About SleepScorer
SleepScorer is a research tool for sleep analysis. It was built to make automated sleep scoring easier to understand, review, and question, rather than presenting results as a black box. The system combines rule-based scoring, machine learning, uncertainty warnings, and human review in one interface.
SleepScorer lets users inspect sleep metrics, review the hypnogram, examine individual epochs, and see where different methods disagree or where the system is uncertain. It presents rule-based and pattern-based estimates side by side and allows manual correction when needed.
The system also includes a chat interface for discussing decisions in plain language. The goal is to support human oversight, not replace it.
SleepScorer was created by Emil Hardarson as part of his doctoral research on AI in regulated sleep analysis at Reykjavik University. It reflects a broader research effort on how AI systems can be made more transparent, auditable, and useful in real clinical workflows.
SleepScorer is free for personal use.
For questions, feedback, or collaboration inquiries, please contact hello@sleepscorer.com.