COVISLEEP: A sleep dataset to investigate the impact of long-COVID on chronic insomnia.

This dataset contains polysomnography recordings of 93 patients suffering from chronic insomnia. Among them, 31 declared suffering from insomnia after experiencing long-COVID symptoms (identified by the 'S' letter before their ID), while the other 62 have not been infected by COVID-19 (identified by the 'T' letter). Sleep staging was performed by trained sleep physicians for each recording following the AASM guidelines.

A metadata file contains additional information about the subjects, such as their age, height, or type of insomnia.

The raw zip file contains unprocessed polysomnography recordings as EDF files. They were acquired through Nox A1 sensors and contain up to 80 channels. Each EDF file corresponds to a night of sleep for each patient.

For each raw file, there is a CSV file in the annotations zip files. Each CSV file contains the sleep stages, with its corresponding epoch ID and time.

Finally, the preprocessed zip file contains the preprocessed polysomnography signals, stored as NPY files.