Effects of Deep Brain Stimulation on Sleep-Wake Disturbances in Patients with Parkinson's Disease: A Narrative Review

Abstract

Sleep-wake disturbances (SWD) are one of the most common non-motor symptoms in the Parkinson's disease (PD) and can appear in the early stage, even before the onset of motor symptoms. Deep brain stimulation (DBS) is an established treatment for the motor symptoms in patients with advanced PD. However, the effect of DBS on SWD and its specific mechanisms are not widely understood and remain controversial. In addition to the circuit-mediated direct effect, DBS may improve SWD by an indirect effect, such as the resolution of nocturnal motor complications and a reduction of dopaminergic medication. Here, the authors review the recent literatures regarding the impact of DBS on SWD in patients with PD. Furthermore, the selection of the DBS targets and the specific effects of applying DBS to each target on SWD in PD are also discussed.

1. INTRODUCTION

Sleep-wake disturbances (SWD) are one of the most common non-motor symptoms in Parkinson's disease (PD), with an estimated prevalence of up to 98%, which can appear in the early stage of PD, even before the onset of motor symptoms [1, 2]. The earliest description of the “Essay on the Shaking Palsy” by James Parkinson included disturbed sleep [3]. Because SWD negatively affects the quality of life of patients and their caregivers [4, 5], understanding SWD is essential to manage PD patients.

Although, the etiology of SWD in PD is not clearly understood, it seems to be multifactorial. In addition to nocturnal sleep disruption due to aggravation of motor symptoms, coexisting non-motor symptoms, such as frequent nocturia, pain, anxiety, and depression and concomitant primary sleep disorders can contribute to SWD in PD [6, 7]. PD pathology affects several brainstem structures and neurotransmitters related to the regulation of sleep [8, 9]. It seems that the degeneration of the sleep-wake regulation-related neurons, associated with the progression of the disease itself, has a considerable impact on SWD in PD. Sleep-wake and rapid eye movement (REM) sleep control systems are damaged in PD [10]. Neuronal loss in arousal systems, including the noradrenergic locus ceruleus, the serotonergic raphe, the cholinergic pedunculopontine nucleus (PPN) and the orexin neurons in the hypothalamus, is observed in the PD brain [11]. Control of the NREM to REM sleep transition and the descending control of muscle tone may be disrupted by alpha-synuclein deposition in PD, which is suggested to induce REM without atonia, or REM sleep behavior disorder (RBD), that precedes the onset of motor symptoms of PD [12-15]. Furthermore, dopaminergic medications may also have an important role in the development of SWD. Levodopa appears to have a differential effect on sleep architecture depending on the dosage [16]. At a low dose, it promotes slow-wave sleep and REM sleep, possibly inducing somnolence [16]. On the other hand, levodopa reduces slow-wave sleep and REM sleep and increasing alertness at a high dose [16]. Dopamine agonists commonly induce excessive daytime sleepiness (EDS) and sleep attack [17, 18].

Deep brain stimulation (DBS) is an established treatment for the motor symptoms in patients with advanced PD. It is suggested that DBS disrupts abnormal information within the cortico-basal ganglia-thalamic neural circuits through its excitatory and inhibitory effects on the targeted nucleus, which results in motor improvement [19, 20]. In several other neurological and psychiatric disorders such as dystonia, Alzheimer’s disease, refractory obsessive-compulsive disorders, eating disorders and addiction, DBS is emerging as an alternative therapeutic strategy by modulating circuit dysfunction [21-25]. Various mechanisms have been suggested for the improvement of SWD after DBS. DBS may improve SWD by an indirect effect, such as the resolution of nocturnal motor complications [26-28] and a reduction in dopaminergic medication [29, 30] in addition to a circuit-mediated direct effect [31, 32]. Hjort et al. suggested that the improvement of SWD is affected by reducing motor symptoms rather than throughcentral sleep modulation [26]. However, a polysomnography (PSG) study showed significant sleep improvement after DBS, despite a rather modest motor improvement at the time of the recording, and the authors suggested that a direct effect of the stimulation on sleep regulatory centers cannot be excluded [33].

