Polysomnographic and neuropsychological characteristics of rapid eye movement sleep behavior disorder patients

Abstract Objectives To compare the sleep characteristics and cognition between rapid eye movement sleep behavior disorder (RBD) patients and non‐RBD (nRBD) healthy control subjects and to determine the correlation between sleep and cognition in RBD patients. Methods Polysomnography (PSG) was performed to confirm and exclude RBD. Fifteen iRBD patients, 12 PD with RBD patients, and 23 matched nRBD healthy control subjects were enrolled. Subjective sleep and neuropsychological evaluations were performed. Results No differences were found in the subjective sleep among the three groups. Compared to the nRBD subjects, iRBD patients showed higher PLMI, the PD with RBD patients showed an increased proportion of N1 sleep, a decreased proportion of N2 and N3 sleep, and an increased periodic leg movement index. The iRBD patients performed worse on RCFT time than the nRBD subjects, the PD with RBD patients performed worse than the nRBD subjects on the following evaluations: the Mini‐Mental State Examination; auditory verbal learning test (AVLT); Rey Complex Figure Test (RCFT) time, Clock drawing test (CDT); delay memory score of RCFT; Symbol digit modalities test (SDMT); Trail Making Test (TMT); and Stroop Test B and C, all significant changes (all p < 0.05). Furthermore, in RBD patients, lower sleep efficiency was correlated with decreased SDMT scores (r = 0.694, p < 0.01), longer time on the TMT A (r = −0.589, p < 0.01), and lower city fluency test scores(r = 0.556, p < 0.01). Less total sleep time was correlated with lower RCFT scores (r = 0.392, p = 0.043), longer time on the TMT A (r = −0.417, p = 0.031), and lower city fluency test scores (r = 0.405, p = 0.036). Conclusions PD with RBD patients suffered from abnormal sleep architecture and extensive cognition dysfunction. Decreased total sleep time and sleep efficiency may contribute to cognitive deterioration in RBD patients.


| INTRODUC TI ON
Rapid eye movements, cortical activation, vivid dreaming, skeletal muscle paralysis (atonia), and muscle twitches are the characteristics of rapid eye movement (REM) sleep. Loss of normal muscle atonia during REM sleep leads to a parasomnia called REM sleep behavior disorder (RBD). During REM sleep, RBD patients act out their dreams violently and forcefully, which may lead to disturbed sleep or injuries to themselves or their bed partners.
The principal cognitive deficits of RBD patients involve visuospatial skills, attention, execution, and memory (Hogl et al., 2018).
Moreover, studies have suggested that mild cognitive impairment (MCI) in RBD may be an early marker of neurodegeneration (Gagnon et al., 2009;Sasai, Matsuura, & Inoue, 2013). However, few studies have examined the features and quality of sleep in RBD patients. It is well-known that sleep plays an important role in memory consolidation. Sleep duration, timing, and continuity can affect cellular ultrastructure, gene expression, metabolic and hormone regulation, mood, and vigilance (Van Someren et al., 2015).
Few studies have focused on the correlation between sleep characteristics and cognition in RBD patients. However, the question of whether disrupted sleep contributes to cognitive impairment in RBD patients remains unanswered. Therefore, the aim of this study was to compare the sleep characteristics and neuropsychological performances of Chinese RBD patients using overnight video PSG and sleep and neuropsychological evaluations focusing on the following:

| Subjects
Participants were recruited from Tianjin Medical University General Hospital from October 2014 to May 2017. The RBD patients were diagnosed according to the standards described in the International classification of sleep disorders (ICSD-3) criteria (Sateia, 2014

| Polysomnography examination
All the subjects were observed with overnight video PSG using a digital sleep laboratory system (Nicolet v32, Natus Medical Incorporated, Pleasanton, CA). Participants were instructed to go to sleep at their usual bedtimes. All PSG sessions were monitored by a trained technician according to standardized criteria (Hakkinen et al., 1993). The following sleep variables were acquired and ana-  including the total score and the short delay recall, long delay recall, and recognition error scores (Guo, Zhao, Chen, Ding, & Hong, 2009); (c) visuospatial skills were evaluated with the Rey Complex Figure Test (RCFT) (copy and time) (Caffarra, Vezzadini, Dieci, Zonato, & Venneri, 2002) and the Clock drawing test (Rouleau, Salmon, Butters, Kennedy, & McGuire, 1992), and visual memory was evaluated with the RCFT delay memory score (Caffarra et al., 2002); (d) verbal information processing was evaluated with the Symbol Digit Modalities Test (SDMT) (Strauss, Sherman, & Spreen, 2006); (e) executive function was evaluated with the Trail Making Test and (f) language was evaluated according to semantic verbal fluency (Strauss et al., 2006) (animal fluency test, city fluency test, and animal-city alternation fluency test) and naming ability [the Boston Naming Test (Chiu et al., 2000)].

