Launching the official phase of the PhysioNet Challenge 2026
PhysioNet Challenge
Dear Challengers,
We are pleased to announce the beginning of the official phase of the George B. Moody PhysioNet Challenge 2026: Screening for Cognitive Impairment During Sleep Studies:
https://physionetchallenges.org/2026/.
For the official phase of this year’s Challenge, we have made many changes to improve the Challenge, including:
updated and simplified criteria for patients with and without cognitive impairment, including a more uniform time window across sites closer to the time of the sleep study;
updated training, validation, and test sets, including "small" (~200 GB) and "large" (~1.2 GB) versions of the training set that teams can choose to use;
new scoring metrics that condition on age, recognizing that older patients are more likely to have cognitive impairment and rewarding algorithms for identifying younger patients with cognitive impairment; and
updated example and scoring code that implement these changes.
These changes, made in response to our observations and your feedback during the unofficial phase, help to improve the Challenge, but we have not changed the fundamental task of the Challenge: using sleep studies to identify future cognitive impairment diagnoses.
Please see the data section of the website for updates to the data:
http://physionetchallenges.org/2026/#data-access
Your code from the unofficial phase may not run as-is in the official phase, but you should be able to run it with minor updates. Please see the example and scoring code for updates to the example and scoring code:
https://github.com/physionetchallenges/python-example-2026
We will start evaluating official phase submissions after about a week to allow for questions and comments about these changes (and potential bug fixes), so please share any feedback that you may have as soon as possible.
Please see the Challenge webpage for details, update and double check your code to include all of the promising ideas that you've had since the end of the unofficial phase, and submit it for evaluation on the validation set when ready (but after a week or so to allow for feedback and potential changes in response to feedback):
https://physionetchallenges.org/2026/submissions/
The reviews of your Computing in Cardiology abstracts are nearly finished. We look forward to sharing the results with you in the coming weeks and seeing you and discussing your work in Madrid in September.
Best,
The PhysioNet Challenge Organizers
---
Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email info at physionetchallenge.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.
PhysioNet Challenge
As we announced earlier this month, we have launched the official phase of the 2026 Challenge, with a few updates since the beginning of the month.
First, we have posted both a small version (~200 GB) of the training set and a large version (~1.2 TB) of the training set:
http://physionetchallenges.org/2026/#data-access
The small version of the training set is a subset of and prevalence matched to the large version of the training set. You can download either or both versions of the training set, and you can ask us to train each entry on either version of the training set. All things being equal, we expect to be able to share results on the small version of the training set more quickly because we expect your training code to complete more quickly, but you may find one version to be a better fit for your approach than the other. Both versions of the training set have the same computational constraints.
Second, we have also begun to score your entries on the hidden validation set:
http://physionetchallenges.org/2026/leaderboard/
You'll notice two tabs on the leaderboard: one tab for a prevalence-based reward metric and another tab for an age-conditioned AUROC metric. Both metrics evaluate model performance while incorporating patient age, which is strongly associated with cognitive impairment diagnoses, to better reflect the context of patient evaluations and to support more useful approaches. (The evaluation code also returns other more traditional metrics as well.) We will share more information soon about how to report these metrics in your preliminary papers and presentations.
https://cinc2026.org/
If your abstract was not accepted, or if you discovered the Challenge too late to submit an abstract, then we still encourage you to participate in the official phase and to compete for a "wild card" entry for a high-scoring team without an accepted CinC abstract:
http://physionetchallenges.org/2026/#wild-card
Best,
Matt
(On behalf of the Challenge team.)
Please post questions and comments in the forum. However, if your question reveals information about your entry, then please email info at physionetchallenge.org. We may post parts of our reply publicly if we feel that all Challengers should benefit from it. We will not answer emails about the Challenge to any other address. This email is maintained by a group. Please do not email us individually.