The purpose of this narrative review is to examine the effect of DBS on SWD in patients with PD in the literature. We conducted a literature search on PubMed using a combination of the following keywords: “Parkinson’s disease”, “deep brain stimulation”, “sleep”, “sleep disturbances”, “sleep quality”, “REM sleep behavior disorder”, “sleep apnea”, “restless leg syndrome”, “periodic limb movements”, “excessive daytime sleepiness”, “subthalamic nucleus”, “pedunculopontine nucleus”, “globus pallidus”, “ventral intermediate nucleus.” Case studies, clinical trials and reviews published between 2000 and 2020 were all reviewed. Furthermore, the selection of the DBS targets and the specific effects of applying DBS to each target on SWD in PD are also discussed.

2. IMPACT OF DBS ON THE OVERALL SLEEP

2.1. Subjective Sleep Quality

For the evaluation of subjective sleep quality, the Parkinson Disease Sleep Scale (PDSS) [34], PDSS-2 [35], and Pittsburgh Sleep Quality Index (PSQI) [36] were commonly used in patients with PD Table (1). Thirteen subthalamic nucleus (STN)-DBS studies used PDSS or PDSS-2 for the assessment of the overall sleep quality [26, 37-48]. Mean follow-up ranged from 3 months [26] to 5 years [43]. Most of the studies showed a significant improvement in the PDSS or PDSS-2 after STN-DBS. However, regarding the long-lasting effect over 5 years after surgery, a recent Chinese study failed to show a significant change in sleep quality measured by the PDSS-CV (Chinese version) [43]. The “rebound phenomenon”, namely the prominent increment of the total PDSS score in the first 6 months after STN-DBS followed by a gradual decrement, was observed in long-term serial follow-up studies [37-40]. In the studies analyzing the PDSS subscores, significant improvements were commonly observed in the overall sleep quality and nocturnal motor symptoms, whereas nocturia and daytime sleepiness showed inconsistent results [26, 37, 39, 41, 44]. In addition, two studies reported a significant relationship between sleep and quality of life (QoL) improvements at 24 months [37] and 36 months [39] after STN-DBS, highlighting the importance of sleep quality for overall clinical status expressed as QoL. There have also been five studies that usedPSQI to assess the impact of STN-DBS on subjective sleep quality in PD patients. Most studies showed an improvement of the total PSQI score with STN-DBS [33, 49-51]. In contrast, a recent study by Torun et al. showed no improvement in the total PSQI score, 3 months after STN-DBS. However, the subscales of the PSQI revealed that the subjective sleep latency was significantly decreased, and sleep duration was significantly increased [52].

Table 1.

The impact of DBS on sleep quality.

Refs.	DBS Target	n	Follow-up(After DBS)	Scale	Mean Pre-operativeTotal Score	Mean Post-operative Total Score	Mean Changein Total Score(%)
Hjort et al.(2004) [26]	STN	10	3 months	PDSS	79.8	105.3	+31.9*
Dafsari et al.(2020) [37]	STN	66	24 months	PDSSa	90.0	98.9	+9.9*
Jung et al.(2020) [38]	STN	33	36 months	PDSS	90.0	98.9	+13.4*
Choi et al.(2019) [39]	STN	46	36 months	PDSS	79.0	93.3	+18.1*
Dulski et al.(2019) [40]	STN	36	12 months	PDSS	80.6	87.9	+9.1*
Breen et al.(2015) [41]	STN	11	6 months	PDSS	95.9	113.2	+18.0*
Nishida et al.(2011) [42]	STN	10	Not specified	PDSS	96.0	113.5	+18.2*
Jiang et al.(2015) [43]	STN	10	60 months	PDSS-CV	103.4	113.4	+9.7
Chahine et al.(2011) [44]	STN	17	6 months	PDSS	94.2	124.9	+32.6*
Peppe et al.(2012) [45]	STN	5	12 months	PDSS	70.0	98.0	+40.0*
	STN + PPN (on)				70.0	123.0	+75.7*
	STN + PPN (cyclic-on)				70.0	126.0	+80.0*
Odekerken et al.(2013) [46]	STN	63	12months	PDSS	81.3	94.7	+16.6*
	GPi	65			83.6	90.8	+8.6*
Deli et al.(2015) [47]	STN	13	12 months	PDSS-2b	24.8	14.2	-42.7*
Ricciardi et al.(2019) [48]	STN	18	24 months	PDSS-2	22.9	11.1	-51.5*
Monaca et al.(2004) [33]	STN	10	3 months	PSQIc	11.7	5.3	-54.7*
Amara et al.(2012) [49]	STN	53	6 months	PSQI	9.3	7.9	-14.7*
Iranzo et al.(2002) [50]	STN	11	6 months	PSQI	14.8	5.4	-63.5*
Liu et al.(2013) [51]	STN	8	26 months	PSQI	10.8	4.8	-56.3*
Torun et al.(2019) [52]	STN	8	3 months	PSQI	7.1	6.5	-8.5

