Wesper Lab home sleep studies used for training and testing the Wesper autoscoring algorithm for the detection of central apnea events

The data is organized into two main folders: one containing Wesper HSAT studies and the other containing Wesper Clinical studies. The Wesper HSAT folder is further divided into two subfolders: one for training sessions and another for testing sessions. Each session is stored in its own folder, and within each session folder, there are two files: a CSV file containing the signals and another CSV file containing the scores.<br><br>Research Description:<br><br>Sleep apnea, a disorder characterized by repetitive disruptions in breathing during sleep, is traditionally diagnosed through in-laboratory polysomnography (PSG). Despite the growing popularity of non-invasive home sleep apnea tests (HSATs) due to their convenience, these devices often encounter limitations in accurately detecting central sleep apnea (CSA). These challenges stem from factors such as suboptimal signal quality in home environments, inaccuracies in surrogate signal interpretations, and variability in manual scoring. To address these issues, this study evaluated a novel HSAT device, Wesper Lab, which utilizes direct respiratory signal measurements and integrates artificial intelligence (AI) for automated scoring. Specifically, the accuracy of the Wesper Autoscoring Algorithm (WAA) in identifying central respiratory events, including the Central Apnea Index (CAI), was compared to Manual Human Scoring (WAA).<br>