many cases journals now treat them as accessibility requirements rather than optional extras.

The common formats for video abstracts in clinical research are narrated slide presentations (the most accessible to create), screencast walkthroughs of a figure or dataset, and talking-head segments recorded by one of the authors. Animated explainer videos are occasionally submitted for high-profile findings but are resource-intensive to produce. Most journals do not have a format preference and will accept any of these, provided the technical specifications are met.

Patient consent is a specific concern with video abstracts in clinical research. If your video includes any identifiable patient information, case images, pathology specimens that could be traced to a specific individual, or footage recorded in a clinical setting, you need documented patient consent before the video can be published. This is no different from the consent requirements for case reports, but it is easy to overlook during video production. The consent should cover the specific platform on which the journal will host the video, which is sometimes YouTube rather than the journal's own website.

Background music in video abstracts is a recurring editorial problem. Many authors add music to make their video feel more professional, then submit a file that contains a track the journal cannot legally distribute because the authors do not hold the rights. Use royalty-free music from clearly licensed libraries, or use no music at all. The journal will remove or reject a video with unlicensed audio, and this is rarely discovered until the post-acceptance stage when it delays publication.

How Graphical Abstracts Are Used After Publication

Understanding where a graphical abstract appears after publication is useful for deciding how much detail to include and how to handle labeling. On ScienceDirect, Elsevier's platform, graphical abstracts appear in search results and in journal browsing views, scaled to the small display size noted earlier. They also appear in table-of-contents alert emails sent to subscribers, where they represent the paper entirely to anyone who has not yet clicked through.

Journals routinely share graphical abstracts on their social media accounts when promoting newly published papers. This is one of the highest-value channels for author visibility, because journal social media accounts can have follower counts that individually exceed the subscriber base of the journal itself. A graphical abstract that reads clearly at 500 x 250 pixels on a smartphone screen is one that actually serves this function. One with fine text or complex multi-panel structure does not.

Authors are also free to share their own graphical abstracts, and many do. The standard route is posting to platforms like X (formerly Twitter) or LinkedIn alongside a link to the published paper. Academic networking sites like ResearchGate and Academia.edu pick up paper metadata automatically, but they do not always display graphical abstracts unless the author uploads one separately. If you are active on academic social platforms, uploading your graphical abstract there is a low-effort way to improve how your paper appears in search results.

A Pre-Submission Checklist for Graphical Abstracts

The most common reason a graphical abstract triggers an editorial revision request is a mismatch between the image specifications and the journal's requirements, discovered at submission. Journals differ on pixel dimensions, aspect ratios, file formats, resolution, and whether the image should contain the paper title. None of these are difficult to fix, but they all require redoing the export if you did not check before building the image.

The second most common problem is visual complexity that makes the image illegible at display size. Authors sometimes confuse the print resolution of an image (which determines how sharp it looks when printed at full size) with the information density appropriate for the image when it is shown at a fraction of that size online. An image at 300 dpi that contains eight panels, three axes, a sample size breakdown, and a legend is a figure, not a graphical abstract.

One practical shortcut that works well: build your graphical abstract in the final week of manuscript preparation rather than after you receive reviewer comments. By the time you are revising the paper in response to peer review, the study's visual story may need to change anyway, and you will have wasted the first version. Waiting until after the first round of review also means you are adding a task during what is usually the most pressured stage of the submission cycle.

What This Means for Your Next Submission

The honest assessment is that journals are not yet uniformly enforcing graphical abstract requirements at the point of submission. You can submit to some Elsevier journals without one and receive no immediate rejection. But the direction of travel is clear. Publishers have invested enough in the infrastructure to display and share these images that they are unlikely to let them become optional again, and journal-level requirements are tightening faster than the author community has noticed.

The researchers who will be most inconvenienced by this trend are those who try to retrofit a graphical abstract to a paper after it has already been accepted. Creating one at that stage, with no flexibility to revise the study design for visual clarity, is harder than building it in parallel with the manuscript. The ones who find the process easiest are those who treat it as a design exercise they run alongside writing, using the constraints of a 500-pixel-wide image as a forcing function that makes the main argument clearer.

If your institution has a BioRender license, use it. If it does not, Canva or PowerPoint will serve you for most clinical journals. Check the specific journal's author instructions for dimensions before you start, build at 300 dpi, simplify aggressively, and test at display scale. That process takes two to four hours the first time and considerably less once you have a template. The resulting image will outlast the paper in certain contexts, appearing in search results and social shares long after most readers have stopped citing the text.

Further Reading