^*Significant change (p < 0.05); Positive values indicate an increase and negative values indicate a decrease in the parameter, respectively.

^aThe total score of the PDSS has a maximum value of 150, and higher scores mean better sleep.

^bThe total score of the PDSS-2 has a maximum value of 60, and higher scores mean more nocturnal disturbances.

^cThe total score of the PSQI has a maximum value of 21, and values higher than 5 indicate poor sleep quality.

Abbreviations: DBS = Deep brain stimulation; STN = Subthalamic nucleus; GPi = Globus pallidus interna; PPN = Pedunculopontine nucleus; SNr = Substantia nigra pars reticulata; PDSS = Parkinson’s disease sleep scale; PDSS-CV = Parkinson’s disease sleep scale-Chinese version; PSQI = Pittsburgh sleep quality index.

Peppe et al. assessed the sleep parameters after STN- and PPN-DBS in three different settings (PPN in OFF, ON, or cyclic ON) with STN-DBS constantly ON [45]. In PPN-ON, the PPN was stimulated 24 hours/day and in PPN- cyclic ON, the PPN was switched ON only at night (12 hours/day, from 20:00 to 8:00) [45]. In that study, PD patients reported a marked improvement in the PDSS score for all DBS settings, with the PPN cyclic ON setting producing the greatest improvement [45]. When comparing the effect of simultaneous DBS of the STN and SNr (substantia nigra pars reticulata) to conventional STN-DBS, the overall PDSS-2 scores were not significantly different between the two groups [53]. Oderkerken et al. reported that the total PDSS score is not significantly different between groups treated with STN- and globus pallidus interna (GPi)-DBS [46].

2.2. Sleep Architecture

Objective assessment of sleep using PSG has been increasingly performed in recent years Table (2), and most studies have reported a positive impact on several sleep architectures, including an increase of the total sleep time and sleep efficiency and a decrease of wake after sleep onset after STN-DBS [27, 31, 33, 42, 50, 54]. On the other hand, Dulski et al. reported a deterioration of objective sleep parameters at 6 months after STN-DBS, characterized by a decrease in total sleep time and N2 sleep and an increase in arousals, in spite of an improvement of subjective sleep parameters [40]. Furthermore, it is generally considered that REM sleep features are not affected by STN-DBS, although a reduction in REM sleep latency with STN-DBS ON was reported in two studies [31, 52].

Table 2.

The impact of DBS on sleep architecture assessed by PSG.

