N2 Stage Impact on REM Sleep Behavior Disorders Detection

Sleep is a natural, essential biological process in which the body and mind undergo rest and recovery. During sleep, consciousness is suspended, the brain cycles through different stages, and vital physiological processes occur that support physical health, emotional well-being, and cognitive function. Sleep architecture includes several stages, each contributing uniquely to overall sleep quality and neurological functions. The sleep cycle is divided into the Rapid Eye Movement (REM) stage and non-REM (NREM) stages N1, N2, and N3. Among these, REM sleep plays a crucial role in dreaming and neurological health. Disruptions in this stage can lead to specific sleep disorders. One such disorder, Rapid Eye Movement Sleep Behavior Disorder (RBD), is a type of parasomnia. It is marked by a lack of normal muscle relaxation during the REM sleep stage, causing individuals to physically act out their dreams.This behavior may involve talking, shouting, limb movements, or even violent actions like punching and kicking. RBD not only disrupts sleep quality but also presents risks of injury to the individuals themselves and their bed partners. RBD can be an early sign of diseases like Parkinson's and Lewy body dementia, so detecting it early is important. While REM sleep is directly implicated in RBD, recent research shows that the N2 stage may also hold an important role in detecting sleep disorders. The N2 stage makes up a large part of total sleep time. It is characterized by specific Electroencephalography (EEG) features, including sleep spindles and K-complexes, which are linked to memory consolidation and sensory processing. Understanding the impact of the N2 stage on RBD detection is necessary for several reasons. First, changes in N2 stage patterns may serve as early signs of sleep disturbances that either come before or occur alongside RBD. Second, including N2 stage analysis could improve the accuracy of diagnostic methods for RBD, enabling earlier treatments. Finally, examining the relationship between the N2 and REM stages may provide a greater understanding of the underlying mechanisms of RBD and its progression toward neurodegenerative conditions. To investigate the impact of the N2 sleep stage on detecting RBD, we performed a detailed analysis beginning with the extraction of 236 features across the time, frequency, time-frequency, and nonlinear metrics. From these, the 5 most important features were selected using the XGBOOST feature selector. We then applied machine learning classifiers to EEG data, using methods such as K-Nearest Neighbors, Random Forest, Decision Tree, and Gaussian Naive Bayes to analyze EEG features from the N2, N3, and REM sleep stages, both individually and in various combinations. Our findings revealed that using only the N2 sleep stage achieved an overall accuracy of 85%, with a sensitivity of 90% for the RBD class. When combining N2 and REM stages, we reached a maximum overall accuracy of 85% and a sensitivity of 100% for the RBD class, outperforming models that considered REM alone. These results underscore the valuable role of the N2 sleep stage in detecting RBD. Including N2 stage analysis enhances the sensitivity and specificity of diagnostic methodologies, enabling earlier interventions.