Excessive daytime sleepiness (EDS) is a common non-motor manifestation of Parkinson’s disease (PD) impacting nearly 50% of patients (Abbott and White, 2005). Characterized by inappropriate drowsiness during wakefulness, EDS impairs cognitive function and quality of life and increases risks such as traffic accidents (Knie et al., 2011). EDS is more frequently observed in advanced stages of PD and has been associated with various non-motor features, such as mood disturbances, autonomic dysfunction, and fatigue (Feng et al., 2021; Maggi et al., 2023). Though its exact mechanisms remain unclear, damage to wake-promoting brain regions and neurotransmitter imbalances—particularly in dopaminergic, cholinergic, and noradrenergic systems—are likely contributors (Liu et al., 2022). Identifying reliable biomarkers and understanding its neurobiological underpinnings is crucial for early intervention and disease management (Siddiqui et al., 2024; Tripathi et al., 2024).

Neuroimaging research has shed light on structural and functional brain changes in PD-EDS patients. Using Voxel-based morphometry (VBM), researchers have detected gray matter disruptions in regions involved in sleep–wake regulation (Kato et al., 2012; Chondrogiorgi et al., 2015; de Schipper et al., 2017). Surface-based morphometry (SBM), which is more sensitive than VBM in detecting subtle structural changes, has shown that cortical folding measurements, especially cortical thickness, better identify PD-related gray matter alterations (Pereira et al., 2012). SBM-based EDS findings include hypertrophy in the putamen and pallidum (Gong et al., 2019), cortical surface expansion in the anterior insula, and subcortical atrophy in the amygdala and putamen (Rosinvil et al., 2024). Research on cortical thickness in PD-EDS is limited, and conflicting findings underscore the need for further investigation of its impact on brain structure. Functional MRI (fMRI) studies have shown abnormal connectivity in cortical and subcortical arousal networks (Wen et al., 2016; Zi et al., 2022; Zheng et al., 2023), especially within the default mode network, where hyperactivity in prefrontal and temporal regions may reflect compensatory or attentional deficits (Ooi et al., 2019; Wang et al., 2020; Zheng et al., 2023).

Neurofilament light chain (NfL), a cytoskeletal protein released during axonal injury, is a well-established biomarker of neuronal damage and degeneration (Gaetani et al., 2019; Sharma et al., 2024). Raised blood NfL concentrations correlate with both motor and non-motor symptoms in PD, offering clinical potential through advancements in ultrasensitive detection techniques like single molecule arrays (Simoa™; Pilotto et al., 2021; Zhu et al., 2021; Yin et al., 2022). Although one study has revealed elevated plasma NfL levels in patients with PD and EDS (Lin et al., 2024), the relationship between NfL levels and specific neuroimaging markers (Preische et al., 2019; Sampedro et al., 2020; Cruz-Gomez et al., 2021; Lee et al., 2022; Clarelli et al., 2024; Yao et al., 2024) has only been explored in patients with cognitive dysfunction (Mielke et al., 2019). Until now, no prior studies have assessed the tripartite relationship among plasma NfL levels, neuroimaging markers, and EDS severity in PD, representing a pivotal knowledge gap.

This study is the first to integrate structural and functional neuroimaging to investigate the neural mechanisms underlying EDS in PD. Using FreeSurfer, we analyzed cortical thickness alterations in key brain regions and examined functional connectivity (FC) patterns across brain networks. Furthermore, we assessed the interrelationships among cortical thickness, FC, and plasma NfL levels, and their associations with EDS severity. Finally, mediation analyses were conducted to determine whether NfL mediates the relationships between cortical thinning, FC alterations, and EDS, providing insights into potential neurodegenerative pathways.

This study is, to our knowledge, the first to assess how cortical thickness, FC, and plasma NfL levels interact in PD-EDS. Relative to PD patients without EDS, individuals with PD-EDS showed elevated NfL levels, thinner cortices in the left SMG and right PoCR, and weakened FC between the left SMG and left PoCR, as well as between the right PoCR and left IFG operc. Further mediation analysis indicated that plasma NfL was a stronger mediator of the connection between structural brain changes and the severity of EDS in PD. Moreover, the association between cortical atrophy in the left SMG and elevated NfL concentrations may help differentiate PD-EDS from PD-non-EDS, providing a novel perspective on the underlying mechanisms of PD-EDS.

Our study provided the first structural MRI evidence of alterations in the left SMG and right PoCR in patients with PD-EDS. Both regions, located in the parietal lobe, are integral to sensory integration and cognitive processing (Vandenberghe et al., 2012). The PoCR, as part of the primary somatosensory cortex, is essential for tactile perception, spatial awareness, and sensorimotor coordination, whereas the SMG, within the inferior parietal lobule, integrates multisensory information to support higher cognitive functions, including attention and social cognition. Given that PD-EDS is closely tied to cognitive impairment, it is plausible that the atrophy of the parietal cortex, a key hub for attention and sensory processing, contributes to its pathogenesis. Similarly, significant cortical thinning in the medial and dorsolateral prefrontal cortices and inferior parietal lobules has been reported in narcolepsy patients with cataplexy, affecting executive attention and working memory (Joo et al., 2011). Additionally, obstructive sleep apnea is associated with a higher likelihood of developing PD-EDS (Jeon and Oh, 2023), with cortical thinning observed in the left parietal, frontal, and temporal lobes, which is negatively correlated with ESS score (Li et al., 2023). Thus, these findings suggest that cortical thinning in specific regions may be associated with attention deficit and memory impairment in sleep disorders, including PD-EDS.

