Retraction Note: Multisensory learning binds neurons into a cross-modal memory engram
Article Date: 25 March 2026
Article URL: https://www.nature.com/articles/s41586-026-10355-4
Article Image: Header image
Summary
The authors have retracted their 2023 Nature paper after failing to reproduce key voltage imaging results. Replication experiments and an independent reanalysis revealed that the voltage imaging data reported in Figs. 3a–3f and 4a–4b, and the section titled “Neurons gain cross-modal activation”, are not reproducible. The team identified errors in the original data-analysis pipeline that may have contaminated the ΔF/F0 signals. Behavioural and connectomic results were replicated, and the authors state the general conclusions remain supported, but they have no confidence in the voltage imaging evidence and therefore retract the article. The authors apologise and say follow-up data will be shared. All authors agree with the retraction except Pedro F. Jacob, who could not be contacted by the corresponding author.
Key Points
- The Nature paper (original DOI: 10.1038/s41586-023-06013-8; published 26 April 2023) has been formally retracted (25 March 2026).
- Voltage imaging results underlying claims about cross-modal neuronal activation (Figs. 3a–3f, 4a–4b) are not reproducible in follow-up experiments.
- Errors in the original analysis pipeline could have contaminated ΔF/F0 data; an independent reanalysis failed to replicate the original voltage-imaging findings.
- Behavioural and connectomic data were replicated by the authors, so some broader conclusions may still hold, but confidence in the voltage-imaging support is lost.
- The authors thank colleagues who helped identify analysis issues; follow-up datasets will be made available. One author (Pedro F. Jacob) could not be contacted regarding the retraction.
Context and Relevance
This retraction matters for researchers using voltage imaging and those studying multisensory memory engrams. The original paper made mechanistic claims about cross-modal activation in identified neurons; those claims relied heavily on the now-questioned imaging data. The case highlights the importance of robust analysis pipelines, independent replication, and open data sharing in neuroscience. Labs that cited or built on the imaging results should re-evaluate those dependent conclusions while noting that behavioural and connectomic findings reported by the authors were replicated.
Author note
Punchy summary: a high-profile claim has been pulled because key experimental signals didn’t hold up and the analysis pipeline had flaws — this is a clear reminder that even flashy published results need rigorous replication and transparent processing.
Why should I read this?
Look, if you work in neuroscience, imaging, or reproducibility, this is one to skim properly. It tells you which piece of evidence collapsed (voltage imaging), what still stands (behavioural/connectomic), and why pipeline checks and open follow-up data matter. We’ve saved you the time of digging through the full notice — but if this work touched your projects, go and read the full retraction and grab the follow-up data when it’s posted.