| Statistical analysis
The results are expressed as the mean ± standard error (SE) for continuous variables and as the probability (percent) for categorical variables.
Differences in the demographic and clinical characteristics among the three groups were compared descriptively using chi-square tests for categorical measures; using one-way ANOVA tests for data that were normally distributed and which homoscedasticity was respected; using nonparametric Kruskal-Wallis or Mann-Whitney U tests for variables that were not distributed normally or for which homoscedasticity was not respected. Neuropsychological test scores and PSG parameters in RBD participants were converted into z scores using the mean and SDs of the RBD participants' scores before correlation analysis. Linear models were applied to compute the Pearson correlation coefficient (r) between sleep and neuropsychological parameters in the RBD group.
Statistical significance was defined as p < 0.05. The statistical analyses were performed using SPSS 16.0. Figures were created using GraphPad Prism Version 5 (San Diego, USA).

| Demographics, self-report questionnaires, and overnight sleep studies of the three groups
The demographic and clinical characteristics are reported in Table 1. There were no differences in age, gender, or education between the three groups. The average RBD duration of iRBD and PD with RBD patients was 5.77 and 11.46 years, respectively.
The results from the self-report questionnaires and polysomnographic parameters are reported in Table 2. The subjective sleep evaluations indicated no significant differences in daytime sleepiness, night insomnia, and quality of sleep among the three groups. PD with RBD patients had a higher proportion of N1 (32.33 vs. 15.61, p < 0.01), lower proportion of N2 (52.0 vs. 63.39, p < 0.01), lower proportion of N3 (2.92 vs. 6.87, p < 0.05) than the nRBD healthy control subjects. IRBD patients had a higher PLMI than PD with RBD patients and nRBD healthy control subjects (26.55 vs. 13.58 vs. 4.87, p < 0.05, p < 0.01), lower proportion of N1 than PD with RBD patients (19.93 vs. 32.33, p < 0.05). No differences were found in the TST, SL, REM SL, SE, proportion of REM sleep, arousal index, average SpO 2 , minimum SpO 2 , or AHI among the three groups.

| Cognitive deficits in the iRBD group and PD with RBD group
The iRBD patients showed deficit on RCFT time (p = 0.043).
PD with RBD patients had significantly lower scores than the nRBD healthy control subjects on the following evalua-  Table 3.

| Correlation between sleep parameters and neuropsychological parameters in the RBD group
Lower sleep efficiency was correlated with decreased SDMT scores

