Impaired Slow Wave Activity and Signal Complexity during NREM Sleep in Levodopa-Induced Dyskinesias

Sleep disturbances are common in Parkinson’s disease (PD) and often emerge early, suggesting that sleep may be a potential modulator of the clinical manifestations of the disease. While levodopa effectively treats motor symptoms, high-dose use can cause motor fluctuations and levodopa-induced dyskinesias (LID). This study investigates whether dyskinetic (DYS) patients differ from non-DYS PD patients and healthy controls in terms of slow-wave activity and brain signal complexity - key markers of sleep homeostasis and neural information processing. The analysis focused on the deepest stage N3 of non-REM sleep, which is considered crucial for the restoration of synaptic balance according to the synaptic homeostasis hypothesis. In order to examine the temporal dynamics of this restorative process, two distinct periods of the night were selected for comparison: the early sleep phase and the late sleep phase. Standard sleep metrics and N3 duration did not differ between groups, prompting a focus on EEG-derived features. Two analytical approaches were used: a subject-level analysis, which averaged signal features across channels for each participant but did not yield significant results due to limited sample size; and a channel-level analysis, which treated channels as observational units. Despite spatial dependencies, appropriate statistical corrections were applied, leading to significant findings. DYS group exhibited significantly reduced slow-wave activity during early sleep compared to other non-DYS PD groups and controls. Unlike the other groups, the DYS patients showed no significant slow-wave activity decline in any brain area from early to late sleep. The slow-wave activity difference between early and late sleep was significantly smaller in the DYS group than in all other groups. For brain signal complexity, both controls and non-DYS PD groups showed a widespread increase from early to late sleep, reflecting normal sleep-dependent neural reorganization. In contrast, DYS patients did not exhibit significant complexity increases in any brain region, their difference in complexity between early and late sleep was significantly the smallest among all groups and their late-night complexity remained lower than that of advanced PD patients. These findings suggest that patients with LID exhibit disrupted slow-wave activity, reflecting impaired synaptic downscaling and an unreduced sleep need, as well as a lack of sleep-related increases in brain signal complexity, indicating impaired cortical reorganization and potentially dysfunctional synaptic plasticity. This evidence highlights an abnormal deep sleep physiology in this subgroup of PD patients. Future studies should investigate brain activity of non-DYS and DYS PD groups after awakening to assess how these markers evolve at different times of the day. In addition, it would be worthwhile to investigate whether specific treatments can restore the physiological dynamics of these metrics and determine whether such interventions effectively reduce dyskinesias.