Back

Does charging for corrections in the bioscience literature disincentivize pre-publication handling of problematic image data? An ImageTwin-AI study.

Brookes, P. S.

2026-01-28 cancer biology
10.64898/2026.01.16.700000 bioRxiv
Show abstract

The Committee on Publication Ethics (COPE) recommends that publishers do not charge for corrections to published papers. Until late 2024 the Journal of Cancer levied a charge on authors (50% of the original article processing charge, APC) for publication of a correction. Herein, it was hypothesized this could disincentivize the discovery and removal of problematic data prior to publication, since post-publication discovery and correction would generate additional revenue. The correction charge policy at J. Cancer was rescinded in 2025, permitting a test of the hypothesis by comparing the prevalence of problematic image data in the journal before and after the policy change. Recently developed artificial intelligence (AI) tools afford the ability to screen scientific publications for problematic image data. As such, the 2024-2025 output of J. Cancer was analyzed using ImageTwin-AI, followed by human verification and annotation of identified problems. Of 754 papers analyzed, 510 contained image data. Of these, 95 (18.6 %) showed evidence of inappropriate image manipulation, with 19 papers (3.7 %) having images that overlapped with unrelated papers. The prevalence of papers with problem images was 20.3% in 2024, and 15.9% in 2025, suggesting only a modest impact of the policy change on pre-publication handling of such problems.

Matching journals

The top 10 journals account for 50% of the predicted probability mass.

1
PeerJ
261 papers in training set
Top 0.1%
19.0%
2
PLOS ONE
4510 papers in training set
Top 30%
5.0%
3
PLOS Biology
408 papers in training set
Top 1%
5.0%
4
Proceedings of the National Academy of Sciences
2130 papers in training set
Top 19%
3.7%
5
Scientific Reports
3102 papers in training set
Top 34%
3.7%
6
eLife
5422 papers in training set
Top 24%
3.7%
7
F1000Research
79 papers in training set
Top 0.5%
3.7%
8
Royal Society Open Science
193 papers in training set
Top 1%
2.4%
9
Patterns
70 papers in training set
Top 0.5%
2.1%
10
npj Systems Biology and Applications
99 papers in training set
Top 0.9%
1.9%
50% of probability mass above
11
Computational and Structural Biotechnology Journal
216 papers in training set
Top 3%
1.9%
12
Cancers
200 papers in training set
Top 3%
1.7%
13
Bioengineering
24 papers in training set
Top 0.5%
1.5%
14
BMC Biology
248 papers in training set
Top 1%
1.5%
15
Heliyon
146 papers in training set
Top 2%
1.5%
16
Scientific Data
174 papers in training set
Top 1%
1.4%
17
Communications Biology
886 papers in training set
Top 12%
1.4%
18
Open Biology
95 papers in training set
Top 0.9%
1.4%
19
PLOS Computational Biology
1633 papers in training set
Top 18%
1.4%
20
Frontiers in Oncology
95 papers in training set
Top 3%
1.0%
21
International Journal of Cancer
42 papers in training set
Top 1%
0.9%
22
GigaScience
172 papers in training set
Top 2%
0.9%
23
Frontiers in Bioinformatics
45 papers in training set
Top 0.7%
0.8%
24
Science
429 papers in training set
Top 19%
0.8%
25
BMC Medicine
163 papers in training set
Top 6%
0.8%
26
Database
51 papers in training set
Top 0.8%
0.8%
27
Frontiers in Genetics
197 papers in training set
Top 9%
0.8%
28
Journal of Translational Medicine
46 papers in training set
Top 2%
0.8%
29
iScience
1063 papers in training set
Top 31%
0.8%
30
Molecular Systems Biology
142 papers in training set
Top 2%
0.8%