Thanks to an amazing (please read below) paper in Scientific Reports we
Thanks to an amazing (please read below) paper in Scientific Reports we now know it is likely that Rambo was having a stroke, which may explain his erratic behavior. In related news, said Scientific Reports paper may (one only hopes) be retracted due to a "a comically bad dataset" (thank you Adrian Barnett for your hard work and Retraction Watch for spreading the word). When images of patients in their natural settings include pictures titled forest-whitaker-instagram.png or ss.png with Sly on it, suspicions may arise. In addition to impeding the progress of science, potentially endangering patients, enabling the propagation and amplification of seriously bad datasets and including pictures of children without explicit consent, this type of article raises depressing questions about Kaggle and Springer Nature's review process integrity for at least some of its journals. Scientific Reports is not building a great track record in that respect. 🤯 I will leave you with the thing that gives me the most anxiety: the "self-referential trap." On the The PubPeer Foundation website, a comment by the author of an earlier (and by now retracted) Scientific Reports article using another dubious dataset should give you pause: "The dataset used in this work is the widely known Stroke Prediction dataset available on Kaggle. This dataset remains publicly accessible and continues to be used by researchers in the machine learning and healthcare analytics community. Numerous studies before and after the publication of our work rely on the same dataset for benchmarking predictive models related to stroke risk. Its continued presence in current literature indicates that it remains a commonly accepted resource for experimental evaluation in this research area. The dataset remains publicly accessible on Kaggle at the time of this response, and identical copies of the dataset are also available on several other open access online repositories."
https://pubpeer.com/publications/2CDA2B937DC37BDE2AB42952AD59B8
https://retractionwatch.com/2026/05/18/kaggle-dataset-clinical-models-stroke-diabetes/