Refs.	DBS Target	n	Follow-up(After DBS)	Increased¶Parameters	Decreased¶Parameters	UnchangedParameters
Arnulf et al.(2000) [27]	STN	10	3 - 6 months	SE, TST, N2 sleep	WASO	RBD
Baumann-Vogel et al. (2017) [31]	STN	50	7.7 months	SE, TST, N3 sleep, PLMS	WASO, REM latency	Sleep latency, N1, N2, REM sleep, Awakening index, Number of body position changes, REM sleep without atonia, RBD, AHI
Monaca et al.(2004) [33]	STN	10	3 months	SE, TST, deep, slow-wave sleep, paradoxical sleep	-	Number of awakenings
Dulski et al.(2019) [40]	STN	24	6 months	N1 sleep, WASO, Sleep latency	SE, TST, N2 sleep	-
Nishida et al.(2011) [42]	STN	10	38 - 99 days	Number of sleep cycle including slow-wave sleep, REM sleep, REM sleep with atonia	WASO,REM sleep without atonia	PLMS
Iranzo et al.(2002) [50]	STN	11	6 months	Number of body position changes, Longest continuous sleep period	Arousal index	PLMS, REM sleep without atonia, RBD, AHI
Torun et al.(2019) [52]	STN	8	3 months	Number of body position changes, central AHI	REM latency	-
Cicolin et al.(2004) [54]	STN	5	3 months	SE	WASO, REM latency	PLMS, REM sleep without atonia, RBD, AHI
Lim et al.(2009) [57]	PPN	3	4 - 12 months	REM sleep	-	REM sleep without atonia, RBD

^¶ Statistically significant increased or decreased parameters are recorded.

DBS = Deep brain stimulation; PSG = Polysomnography; STN = Subthalamic nucleus; PPN = Pedunculopontine nucleus; SE = Sleep efficiency; TST = Total sleep time; WASO = Wake after sleep onset; REM = Rapid eye movement; RBD = REM sleep behavior disorder; PLMS = Periodic limb movement in sleep; AHI = Apnea-hypopnea index.

It was reported that low-frequency stimulation increased alertness, whereas high-frequency stimulation-induced NREM sleep in two PD patients who underwent PPN-DBS was reported [55]. In a patient with simultaneous stimulation of the STN and PPN, both STN-DBS alone and PPN-DBS alone improved sleep efficiency and sleep latency to a similar degree, but STN-DBS alone had a insignificant effect on REM sleep [56]. Strikingly, PPN-DBS was associated with an increase in REM sleep in that study [56]. In the same context, Lim et al. reported an increase in the total duration and percentage of REM sleep with PPN-DBS [57]. However, their study included only 5 patients, of which 2 were patients with progressive supranuclear palsy, therefore, sthe results cannot be generalized to PD patients with PPN-DBS [57].

In a small pilot trial by Tolleson et al., although statistically not significant, Gpi-DBS showed trends for improvement in several PSG measures, including sleep efficiency and sleep latency [58]. The continuous stimulation of the ventral intermediate nucleus of the thalamus (Vim) failed to modify sleep quality or architecture [59].

2.3. Excessive Daytime Sleepiness

EDS has been actively investigated in patients with PD, but studies on DBS-treated PD patients are relatively scarce. Eleven studies on the effects of DBS on EDS have been included in the present review [26, 28, 38, 39, 44, 45, 54, 60-63]. All the studies used the Epworth Sleepiness Scale (ESS) for the evaluation of EDS and EDS was defined as an ESS score of ≥ 11 [64]. We found one study that assessed EDS using an objective measure, such as the multiple sleep latency test and the maintenance of wakefulness test, but these objective measures were available only preoperatively, not postoperatively [60].

Although, there have been a few studies showing an improvement of the EDS after STN-DBS [44, 54, 60], most studies showed no changes in EDS after surgery [26, 28, 38, 39, 61-63]. Whereas patients with persistent EDS gradually decreased postoperatively, patients with newly developed or worsened EDS were added as the disease progressed, and there seemed to be no significant change in the overall number of patients with EDS [38, 62]. Moreover, an association between the decrease in dopaminergic medications and changes in the ESS score was not detected [38, 62].

In a study on PD patients who underwent STN- and PPN-DBS simultaneously, PPN-DBS produced not only a remarkable improvement in nighttime sleep, but also a significant amelioration of daytime sleepiness, unlike STN-DBS [45]. The marked improvement of SWD was still observed at oneyear after surgery, suggesting that PPN-DBS led to the reorganization of the neural activity of the mesencephalic region [45].

