A multimodal dataset for training deep learning models aimed at detecting and analyzing sleep apnea

Area of Science:

• Respiratory Medicine
• Medical Informatics
• Artificial Intelligence in Healthcare

Background:

• Sleep Apnea Syndrome (SAS) is a significant respiratory disorder with severe health implications like hypertension and cognitive impairment.
• The subtle nature of SAS symptoms often leads to delayed diagnosis, while clinical screening is resource-intensive.
• Existing diagnostic methods for SAS can be time-consuming and require specialized equipment.

Purpose of the Study:

• To introduce a novel, comprehensive dataset for Sleep Apnea Syndrome (SAS) research.
• To facilitate the development and application of deep learning models for SAS diagnosis.
• To provide a standardized, high-quality, and publicly available data resource for the scientific community.

Main Methods:

• Integration of data from Polysomnography (PSG) devices with synchronized audio recordings.
• Rigorous annotation of the dataset by expert medical professionals based on PSG monitoring.
• Creation of a publicly accessible, standardized data resource.

Main Results:

• A high-quality, annotated dataset combining PSG and synchronized audio data for SAS.
• The dataset is designed to support the training and validation of deep learning algorithms.
• Enables more accurate and efficient diagnostic approaches for Sleep Apnea Syndrome.

Conclusions:

• The developed dataset is a valuable resource for advancing research in Sleep Apnea Syndrome.
• It is expected to enhance diagnostic accuracy and efficiency through deep learning applications.
• Promotes scientific innovation in the detection and management of SAS.