Background: Polysomnography (PSG), the gold standard diagnostic tool in sleep medicine, typically evaluates sleep patterns and sleep-related disorders in an artificial environment, potentially leading to distorted sleep. Sleep macro-structures such as sleep stages are particularly sensitive to environmental effects, whereas micro-structures (i.e., sleep spindles, REM sleep without atonia, RSWA) tend to remain more stable.
Objective: This study aimed to apply automated algorithms to capture RSWA and sleep spindles and to compare between values measured in the lab and those recorded under natural sleep conditions at home.
Methods: The analysis included 107 full-night recordings from 55 subjects: 24 healthy adults , 28 patients with Parkinson’s disease (PD), and three individuals with isolated Rem sleep behavior disorder (RBD). The PD population included 15 PD patients with RBD and 13 PD patients without RBD. All sessions were manually scored for sleep staging. An automatic algorithm for quantifying RSWA was developed and tested against manual scoring. Automatic artifact removal was incorporated into the process.
Results: Home versus lab RSWA, derived using the semi-automatic algorithm were moderately correlated (60% correlation).RBD detection in the home environment was achieved with 83% sensitivity, 79% specificity, and 81% balanced accuracy.
Discussion: A semi-automated algorithm was found to accurately quantify RSWA in the lab and in the home environment, enabling the detection of RBD patients. These findings could enhance clinical research, facilitate more frequent and accessible sleep testing, and provide a possible alternative for screening RBD.