Graphical Abstract

A graphical abstract is a single image — typically one panel, combining a small amount of text with icons, diagrams, or illustrations — that summarizes a paper’s central finding or methodology visually, published alongside (occasionally instead of) the manuscript’s text abstract. Its function is to let a reader viewing a search-results list, a journal’s online table of contents, or a social-media share understand the paper’s core message in seconds, without reading the full text. Because that function depends on the image working on its own, a graphical abstract has to be self-explanatory: no reliance on a caption printed elsewhere, no assumption the viewer has already read the manuscript, and (per most publishers’ own instructions) little to no free-standing prose beyond short in-image labels.

How it differs from a regular figure and from the text abstract

A graphical abstract is not simply “one of the paper’s figures.” An ordinary numbered figure — a bar chart, a micrograph, a flow diagram in the Methods section — documents one specific piece of evidence and is written to be read in sequence with the surrounding Results or Methods prose; on its own it usually isn’t interpretable, and it isn’t designed to represent the whole study. See CASRAI’s Scientific Figure Rules for Journal Submission guide for the general preparation, resolution, and integrity requirements that apply to those in-text figures. A graphical abstract is also distinct from the text abstract: the text abstract is prose — a structured or unstructured paragraph stating purpose, methods, results, and conclusions, submitted as manuscript text and indexed as searchable text by databases such as PubMed and Scopus. A graphical abstract is an image file, submitted separately from the manuscript, that a publisher displays alongside the text abstract (in the online table of contents, search-result listings, and the article landing page) — it supplements the text abstract rather than replacing the indexing and searchability that prose provides.

Which journals and publishers require one

Whether a graphical abstract is required, encouraged, or optional is set at the level of the individual publisher and often the individual journal — there is no single cross-industry rule, so authors need to check the specific target journal’s guide for authors. Real, verifiable examples of current practice:

• Elsevier supports graphical abstracts across ScienceDirect, displaying an eligible article’s graphical abstract in online search-result lists, the online table of contents, and the article page (not, typically, in the print or PDF version) — but whether a specific Elsevier journal requires, encourages, or doesn’t use one is set per journal, per its own guide for authors.
• Cell Press requires a graphical abstract for most of its journals; a specified set of exceptions (including Cell Systems, Current Biology, Patterns, Neuron, and Newton) treat it as optional rather than required.
• The Royal Society of Chemistry (RSC) requires a “table of contents entry” — RSC’s own name for the same concept, an eye-catching graphic plus one to two sentences of summary text — submitted at the revision stage for its journals, appearing in the table of contents and RSS feeds.
• ACS (American Chemical Society) journals require a “Table of Contents (TOC) graphic”: original, unpublished artwork (not a figure reused from the manuscript) that gives a quick visual impression of the paper’s essence without revealing specific results, labeled “For Table of Contents Only” on submission.

Outside chemistry and life-sciences publishing, a related but distinct informal format called a visual abstract — usually a more heavily templated, multi-panel graphic built for conference or social-media promotion rather than the official table-of-contents entry — has become common in medical-journal publishing and post-publication promotion; the two terms are frequently used loosely as synonyms, but a journal’s formal submission requirement (if it has one) is the graphical abstract or TOC graphic as defined in its own author guidelines, not a self-made social-media visual abstract.

Design and technical requirements

Exact specifications are publisher-specific and change over time, so always confirm the current numbers in the target journal’s guide for authors before finalizing a file. As illustrative, currently-published examples of how specific those requirements get:

• Elsevier’s own guidance specifies a minimum image size of 1328 × 531 pixels at a minimum resolution of 300 dpi, using Times, Arial, Courier, or Symbol font at a large enough size to remain legible once reduced for the table-of-contents display, submitted as a separate file (preferred formats: TIFF, EPS, PDF, or an MS Office file) with no additional caption, outline, or synopsis outside the image itself.
• Cell Press specifies a single-panel image roughly 5.5 inches square at 300 dpi, using Arial at 12–16 point type, submitted as TIFF, PDF, or JPG.
• RSC Advances specifies a graphic approximately 8 cm wide by 4 cm tall at a minimum of 300 dpi, with the one-to-two-sentence description entered separately in the submission system rather than embedded in the image.

Across publishers, the recurring design principles are consistent even where exact dimensions differ: use a single panel with a clear reading path (top-to-bottom or left-to-right); keep any in-image text sparse and legible at small display sizes; do not embed a caption, title, or the words “graphical abstract” inside the image; emphasize the paper’s new finding rather than background already covered in the introduction; and secure permission for any third-party image content before use, the same requirement that applies to any other published figure. Several publishers, including Elsevier and Cell Press, have published explicit policies restricting or disclosing the use of general-purpose generative-AI image tools for graphical-abstract creation (Cell Press, for example, permits AI-assisted features inside dedicated scientific-illustration tools but not general-purpose generative-AI image generators) — see CASRAI’s generative-AI disclosure statement entry for the related disclosure obligation, and consult the target journal’s current policy before using any AI-assisted tool.

Related concepts

For general figure preparation, resolution, and image-integrity rules that also apply to figures used within the manuscript body, see CASRAI’s Scientific Figure Rules for Journal Submission guide. For the manuscript’s required prose summary, see the Abstract entry. For CASRAI’s broader scholarly-writing coverage, see the Scholarly Writing & Publishing hub.

Frequently Asked Questions

Is a graphical abstract the same as a TOC (table of contents) graphic?

Not exactly, though the two are close equivalents under different publisher terminology. RSC calls its version a “table of contents entry” — a graphic plus one to two sentences of summary text — and ACS calls its version a “Table of Contents (TOC) graphic,” original artwork labeled “For Table of Contents Only.” Both serve the same purpose as a graphical abstract: a single image representing the paper’s essence.

How is a graphical abstract different from a regular figure in the paper?

An ordinary numbered figure documents one specific piece of evidence and is written to be read in sequence with the surrounding Methods or Results prose — on its own it usually isn’t interpretable. A graphical abstract is a separate, self-explanatory image meant to stand alone without the reader having read the manuscript, summarizing the whole study’s core finding rather than a single piece of evidence.

Do all journals require a graphical abstract?

No. Whether one is required, encouraged, or optional is set by the individual publisher and often the individual journal, not by a single cross-industry rule. Cell Press requires one for most of its journals (with named exceptions such as Cell Systems and Neuron), ACS and RSC both require an equivalent under their own names, and Elsevier supports them but leaves the requirement up to each journal’s own guide for authors.

Can a graphical abstract include text?

Only sparingly. Most publishers’ instructions call for little to no free-standing prose beyond short in-image labels, since the image is meant to work on its own without a caption. Some publishers keep any written summary out of the image file entirely — RSC, for example, has authors enter the one-to-two-sentence description in the submission system rather than embedding it in the graphic.

Is a graphical abstract the same as a visual abstract?

They’re related but not officially the same thing. “Visual abstract” is an informal term for a more heavily templated, multi-panel graphic built for conference or social-media promotion, common in medical-journal publishing. A journal’s actual submission requirement, if it has one, is the graphical abstract or TOC graphic defined in its own author guidelines — not a self-made visual abstract.

Can I use AI tools to create a graphical abstract?

It depends on the publisher and the tool. Several publishers, including Elsevier and Cell Press, have published policies restricting or requiring disclosure of general-purpose generative-AI image tools for graphical-abstract creation — Cell Press, for instance, permits AI-assisted features inside dedicated scientific-illustration tools but not general-purpose generative-AI image generators. Always check the target journal’s current policy first; see CASRAI’s generative-AI disclosure statement entry for the related disclosure obligation.