2.4. Proposed Mechanisms of the Impact of DBS on Overall Sleep

The subjective improvement in sleep quality might be explained by the improvement in nocturnal mobility, especially in the early stages after surgery. In addition, considering the “rebound phenomenon” reported in the long-term follow-up studies, the improvement of the subjective sleep quality could be partly accounted for, by the placebo effect of the surgery. The placebo effect has been widely documented in patients with PD [65], and the surgical interventions are associated with increased placebo responses compared to oral medications [66]. Whereas the sleep quality was not significantly correlated with mood but with nocturnal motor symptoms in a study by Dafsari et al. [37], however, a significant correlation between subjective sleep quality and mood, especially depression, was observed in the other two studies [39, 40]. In addition, the discrepancy between the improvement of the subjective sleep parameters and the deterioration of the objective sleep parameters could partially be explained by the recognized improvement of the mood [40]. A positive mood, when waking up was demonstrated to be predictive of an overestimation of the total sleep time, whereas a negative mood had the opposite effect [67, 68]. The change in mood may affect subjective sleep quality directly as well as indirectly through the placebo effect.

The etiology of EDS in PD seems to be multifactorial: prior nocturnal sleep disruption, degeneration of sleep-wake regulation related neurons associated with the progression of the disease itself, and the effect of dopaminergic medication [6-9, 17, 18, 69]. It remains unknown which factors contribute to EDS after DBS surgery, and multiple factors may affect the post-DBS changes of EDS. Changes in nocturnal manifestations, including mobility at night, nocturia and concomitant other sleep problems, can contribute to daytime somnolence in the substantial part, but other factors, such as a reduction in dopaminergic drugs or a direct effect of the electrodes stimulation, should be considered.

3. IMPACT OF DBS ON CONCOMITANT SLEEP DISORDERS

3.1. REM Sleep Behavior Disorder

Only a small number of studies assessed the effect of DBS in RBD, with most studies reporting no change in the prevalence of RBD before and after STN-DBS [27, 31, 33, 50, 54]. On the other hand, Nishida et al. suggested that STN-DBS helps to reduce RBD and has a beneficial effect on sleep disturbances, including improved daytime sleepiness by restoring normal REM sleep [42]. However, this study was limited by the small sample size and the authors did not evaluate the percentage of REM sleep atonia relative to the total REM sleep [42]. In contrast, a retrospective study by Kim et al. reported that the incidence of probable RBD increased after STN-DBS because de novo RBD developed, and pre-existing RBD persisted 1 year after surgery [70]. Considering that the levodopa equivalent daily dose (LEDD) reduction after DBS was greater in the de novo RBD group than in the non-RBD group; the decrease in the dopaminergic drug could unmask or contribute to the occurrence of RBD [70]. Although, this study included a large number of patients (n = 90), an important limitation is that the diagnosis of RBD was only based on a clinical interview, which was not confirmed by PSG [70].

An aforementioned study showed a significant increase in REM sleep in the PPN-DBS ON state [57]. However, in that study, dream enactment behaviors and REM without atonia persisted in 2 PD patients with RBD [57]. Moreover, PPN-DBS did not induce RBD in any PD or progressive nuclear palsy patients who did not have RBD before DBS surgery [57]. The effectiveness of PPN-DBS on RBD needs to be further evaluated.

3.2. Sleep-disordered Breathing

We found only two studies investigating the impact of DBS on sleep apnea/hypopnea assessed by PSG, and the results showed that the apnea-hypopnea index remained unchanged before and after STN-DBS [50, 54].

3.3. Restless Leg Syndrome

Considering that impaired central dopaminergic transmission is thought to be a common pathophysiologic mechanism of both Restless leg syndrome (RLS) and PD [71], an improvement of RLS in PD patients after DBS could be expected. However, there has been conflicting evidence regarding the effects of DBS on RLS in patients with PD. Studies evaluating RLS used the International Restless Legs Syndrome Study Group (IRLSSG) rating scale, and a high score indicatessevere RLS symptoms [72, 73].

