PhysioEx

PhysioEx provides an extensible and modular API for standardizing and automating the sleep staging pipeline, covering data preprocessing, model training, testing, fine-tuning, and explainability using deep learning and Explainable AI techniques.

Key Features:

• Standardized pipeline: PhysioEx provides a standardized pipeline for sleep staging analysis with customizable workflows.
• Pretrained models: Includes pretrained models based on the Sleep Heart Health Study dataset supporting single-channel EEG and multichannel EEG-EOG-EMG configurations.
• Explainable AI methods: Offers a range of XAI post-hoc methods to explain model decisions and align them with expert knowledge, linking model decisions to human-understandable concepts such as American Academy of Sleep Medicine-defined rules.
• Benchmarking capability: Benchmarks state-of-the-art sleep staging models in a standard pipeline, enabling fair comparison between them both on the training source and out-of-domain sources.
• Modular architecture: Features an extensible and modular API that is easily adaptable to custom datasets.

Scientific Applications:

• Sleep disorder diagnosis: Supports sleep staging analysis for diagnosing and understanding sleep disorders in clinical and research contexts.
• Model comparison: Enables fair comparison of state-of-the-art sleep staging models on both training source and out-of-domain sources.
• Clinical interpretability: Provides explainable insights that bridge the gap between machine learning models and clinical expertise by linking model decisions to American Academy of Sleep Medicine-defined rules.

Methodology:

PhysioEx implements deep learning models for sleep staging with post-hoc Explainable AI methods; specific computational algorithms are not explicitly detailed in the sources.