The University of Zurich (UZH) has cleared neuropathologist Adriano Aguzzi, longtime director of the Institute of Neuropathology at University Hospital Zurich, of research misconduct — but its investigation found that 7 of the 36 studies he co-authored between 1996 and 2023 contain “scientifically significant” errors, primarily inappropriate alteration or reuse of figures, microscopy images, and other visual representations of data. The university’s report, released July 2, 2026, states it is now incumbent on Aguzzi to retract or correct those seven papers, and that failing to do so “would constitute a separate violation of good scientific practice.”
For research-administration offices, the case is a useful, real-world illustration of three things at once: how a formal misconduct inquiry under an institutional integrity ordinance actually reasons through intent and negligence, what corresponding-author responsibility means in practice when a lab’s junior staff are implicated, and how exposed image-based methods (microscopy, histology, Western blots) remain to both genuine error and deliberate manipulation across a career-length body of work.
What the investigation found
UZH’s integrity proceedings, formally opened in March 2024 and led by legal scholar Wolfgang Ernst (University of Zurich and University of Oxford), reviewed 36 publications spanning a 28-year period. Aguzzi is listed as corresponding author on 24 of them. The university’s own account, published as “Integrity Proceedings Clear Adriano Aguzzi”, concluded that it could not establish Aguzzi acted intentionally or negligently under UZH’s Integrity Ordinance — the standard the university applies to distinguish sanctionable misconduct from honest error or the actions of others in a lab that a supervisor could not reasonably have caught.
At the same time, the report identified seven papers with errors serious enough to require correction, centered on figure and image handling. Coverage of the report from The Transmitter and the research-integrity blog For Better Science both describe the flagged issues as involving figures that were misassigned, inadequately processed, or otherwise modified — the same category of problem covered generally under CASRAI’s image manipulation entry. The university’s finding places the burden of correction on Aguzzi directly: not retracting or correcting the seven papers, the report states, would itself be treated as a further breach of good scientific practice, independent of the underlying misconduct finding.
How the case reached a formal inquiry
The proceedings trace to a dossier of concerns delivered to Aguzzi and UZH in February 2024, compiled with contributions from research-integrity analyst Elisabeth Bik and science journalist Charles Piller. The dossier alleged that a former laboratory employee had fabricated or falsified mouse-brain experiments by reusing microscope images from earlier studies. UZH opened a formal investigation under its Integrity Ordinance the following month.
Two papers co-authored by Aguzzi have already been retracted in the past two years, independent of this inquiry. One traces to a separate University of Minnesota investigation involving co-author Sylvain Lesné, who resigned his tenured position in February 2025 after his own image-integrity findings drew sustained scrutiny. Aguzzi reached retirement in January 2026, while the UZH proceedings were still underway.
Bik has publicly said she remains unsatisfied with the outcome, arguing that as corresponding author on many of the papers under review, Aguzzi bears responsibility for their integrity regardless of who performed the underlying image manipulation.
Why corresponding-author status is the crux of the disagreement
The gap between UZH’s finding (no proven intent or negligence) and Bik’s public reaction (responsibility regardless of who manipulated the images) is really a disagreement about what corresponding-author responsibility should mean in practice, not a factual dispute about what happened to the images. ICMJE’s authorship criteria and most journals’ policies hold the corresponding author accountable for the integrity of the work submitted under their name, but institutional misconduct findings — UZH’s included — generally require evidence of the individual’s own intent or negligence, not just formal authorship position, before a misconduct finding attaches personally. Those are two different bars, and a case can clear one without satisfying the other. That distinction is exactly why the university’s report can simultaneously clear Aguzzi of misconduct and still require him to retract or correct the flagged papers: correction is a scientific-record obligation tied to authorship, separate from a disciplinary finding tied to conduct.
For institutional research-integrity offices, the practical takeaway is to keep these two tracks explicit in policy and in communication with faculty and the public: a “not guilty of misconduct” finding is not the same statement as “the record needs no correction,” and conflating the two — in either direction — misrepresents what an integrity inquiry actually determined. See CASRAI’s guide on how a research misconduct investigation actually works for the general process most institutions follow, and the research misconduct and research integrity dictionary entries for the underlying definitions.
Image integrity across a 28-year body of work
The scale of this review — 36 studies across 28 years — is also a reminder of how image-based methods concentrate integrity risk over a long career. Microscopy, histology, and Western blot figures are reused, cropped, relabeled, and reformatted repeatedly as a lab publishes follow-on work, and each of those steps is a point where an image can be duplicated, mislabeled, or altered without leaving an obvious trace to a reader — which is exactly why image-forensics tools and dedicated screeners (the kind of work Bik and others in the research-integrity community do systematically) have become a standard part of post-publication scrutiny, alongside the pre-publication figure checks some journals now run. CASRAI’s guide on research integrity red flags and its image manipulation entry both cover the broader pattern this case sits within.
The practical implication for research offices and PIs running image-heavy labs is that figure provenance — knowing exactly which raw image underlies which published panel, across years and across projects — is not a bureaucratic nicety but the actual defense against exactly this kind of long-horizon review. Labs that retain and can trace original, unprocessed image files are far better positioned to resolve a flagged figure quickly than labs that cannot reconstruct which image was used where.
What this means for research-administration offices
- Distinguish “no misconduct finding” from “no correction needed” in any communication about a cleared case — the two are legally and scientifically separate determinations, as this case demonstrates directly.
- Revisit expectations for corresponding authors specifically, not just PIs generally — journal and funder policy already treats the corresponding-author role as carrying integrity accountability for the submitted work; institutional policy should be at least as clear.
- Build (or audit) a raw-image retention requirement into lab data-management practice — the CASRAI data management plan entry covers the general DMP framework this fits into; image provenance is a natural, underused component of it for any image-heavy lab.
- Expect post-publication review to keep reaching further back — this inquiry covered work published as early as 1996. Institutional records-retention and correction workflows need to be able to act on decades-old publications, not just recent ones.
Sources
- University of Zurich, “Integrity Proceedings Clear Adriano Aguzzi” (July 2, 2026) — primary institutional report.
- The Transmitter, “Neuropathologist not guilty of research misconduct, says university probe” (July 2026).
- For Better Science, “Schneider Shorts 10.07.2026” (July 10, 2026).