Several studies reported a significant improvement in RLS symptoms after STN-DBS, despite a reduction of dopaminergic medications [44, 74, 75]. A recent study of 22 STN-DBS treated PD patients confirmed a marked improvement of the IRLSSG rating scale up to 2 years postoperatively, despite a decrease in LEDD by 34% [75]. There were no correlations between RLS symptom improvement and PD motor symptoms improvement or reduction in dopaminergic medications [75]. In contrast, one retrospective study showed that 11 out of 195 (5.6%) patients had new-onset RLS symptoms, 3 to 12 months after STN-DBS [76]. The reduction in dopaminergic medication after surgery may unmask symptoms of RLS [76]. However, in another prospective study, 6 out of 31 (19%) PD patients with de novo RLS 6 months after STN-DBS had a higher dose and a lower percentage of decrease in dopamine agonists compared to the PD patients without the emergence of RLS [77].

While the effect of STN-DBS on RLS is still controversial, several studies have shown evidence for the positive effect of GPi-DBS on RLS. In a small pilot trial on 5 PD patients, the IRLSSG rating scale improved from 9.8 ± 14.0 to 3.4 ± 4.7, but it was not statistically significant [58]. Okun et al. reported on a generalized dystonia patient whose RLS symptoms resolved after bilateral GPi-DBS and recurred after the removal of one of the devices due to the infection [78]. Moreover, another study suggested that GPi-DBS improves refractory idiopathic RLS [79]. Further studies are needed to elucidate the impact of pallidal DBS on RLS.

3.4. Periodic Limb Movements in Sleep

Three studies showed a non-significant post-operative deterioration in the periodic limb movements in sleep (PLMS) index [27, 42, 50]. Baumann-Vogel et al. reported that the PLMS indices almost doubled after STN-DBS even though the prevalence of RLS remained unchanged [31]. In contrast, a trend towards an improvement in the PLMS index at 3 months after STN-DBS was reported in another study, but there was no statistical significance [54].

There is a case report on the improvement of the PLM during GPi stimulation [80]. Intraoperatively, the patient's PLM responded to voltage stimulation, higher than those required for the improvement of his parkinsonian symptoms, suggesting that the pallidal stimulation directly ameliorates the PLM [80].

3.5. Proposed Mechanisms of the Impact of DBS on Concomitant Sleep Disorders

RBD is considered to result from a dysfunction of the PPN [81], having strong anatomical connections with the STN and SN, which are functionally impaired in PD [82]. The persistence of RBD after STN-DBS in most of the previous studies [27, 31, 33, 50, 54] suggests that STN stimulation does not restore the PPN activity, promoting muscle atonia during REM sleep. However, there is a case report in which RBD developed immediately after implantation for subthalamic stimulation, suggesting that some fiber tracts to the pontomedullary area, which is responsible for REM sleep with atonia, are damaged by STN-DBS [83]. Even though the role of the STN in REM sleep regulation has not been determined yet, the STN may have a partial role in the mechanism of RBD.

Given that dopaminergic medication is used to treat idiopathic RLS, it is intriguing that RLS symptoms improved after DBS despite the substantial reduction of dopaminergic medications. It has been suggested that STN stimulation modulates the basal ganglia outflow, which decreases the downstream disinhibition at the spinal level, thus alleviating abnormal sensations and motor restlessness [84]. Another possible explanation is that pre-operative RLS symptoms during nighttime were non-motor sensory symptoms, fluctuating with motor symptoms rather than true RLS symptoms, which improved with motor fluctuations. On the other hand, in terms of de novo RLS after STN-DBS, the reduction in dopaminergic medication after surgery may unmask symptoms of RLS [76]. Another hypothesis has been proposed that overstimulation resulting from the cumulative effects of the dopamine agonists and STN-DBS may induce a change in the spinal cord excitability, leading to an emergence of RLS by a mechanism similar to that suggested for RLS augmentation [77].

While the effect of STN-DBS on PLMS is still controversial, the persistence or increase of the PLMS might be associated with the reduction of post-operative dopaminergic agents.

