1.2.3. Polysomnography

A standard PSG recording (6 electroencephalographic [EEG] leads: F3/4, C3/4, O1/2 referenced to the contralateral mastoids; submental electromyogram [EMG]; bilateral electrooculogram [EOG]; electrocardiogram [ECG]) was performed according to the American Academy of Sleep Medicine (AASM) rules [30]. EEG, EOG and EMG were sampled at 256Hz. EEG and EOG were band-pass filtered at 0.3–35Hz and EMG was filtered at 10–100Hz. The ECG signal was collected at 512Hz using Ag/AgCl Meditrace surface spot electrodes in a modified lead II Einthoven configuration, and pass-band filtered at 0.3–70Hz. R-waves were automatically detected by a customized MatLab (MathWorks, Natick, MA, USA) algorithm and visually inspected; normal-to-normal interbeat-intervals (ms) calculated and heart rate derived (HR, bpm).

Sleep was scored (Wake, N1, N2, N3 and REM) in 30s epochs according to AASM criteria [30] and standard sleep parameters were calculated including: time in bed (TIB, min), sleep onset latency as the first epoch of any sleep stage (SOL, min), REM latency (REML, min), wake after sleep onset (WASO, min), total sleep time (TST, min), sleep efficiency (SE, TST/TIB*100, %), percentage of time spent in each stage of sleep calculated as a percentage of TST (N1, N2, N3 and REM, %), and total number of awakenings (awakening index) and arousals (arousal index) per hour of sleep.

During the clinical/adaptation night, bilateral anterior tibialis EMG (to assess leg movements), thoracic and abdominal piezoelectric bands (to measure respiratory effort), airflow using nasal cannula and thermistor, and oxygen saturation (pulse oximeter) were also collected and used to confirm the absence of sleep disorders.