SleepVLM: Explainable and Rule-Grounded Sleep Staging via a Vision-Language Model

1. 1Pre-training the model on waveform-perceptual data
2. 2Utilizing a vision-language model to stage sleep from PSG waveform images
3. 3Generating clinician-readable rationales based on AASM scoring criteria
4. 4Evaluating the model's performance and accuracy in clinical settings

Data scientists and AI engineers on a healthcare team can benefit from SleepVLM's explainable sleep staging, enabling more accurate and trustworthy clinical decisions

💡 SleepVLM's rule-grounded approach enables auditable reasoning and clinician-readable rationales, increasing trust in automated sleep staging

💡 Explainable sleep staging with SleepVLM! 🛋️