4. DBS TARGETS AND MECHANISM OF SLEEP-WAKE DISTURBANCES IN PD

An important question is the selection of a DBS target and the specific effects of applying DBS to each target on SWD in PD. Regarding this topic, a limited number of studies have evaluated the effect of DBS at various targets on sleep in advanced PD patients Table (3). Furthermore, direct comparisons of DBS targets on overall sleep effects are very limited. There have been only two studies from the same group that evaluated the impact of DBS on sleep, in which patients were randomized to the DBS target (STN or GPi) [46, 85]. In a randomized clinical trial comparing the results of STN- vs. GPi-DBS (NSTAPS), no significant differences were seen in subjective sleep quality assessed by the PDSS score between the two groups at 12 [46] and 36 months [85] after DBS.

Table 3.

The impact of DBS at various targets on sleep-wake disturbances.

DBS Target	STN	PPN	GPi	Vim
Effects	Improvement in sleep quality, sleep architecture (SE, TST, WASO, SL)No change in EDS, PLMS, RBDConflicting results in RLS	Improvement in sleep architecture (SE, WASO, REM sleep duration), EDS	Improvement in sleep quality, sleep architecture (SE, SL), EDS, RLS	No change in sleep quality, sleep architecture
References	[26-28, 31, 33, 41, 42, 44, 47, 49, 50, 54, 59, 74]	[45, 55-57, 94]	[58, 104, 105]	[59]

Abbreviations: DBS = Deep brain stimulation; STN = Subthalamic nucleus; GPi = Globus pallidus interna; PPN = Pedunculopontine nucleus; Vim = Ventral intermediate nucleus of thalamus; SE = Sleep efficiency; TST = Total sleep time; WASO = Wake after sleep onset; SL = Sleep latency; EDS = Excessive daytime sleepiness; PLMS = Periodic limb movement in sleep; REM = Rapid eye movement; RBD = REM sleep behavior disorder; RLS = Restless leg syndrome.

The DBS effect on SWD may depend on the target of the stimulation; however, the mechanism by which the stimulation at different targets regulate sleep and wakefulness has not been elucidated. STN has great potential as a target for DBS, and multiple aforementioned studies have shown that STN-DBS improves subjective and objective measures of sleep in patients with PD. The STN is a small, glutamatergic nucleus located at the diencephalon-mesencephalic junction, posterolateral to the hypothalamus, and medial to the SN and red nucleus [86]. The nucleus has reciprocal projections to the sleep-wake regulatory centers, including the basal ganglia, thalamus, and frontobasal cortex [87, 88], and to the PPN, which regulates REM sleep and muscular tone [81, 89, 90]. Local field potential recordings from the STN have shown significant differences in band-power across different stages of sleep, suggesting a role of STN-DBS in the sleep regulatory network and the ascending activating network during REM sleep [91, 92]. Lesions of the STN in normal monkeys decrease the neuronal inhibitory activity from basal ganglia to the thalamus and PPN [93]. In PD patients with STN-DBS, correcting overactive inhibitory influence on PPN via the stimulation of STN could influence the improvement of SWD.

Although, the effect of PPN-DBS has been evaluated in a limited number of patients, the results have shown that PPN-DBS improved sleep efficiency and decreased wake after sleep onset and nocturnal awakenings, similar to the effect of STN-DBS in PD [56, 57]. Moreover, PPN-DBS also increased REM sleep [56, 57] and improved EDS, unlike STN-DBS [45, 55, 94]. The ascending PPN cholinergic neurons project to the thalamus and contribute to the regulation of arousal state maintenance and REM sleep generation [81]. Given the loss of cholinergic neurons in PD [11], the PPN degeneration can be involved in SWD. A number of animal studies have supported the potential role of the cholinergic PPN neurons in the control of arousal and REM sleep [95-97]. In 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP)-treated monkeys, a reduction in REM and slow-wave sleep and an increased number of nocturnal awakening were observed [98, 99]. In addition, cholinergic PPN lesion improved sleep quality after the transient decrease in REM and slow-wave sleep in the in vivo MPTP-PD model [99]. However, the PPN is a heterogeneous nucleus that consists of GABAergic, glutamatergic neurons as well as cholinergic neurons that project to a diverse array of brain structures [100]. Given the complexity of the PPN region and the heterogeneity in the exact location of electrodes among studies, PPN-DBS is unlikely to yield consistent clinical outcomes on the improvement of SWD in PD patients.

