Further evidences for sleep instability and impaired spindle-delta dynamics in schizophrenia: a whole-night polysomnography study with neuroloop-gain and sleep-cycle analysis

Highlights

• Replicated gross sleep-architecture abnormalities and sleep spindle deficits.

• Highlighted sleep-cycle wise abnormalities in schizophrenia.

• Showed higher vulnerability of third REM-period to sleep disruptions.

• Compensatory change in spindle-delta dynamics in second NREM-period.

• Supports role of inefficient thalamocortical network in schizophrenia.

Abstract

Objective

Sleep offers a unique window into the brain dysfunctions in schizophrenia. Many past sleep studies have reported abnormalities in both macro-sleep architecture (like increased awakenings) as well as micro-sleep-architecture (like spindle deficits) in patients with schizophrenia (PSZ). The present study attempts to replicate previous reports of macro- and micro-sleep-architectural abnormalities in schizophrenia. In addition, the study also examined sleep-stage changes and spindle-delta dynamics across sleep-cycles to provide further evidence in support of the dysfunctional thalamocortical mechanisms causing sleep instability and poor sleep maintenance associated with schizophrenia pathophysiology.

Methods

Whole-night polysomnography was carried out among 45 PSZ and 39 age- and gender-matched healthy control subjects. Sleep-stage dynamics were assessed across sleep-cycles using a customized software algorithm. Spindle-delta dynamics across sleep-cycles were determined using neuroloop-gain analysis.

Results

PSZ showed macro-sleep architecture abnormalities such as prolonged sleeplessness, increased intermittent-awakenings, long sleep-onset latency, reduced non-rapid eye movement (NREM) stage 2 sleep, increased stage transitions, and poor sleep efficiency. They also showed reduced spindle density (sigma neuroloop-gain) but comparable slow wave density (delta neuroloop-gain) throughout the sleep. Sleep-cycle-wise analysis revealed transient features of sleep instability due to significantly increased intermittent awakenings especially in the first and third sleep-cycles, and unstable and recurrent stage transitions in both NREM (first sleep-cycle) and rapid eye movement (REM) sleep-periods (second sleep-cycle). Spindle deficits were persistent across the first three cycles and were positively correlated with sleep disruption during the subsequent REM sleep.

Conclusions

In addition to replicating previously reported sleep deficits in PSZ, the current study showed subtle deficits in NREM–REM alterations across whole-night polysomnography. These results point towards a possible maladaptive interplay between unstable thalamocortical networks, resulting in sleep-cycle-specific instability patterns associated with schizophrenia pathophysiology.

Introduction

Sleep represents a unique behavioural state characterized by functional disconnections from external stimuli modulated through complex thalamocortical oscillatory events [1]. Thalamocortical dysfunctions are implicated in the pathophysiology of schizophrenia, a debilitating mental illness that afflicts around 0.5–1% of the world's population [2], and sleep abnormalities are consistently being reported in this disorder. A proper evaluation of sleep abnormalities hence would help to elucidate the abnormal thalamocortical oscillatory events in addition to providing an understanding of schizophrenia pathophysiology itself. Sleep abnormalities such as increased sleep latency, reduced sleep efficiency and increased intermittent awakenings [3], [4], [5], [6], indicate sleep initiation and maintenance abnormalities. This is in addition to the non-rapid eye movement (NREM) [7], [8], [9], [10] as well as rapid eye movement (REM) sleep abnormalities [5], [11], such as reduced slow wave sleep duration (N3), increased light stages (N1 and N2), increased intermittent awakenings, decreased delta waves (0.5- to 4-Hz oscillations predominant during N3 stages) and decreased spindles (11- to 16-Hz waxing–waning oscillations during N2 and N3 stages), reduced REM sleep onset latency, enhanced REM fragmentation with reduced REM duration and REM density (frequency of occurrence of rapid eye movements during REM sleep). Long-term sleep–wake studies using actigraphy [12] have reported altered sleep and wake timings (circadian dysfunction) along with abnormal serum melatonin levels (a biomarker for circadian rhythm) in almost 50% of patients with schizophrenia (PSZ). Nevertheless, the sleep abnormalities are neither consistent nor unique to schizophrenia [13] as these are also reported in other mental disorders such as in major depression [14].

Most of the previous sleep studies in schizophrenia have tried to look at the macro- and micro-sleep-architecture abnormalities averaged across the entire sleeping period. Such approaches do not provide much information on the dynamic changes associated with NREM–REM sleep alternations. By considering the complexity of electroencephalogram (EEG) synchronization associated with sleep deepening and the dynamicity involved in NREM–REM sleep alternations [15], understanding micro-sleep-dynamics on a smaller time-scale would provide additional information on the dysfunctional thalamocortical mechanisms associated with schizophrenia [1], [16]. Such approaches provide us with details of the NREM–REM sleep cyclicity [17] and NREM and REM interdependencies [10], [18], which are important determinants of sleep continuity, sleep satiety, quality, etc. From the two-process model of sleep regulation [19] it could be speculated that sleep initiation and maintenance abnormalities in schizophrenia could be due to ineffective implementation of homeostatic process (‘process S’), whereas circadian rhythm abnormalities associated with schizophrenia could impact the circadian process (‘process C’) via REM sleep disturbances. Additionally, dysfunctional aminergic–cholinergic interactions associated with the disorder [20] would interfere with NREM–REM alternations [21]. Therefore, examining cycle-wise dynamics of sleep patterns would help us to elucidate the transient features of sleep abnormalities in schizophrenia that are characteristic of its pathophysiology, but have rarely been given importance.

Similarly, a proper understanding of micro-sleep-dynamics, especially the spindle characteristics and spindle-delta dynamics, would provide ample evidence of sleep deepening mechanisms essential for restorative sleep, and hence there is a need to elucidate how such events are affected in schizophrenia. Spindle oscillations are generated by thalamic reticular neurons, but their synchronization, duration and propagation are determined by thalamo-thalamic and cortico-thalamic circuits [22], [23]. Within the NREM sleep-period, spindle and delta oscillations exhibit a dynamic relation, with a positive correlation in the initial portion, followed by a negative correlation when delta rises beyond a threshold, and finally a coherent reduction of both before the end of that period [24], [25]. Spindle-delta dynamics are also differentially affected by circadian and other physiological measures [26]. Reduction in various parameters of sleep spindles, such as peak spindle frequency, spindle density, duration and amplitude, have been reported among patients with autism, mental retardation [27] and schizophrenia [28], [29], [30]. Any imbalance in these dynamics could affect the duration and integrity of subsequent sleep-stages, particularly the REM sleep [31]. Therefore, assessing cycle-wise dynamics of sleep micro-architecture patterns would provide us with ample evidences of how these are altered in schizophrenia.

Hence, in the present study, we investigated sleep-stage dynamics as well as delta and spindle density measures (micro-sleep-architecture) across sleep-cycles among PSZ and in age-matched healthy control subjects (HCS). We aimed to replicate some of the sleep deficits reported among PSZ by the earlier mentioned pioneering studies, as well as to carry out sleep-cycle analysis that has not already been done. As sleep electroencephalography (EEG) oscillatory patterns are believed to be an emergent property of the thalamocortical network, it was thought that neuroloop-gain analysis would reveal significant alterations in the cycle-wise dynamics of spindle and delta activity, by providing a continuous measure of these sleep events. Furthermore, such measures would permit replication of the spindle-deficits reported by previous studies.