Abstract
Objective: To investigate micro‐sleep stability in patients with Parkinson's disease (PD) with REM sleep behavior disorder (PD+RBD), PD patients without RBD (PDnoRBD), patients with isolated RBD (iRBD) and neurologically healthy controls (HC) using an automated algorithm.
Background: REM and NREM sleep stability has been seen to progressively decrease in iRBD and PD when using a data‐driven method to score 30‐s sleep epochs [1]. No study has analyzed micro‐sleep stability in relation to neurodegeneration.
Methods: A total of 30 de novo PD+RBD patients, 57 de novo PDnoRBD patients, 31 patients with iRBD, and 41 HC were investigated with video‐polysomnography. For micro‐sleep staging, we used a validated algorithm that scored each 5‐s sleep mini‐epoch as either wakefulness (W), REM or NREM sleep [2]. For each mini‐epoch, the algorithm returned a prediction confidence score. To reduce the algorithm bias we removed from further analysis the mini‐epochs predicted with low confidence. The following micro‐sleep stability indices were computed: wake‐sleep (W‐S) transition index (the number of transitions from wake to sleep, or vice‐versa, per total sleep time), and the REM and NREM fragmentation indices (the number of transitions from REM or NREM to another stage per hour of REM or NREM sleep respectively).
Results: After removing mini‐epochs where the sleep stage prediction had low confidence, the micro‐sleep stability indices were analysed in 90.05±4.15% of the total mini‐epochs in HC, 88.87±7.86% in PDnoRBD, 81.61±7.38% in iRBD and 81.10±8.00% in PD+RBD. All indices were significantly higher in iRBD and PD+RBD patients than in HC and PDnoRBD (Table 1).
Conclusions: Significantly increased micro‐sleep instability in iRBD and PD+RBD compared to PDnoRBD and HC suggests that RBD may compromise the networks responsible for keeping sleep stable. The pathophysiology of micro‐sleep instability should be investigated in future studies.
Background: REM and NREM sleep stability has been seen to progressively decrease in iRBD and PD when using a data‐driven method to score 30‐s sleep epochs [1]. No study has analyzed micro‐sleep stability in relation to neurodegeneration.
Methods: A total of 30 de novo PD+RBD patients, 57 de novo PDnoRBD patients, 31 patients with iRBD, and 41 HC were investigated with video‐polysomnography. For micro‐sleep staging, we used a validated algorithm that scored each 5‐s sleep mini‐epoch as either wakefulness (W), REM or NREM sleep [2]. For each mini‐epoch, the algorithm returned a prediction confidence score. To reduce the algorithm bias we removed from further analysis the mini‐epochs predicted with low confidence. The following micro‐sleep stability indices were computed: wake‐sleep (W‐S) transition index (the number of transitions from wake to sleep, or vice‐versa, per total sleep time), and the REM and NREM fragmentation indices (the number of transitions from REM or NREM to another stage per hour of REM or NREM sleep respectively).
Results: After removing mini‐epochs where the sleep stage prediction had low confidence, the micro‐sleep stability indices were analysed in 90.05±4.15% of the total mini‐epochs in HC, 88.87±7.86% in PDnoRBD, 81.61±7.38% in iRBD and 81.10±8.00% in PD+RBD. All indices were significantly higher in iRBD and PD+RBD patients than in HC and PDnoRBD (Table 1).
Conclusions: Significantly increased micro‐sleep instability in iRBD and PD+RBD compared to PDnoRBD and HC suggests that RBD may compromise the networks responsible for keeping sleep stable. The pathophysiology of micro‐sleep instability should be investigated in future studies.
| Originalsprog | Engelsk |
|---|---|
| Artikelnummer | 589 |
| Tidsskrift | Movement Disorders |
| Vol/bind | 34 |
| Udgave nummer | S2 |
| Sider (fra-til) | S240 |
| ISSN | 0885-3185 |
| Status | Udgivet - 2019 |