Cortical atrophy in key parietal regions may disrupt network connectivity and contribute to EDS in PD. Specifically, reduced FC between the left SMG and PoCG may impair sensory integration and attentional regulation. As components of the somatosensory network, the PoCG (S1) and SMG (S2) are involved in processing tactile input and coordinating sensorimotor functions (Potok et al., 2019). Their weakened interaction may reduce arousal-related signaling, thereby increasing vulnerability to EDS (Hitchcott et al., 2019). Furthermore, weakened FC was detected between the right PoCR and left IFG operc. As a subregion of the left inferior frontal gyrus, the IFG operc plays a key role in information processing during audiovisual perceptual decision-making (Li et al., 2020). These findings align with evidence linking frontal cortex dysfunction to PD-EDS (Wen et al., 2016; Ooi et al., 2019; Zi et al., 2022). The impaired white matter tract integrity between the frontal and parietal lobes in patients with PD-EDS further supports the neural downregulation mechanisms underlying EDS (Chondrogiorgi et al., 2015). Reduced FC between these key brain regions may contribute to deficits in sensory processing, attention, and regulation of arousal. Moreover, the weaker connectivity between these brain regions was correlated with higher ESS scores, indicating greater EDS severity. Disruptions in the FC within the left parietal cortex may serve as targets for neuromodulation interventions such as repetitive transcranial magnetic stimulation. Further investigation is necessary to determine how effectively these approaches modulate the arousal and attentional networks in PD-EDS.

NfL is a key cytoskeletal protein in neurons and a well-established biomarker of axonal damage (Petzold, 2005). Our findings indicate that plasma NfL levels are elevated in PD-EDS compared to PD-non-EDS, aligning with previous studies (Lin et al., 2024). Mediation analysis suggested that reduced SMG thickness may exacerbate EDS severity by driving NfL elevation, implying that neurodegeneration-induced cortical thinning promotes axonal injury biomarker release, which subsequently worsens EDS. Previous studies have demonstrated that serum NfL is related to posterior cortical atrophy in early PD, particularly in the parietotemporo-occipital regions, and is a marker of non-dopaminergic neurodegeneration linked to cognitive decline and PD progression (Sampedro et al., 2020). Elevated NfL concentrations have been shown to mediate the relationship between cortical atrophy and cognitive decline in multiple sclerosis patients (Cruz-Gomez et al., 2021). In Alzheimer’s disease, a bidirectional relationship exists, in which higher NfL predicts faster cortical thinning, whereas reduced cortical thickness accelerates NfL elevation, reflecting ongoing axonal degeneration (Mattsson et al., 2019). Combining NfL level with MRI-based cortical thickness measurement enhances the assessment of neuroaxonal injury. However, the exact relationship between EDS-related cortical thinning and increased NfL levels remains unclear. One plausible mechanism is that low perfusion and reduced blood flow may contribute to cortical thinning by causing metabolic dysfunction and oxidative stress, leading to neuronal shrinkage and cortical atrophy. This thinning, in turn, could exacerbate neurodegeneration and worsen EDS severity (Morrison and Hof, 1997). Consistent with our findings, attention deficits and reduced blood flow in the left parietal cortex have been associated with PD-EDS (Matsui et al., 2006).

Our study explored the contributors to EDS in individuals with PD. Consistent with previous research, EDS was strongly associated with non-motor symptoms, particularly cognitive impairments, anxiety, depression, and poor nighttime sleep quality (Kurtis et al., 2013; Zhu et al., 2016; Feng et al., 2021; Maggi et al., 2023). However, in contrast to previous studies that linked EDS to age, disease duration, and PIGD phenotype, there were no significant associations with these variables in our cohort. Notably, mediation analysis revealed that the indirect effect of cortical thinning on EDS through NfL was reduced and became non-significant after controlling for MMSE scores, suggesting that cognitive dysfunction may partially account for this pathway. These findings highlight the predominant influence of non-motor symptoms—especially cognitive impairment—on EDS, relative to disease duration and motor phenotype. Further research with detailed neuropsychological assessments and rigorous control of confounders is required to clarify these relationships.

Despite these important findings, this study had several limitations. First, although ESS is a commonly employed scale for evaluating EDS, it remains a subjective measure. Polysomnography and other objective tests enable a more comprehensive evaluation. However, these tests are resource-intensive and impractical for large-scale studies, making ESS a feasible alternative for screening EDS in clinical and research settings. Second, the EDS group had a relatively small sample size, especially after we excluded participants with excessive head motion, which may have affected statistical power. Validating these findings necessitates larger, multicenter studies. Finally, the cross-sectional design limits causal inferences between NfL, cortical thinning, and EDS severity in patients with PD. To gain a better understanding of PD-EDS mechanisms, longitudinal studies are essential in determining whether elevated NfL levels precede cortical thinning or result from neurodegeneration.

Collectively, our results indicate that elevated plasma NfL partially mediated the interaction between cortical thinning in the left SMG thickness and the severity of EDS, possibly reflecting the crucial role of neurodegeneration in linking cortical atrophy in this region to EDS. Cortical thinning may impair FC and exacerbate EDS in patients with PD. By integrating cortical thickness, FC, and NfL, our study enhances the comprehension of the potential mechanisms of PD-EDS as well as provides insights into the clinical implications and possible therapeutic targets.