There are limited data on the effect of GPi-DBS on SWD in PD. GPi-DBS revealed improving trends in several PSG measures, including sleep efficiency and sleep latency and RLS, but these improvements were not statistically significant [58]. Other GPi-DBS studies showed improvements in subjective sleep quality and EDS, however, sleep-related parameters were not the primary outcome measures in these studies [101, 102]. GPi is an important output nucleus of the basal ganglia, which also has projections to sleep-wake modulating centers, including the thalamus and PPN [103, 104]. Although, its specific role remains unknown, it is possible that GPi-DBS modulates the sleep-wake network and directly affects sleep physiology [103, 104]. In addition, both STN and GPi project to the GPe, and the improvement in sleep disturbances after STN- and GPi-DBS could be regulated by an indirect effect on the GPe, which is a key basal ganglia hub that directly interacts with the cortex [105].

The role of Vim-DBS on sleep was evaluated in only one study and Vim-DBS failed to modify sleep quality or architecture in this study [59].

5. COMPARISON OF THE EFFECTS OF DBS AND DOPAMINE REPLACEMENT THERAPY ON SLEEP-WAKE DISTURBANCES IN PD

Both levodopa and DBS lead to a major improvement in motor symptoms in PD. DBS is the surgical intervention of choice when motor complications are inadequately managed with medications in advanced PD patients [106]. DBS is more effective than the best medical therapy in improving the 'on' time without troublesome dyskinesia, motor fluctuation, and quality of life but is associated with an increased risk of adverse events [107-109]. Serious adverse events, including a fatal intracerebral hemorrhage, infections and general surgery problems, were more common with DBS than with medication alone [107-109]. Therefore, discussions with candidates for DBS should include the potential risks and benefits of surgery.

Studies comparing the impact of DBS and medical therapy on non-motor symptoms, especially on sleep, are limited. A previous review comparing the effects of DBS and dopamine replacement therapy (DRT) on non-motor symptoms of PD concluded that DBS is more effective than DRT in reducing sensory, sleep, gastrointestinal and urological symptoms [110]. In a recent controlled study, the score for the Non-Motor Symptom Scale (NMSS) sleep/fatigue domain significantly improved in the STN-DBS group, whereas it worsened in the DRT group at the 36-month follow-up after surgery [111]. In a similar vein, the mean total PDSS improved significantly at 3 months after STN-DBS, but no change was found for the DRT group [26]. Improvement of sleep quality by STN-DBS might primarily be due to a significant reduction of nighttime motor symptoms, whereas nocturia, sleep fragmentation, and EDS seemed to be unaffected [26]. In an Indian case-control study comparing frequencies of non-motor symptoms on DBS versus DRT, insomnia was significantly lower in the DBS group compared to the DRT group [112].

Furthermore, longitudinal cohort studies or randomized controlled studies are required to determine the effect of DBS on various SWD in patients with PD compared to that of DRT.

CONCLUSION

Recent evidence shows that DBS, particularly STN-DBS and possibly PPN-DBS, improve several aspects of SWD in patients with PD. In terms of the effects of DBS on SWD in PD patients, the STN and PPN are recognized as preferred sites due to their anatomical proximity and functional connection with the structures known to control SWD. DBS tends to improve both subjective and objective sleep parameters, but a discrepancy between the improvement of the subjective sleep quality and the deterioration of the objective sleep architecture has also been reported. EDS can either persist or newly develop after DBS. Given the paucity of the data, the impact of DBS on RBD is inconsistent and requires further investigation. There has been conflicting evidence regarding the effects of DBS on RLS. Because dopaminergic agents not only treat RLS symptoms but also induce RLS augmentation, the influence of a post-operative decrement in dopaminergic agents on RLS is difficult to evaluate. Further studies with a large number of participants and well-established sleep questionnaires and tests are needed. Especially, studies comparing the effects of the stimulation at different targets are required, and the impact of PPN-DBS on sleep needs to be further explored.

CONSENT FOR PUBLICATION

Not applicable.

FUNDING

None.

CONFLICT OF INTEREST

The authors have no conflicts of interest, financial or otherwise.