Purpose and Status
This is a Journal Implementation Standard: LMI defines the integrity rules; each journal implements screening and verification through its own procedures. The governing principle is that an image presented as research evidence shall accurately represent what was observed. Processing that alters the information an image conveys about the underlying result is prohibited, however it is achieved.
Scope and Boundaries with C2 and C4
This standard governs the manipulation of genuine research images and figures — images produced from the authors' own observation or measurement. Two adjacent matters are governed elsewhere and are not redefined here:
- Reuse and attribution of images or figures originating from another source, including republication of one's own previously published figures, is governed by C2 and, for permissions and licensing, by F1.
- AI-generated or materially AI-modified images, and disclosure of AI assistance, are governed by C4. Generated imagery shall not be presented as research observation.
Where a single figure raises more than one issue — for example, a reused figure that has also been altered — each applicable policy applies on its own terms.
Permitted Processing
Adjustments to brightness, contrast, colour balance or sharpness are acceptable where they are applied uniformly to the entire image, do not obscure, eliminate or misrepresent any feature, and do not alter the interpretation of the result. Cropping is acceptable where it does not remove information material to interpretation or conceal the removal of relevant context.
Prohibited Processing
- selective enhancement, obscuring, moving, removing or introduction of any feature, band, cell, particle or region within an image;
- adjustment applied to part of an image only, or to one panel of a comparison but not another, in a way that affects interpretation;
- splicing of lanes, fields or panels — from the same or different sources — presented as a contiguous image;
- duplication of an image, panel or region to represent a different sample, condition, time point or experiment;
- presenting an image of one specimen, condition or experiment as another;
- adjustments that eliminate weak signal, background or unexpected results so as to strengthen an apparent effect.
Composite Images and Disclosure
Composite figures, spliced gels or blots, stitched fields and merged channels may be legitimate where the assembly is scientifically appropriate. Where used, the composition shall be disclosed in the figure or its legend, with dividing lines or clear demarcation showing where images have been joined, and the methods shall describe the processing applied. Undisclosed composition is prohibited.
Retention of Source Files
Authors shall retain unprocessed, original image and instrument files and the processing record for a period no shorter than the retention expectation the journal states, and shall provide them promptly to the journal on request, at any stage before or after publication. Inability or refusal to produce original files where a genuine concern has arisen is a material consideration in assessing that concern.
Screening
Each journal shall conduct image-integrity screening proportionate to the risk presented by its article types, at a stage of its own choosing. LMI does not prescribe software, thresholds, sampling rates or forensic workflow; those are journal implementation matters, and journals may use publisher-supported tools where LMI provides them.
Screening output is an indicator, not a determination. Apparent duplication or irregularity frequently has innocent explanations — legitimate composition, similar specimens, compression artefacts, formatting errors. The editor shall put the observation to the authors, request original files, and assess the explanation. Where the matter appears to involve deliberate manipulation, it proceeds under C5 with the due-process protections that policy establishes; record action follows G2. Correctable errors of assembly or labelling are ordinarily remedied by correction rather than treated as misconduct.
Standards and Guidance
Informed by COPE guidance on image manipulation and recognised image-integrity guidance from research-integrity bodies including the US Office of Research Integrity.