Sleep Spindles as Predictors of Prognosis in Patients with Disorders of Consciousness

Disorders of consciousness (DoC) are abnormal states of awareness caused by severe brain injuries, commonly manifesting as coma, vegetative state (VS), or minimally conscious state (MCS). Assessing the consciousness state and predicting the prognosis of these patients remained a clinical challenge. This study proposed a prognostic prediction method based on sleep spindle power, using polysomnography (PSG) to extract sleep spindle features from DoC patients and applying statistical analysis and machine learning models to predict prognosis. Significant differences were found in spindle amplitude, RMS, absolute power, and relative power in the 10-13 Hz frequency band between the "improved" and "no improvement" groups. Machine learning models, such as support vector machines and random forests, demonstrated high accuracy in predicting consciousness state changes. The results highlighted the potential of sleep spindle features, especially spindle relative power, in predicting the prognosis of DoC patients, providing a novel and effective tool for clinical evaluation.Clinical relevance- This study demonstrated that the relative power of sleep spindles can accurately predict the prognosis of DoC patients.