Sleep spindle density and temporal clustering are associated with sleep-dependent memory consolidation in Parkinson's disease

Abstract

Study objectives: Sleep is required for successful memory consolidation. Sleep spindles, bursts of oscillatory activity occurring during non-rapid eye movement sleep, are known to be crucial for this process and, recently, it has been proposed that the temporal organization of spindles into clusters might additionally play a role in memory consolidation. In Parkinson's disease, spindle activity is reduced, and this reduction has been found to be predictive of cognitive decline. However, it remains unknown whether alterations in sleep spindles in Parkinson's disease are predictive of sleep-dependent cognitive processes such as memory consolidation, leaving open questions about the possible mechanisms linking sleep and a more general cognitive state in Parkinson's patients.

Methods: The current study sought to fill this gap by recording overnight polysomnography and measuring overnight declarative memory consolidation in a sample of 35 patients with Parkinson's. Memory consolidation was measured using a verbal paired-associates task administered before and after the night of recorded sleep.

Results: We found that lower sleep spindle density at frontal leads during non-rapid eye movement stage 3 was associated with worse overnight declarative memory consolidation. We also found that patients who showed less temporal clustering of spindles exhibited worse declarative memory consolidation.

Conclusions: These results suggest alterations to sleep spindles, which are known to be a consequence of Parkinson's disease, might represent a mechanism by which poor sleep leads to worse cognitive function in Parkinson's patients.

Citation: Lahlou S, Kaminska M, Doyon J, Carrier J, Sharp M. Sleep spindle density and temporal clustering are associated with sleep-dependent memory consolidation in Parkinson's disease. J Clin Sleep Med. 2024;20(7):1153-1162.

Keywords: Parkinson’s disease; cognition; memory consolidation; sleep spindles.

Figure 1. Schematic of the experimental design.

Participants learned 50 semantically related word pairs in the evening. During the initial learning block, all 50 word pairs were presented one at a time and participants were instructed to remember them. This was followed by a practice block where one word was presented and participants had to verbally recall its associate. After a tone indicated the end of the recall period, the correct word was presented on the screen. The practice block was immediately followed by the presleep recall test where memory for one half of the pairs (randomly selected) was tested. Participants completed simple arithmetic problems for 1 minute between each block. The postsleep recall test took place the following morning and memory for the remaining 25 pairs was tested. Overnight polysomnography included 6 EEG leads (F3, F4, C3, C4, O1, O2) referenced to the contralateral mastoid, bilateral electrooculogram, and submental electromyography.

Figure 2. Relationship between sleep spindle density and memory consolidation.

Spindle density is shown for frontal (A) and central (B) leads during both NREM-2 and NREM-3 sleep in association with the overnight relative change in memory, where negative values represent worse memory performance in the morning than at night. Higher density (number of spindles per minute of sleep) of frontal sleep spindles in NREM-3 was associated with less overnight forgetting (r = .46, P = .006). The red line denotes the regression line and shaded areas represent the 95% confidence interval.

Figure 3. Relationship between memory consolidation and clustering of spindles in NREM-3 at frontal leads.

(A) Relationship between density of spindles occurring inside trains (count per minute) and change in memory overnight, where negative values indicate worse performance in the morning than at night. Increased density of spindles occurring inside trains was associated with less reduction in memory overnight (r = .52, P = .001). (B) Density of spindles occurring outside of trains was not associated with the overnight change in memory. (C) Relationship between proportion of spindles occurring inside trains (relative to the total number of spindles) and change in memory overnight. A higher proportion of spindles occurring inside trains was associated with less reduction in memory overnight (r = .46, P = .005). (D) Total train count was not associated with change in memory overnight. Red lines denote regression lines and shaded areas denote the 95% confidence intervals.