Obstructive Sleep Apnea Prediction Data

The database consists of two full datasets. The first dataset is the collection of 20 records gathered from 14 OSA patients as part of the Apnea-ECG Database.The second dataset consists of 10 records from 10 subjects, who were selected based on the quality of the signals, from St. Vincent’s University Hospital/ University College Dublin (UCD Database) sleep apnea database (http://www.physionet.org/pn3/ucddb/). The OSA annotations are in minute-wise scored offline by experienced sleep physicians using polysomnography (PSG) signals. The PSG signals recorded were EEG, EOG, EMG, ECG, nasal airflow, ribcage movements, abdomen movements, oxygen saturation (finger pulse oximeter), snoring (tracheal microphone), and body position. All subjects have been clinically diagnosed with mild to severe OSA.