| D ISCUSS I ON
Our study demonstrated the following results: (a) sleep architecture was disrupted in patients with PD with RBD patients, as indicated by the increased percentage of N1, the decreased percentage of N2 and N3, and the higher PLMI; (b) iRBD showed higher PLMI than the nRBD healthy control subjects; (c) PSG examination could be used to iRBD (n = 15) PD + RBD (n = 12) nRBD (n = 23) p Age  Polysomnographic characteristics in RBD patients are controversial. Our data are consistent with those of a previous study of Chinese RBD patients that showed similar results, with more stage N1 sleep, less stage N2 and N3 sleep, and a higher PLMI than observed in the controls (Zhou et al., 2014). The SE of the RBD group appeared to be less than that of the nRBD group, but there was no significant difference between the two, most likely due to the small sample size. In contrast, a reanalysis found that the percentage of slow-wave sleep (SWS) increased in RBD patients (Schenck, Callies, & Mahowald, 2003). The younger iRBD group (younger than 70 years) showed a significantly lower proportion of N3, while the older iRBD group (70 years or older) showed a significantly higher proportion of N3 than the control patients (Sasai et al., 2013 some studies with small sample sizes that did not find differences between iRBD patients and controls in polysomnographic characteristics, except in the PLMI (Fantini et al., 2011;Ferini-Strambi et al., 2004) and sleep duration (Massicotte-Marquez et al., 2008). IRBD is considered a possible marker of neurodegenerative diseases and does not include RBD combined with neurodegenerative diseases.
In our study, no difference was found in sleep architecture between iRBD and nRBD subjects, while PD with RBD patients showed more   (Raggi & Ferri, 2010). Previous studies of the relationship between circadian/sleep disruption and neurodegenerative disorders have supported that neurodegeneration may impact the brain centers that control sleep and circadian behavior; sleep and circadian disruption may also lead to oxidative damage, metabolic disruption, and decreased clearance of metabolites, such as β-amyloid, all of which may accelerate neurodegeneration (Mattis & Sehgal, 2016). Thus, sleep disruption in RBD patients should be treated in the clinic and managed if possible.
Cognitive deficits in RBD have been frequently reported in previous studies. MCI was found in 50% of iRBD patients and 73% of PD patients with RBD, whereas this condition was observed in only 11% of PD patients without RBD and 8% of control patients (Gagnon et al., 2009). RBD is an important risk factor for MCI, as deficits in executive function, verbal delayed memory and visuospatial function have been consistently associated with more severe RBD symptoms (Zhang et al., 2016). IRBD patients have been shown to perform more poorly on the word span, Rey-Osterrieth Complex Figure Recall, digit span and logic memory tests, and on visuo-constructional learning, sharing common features of cognitive deficits with Lewy body disease in particular (Terzaghi et al., 2008). But we found that only iRBD patients performed worse on RCFT time than the nRBD healthy control subjects, the reason may be that the RBD duration of iRBD patients is shorter in our study. The deterioration in nonverbal logic, attention, executive function and memory observed in RBD follow-ups suggests an underlying evolution of the degenerative process (Terzaghi, Zucchella, Rustioni, Sinforiani, & Manni, 2013;Youn et al., 2016). In our study, multidomain cognitive impairment, including in verbal memory, visuospatial abilities and memory, verbal information processing, execution and language, was found in PD with RBD patients, more extensive domains of impaired cognition than iRBD patients. This indicated that PD with RBD patients exhibit more serious and extensive neurodegeneration from brainstem to mesocortex and neocortex than iRBD patients. Several pathological bases of cognitive impairment in RBD patients have been proposed. First, cholinergic dysfunction is thought to be associated with cognitive decline in iRBD (Nardone et al., 2012). Second, many imaging studies have supported abnormalities in both structure and function in brain regions other than the brainstem. Decreased GMV of the left posterior cingulate, hippocampus (Lim et al., 2016) and superior frontal sulcus (Rahayel et al., 2015) and decreased cortical thickness in the frontal cortex, lingual gyrus, and fusiform gyrus  was detected using single-photon emission computed tomography (Vendette et al., 2012). Furthermore, cortical and subcortical GM abnormalities are associated with cognitive status in patients with RBD, with more extensive patterns in patients with MCI (Rahayel et al., 2018). Imaging abnormalities in RBD patients explain the impairment in information extraction, visuospatial abilities and executive function. Third, lack of delta-band functional connectivity (Sunwoo et al., 2017) and electroencephalographic slowing, especially during REM sleep (Sasai et al., 2013), is associated with cognitive decline in iRBD.
However, no studies have considered that disturbed sleep may lead to deterioration in cognition among RBD patients. The relationship between sleep disorders and cognition has been widely studied. Sleep promotes learning-dependent synapse formation and maintenance of selected dendritic branches, both of which contribute to memory storage (Yang et al., 2014). Convective fluxes in interstitial fluid have been shown to increase the rate of β-amyloid clearance during sleep (Xie et al., 2013 (Rasch & Born, 2013). A correlational analysis showed that lower SE was correlated with worse performance in verbal information processing, execution and language fluency, and TST was correlated with visuospatial abilities, execution and language fluency in RBD patients. Based on previous studies of sleep and cognition, we speculated that sleep abnormalities in RBD patients may contribute to cognitive deterioration.
There were some limitations to our study. First, the sample size of the study was small; thus, differences in some of the PSG and neuropsychological parameters may not have been detected. Second, this was a cross-sectional study. A study with a large sample and a longitudinal study should be conducted to obtain more reliable and robust findings in the future.

| CON CLUS ION
Our study identified increased N1 sleep, decreased N2 and N3 sleep, a higher PLMI and impairments in multidomain cognitive functions, including general cognition, verbal memory, visuospatial abilities and memory, verbal information processing, execution and language, in PD with RBD patients. The sleep disturbance was more serious and the cognitive impairment was more widespread than the iRBD patients. In RBD patients, lower SE was

ACK N OWLED G M ENTS
This study was supported by the Major project of Tianjin science and technology, prevention and control technology of chronic disease (17ZXMFSY00180).