Brain metabolic and network profiles characterized by comorbid depressive symptoms and sleep disturbances in Parkinson’s disease: an 18F-FDG PET study

Abstract

Background Parkinson’s disease (PD) is a globally increasing neurodegenerative disorder. Its complex, multisystem pathophysiology extends beyond dopaminergic neuron loss and gives rise to diverse clinical manifestations. This exploratory study examined brain metabolic patterns in patients with PD and depressive symptoms (PD-Dep) or sleep disturbances (PD-SD). Methods Twenty-eight patients with PD assessed between December 2019 and January 2024 were included. Sixteen met the HAMD threshold for PD-Dep, 21 met the PSQI threshold for PD-SD, and nine met both criteria; the cohorts were therefore overlapping symptom-defined groups. Seventeen healthy participants served as controls (HC). All participants underwent 18F-FDG PET. Voxel-wise group contrasts and exploratory AAL-based partial correlations with HAMD, PSQI, and UPDRS-III were examined. Results At the prespecified liberal voxel-wise cluster-forming threshold, both symptom-defined PD groups showed extensive frontoparietal SUVR reductions relative to HC; these spatial findings should be considered exploratory. In PD-Dep, nominal HAMD associations included cortical and subcortical regions, whereas UPDRS-III associations were concentrated in frontoparietal regions. In PD-SD, nominal PSQI associations showed negative frontal/somatomotor and positive temporo-limbic patterns. The Yeo-network summaries were derived from symptom-correlated AAL regions, not from the voxel-wise group-contrast clusters. None of the regional correlations survived correction for multiple comparisons. Conclusion The findings identify exploratory metabolic and network-level patterns associated with depressive symptoms and sleep disturbances in PD. The symptom-correlated regional distributions are compatible with, but do not validate, aspects of the body-first/brain-first framework. Larger independent cohorts, conventional voxel-wise thresholding, multiplicity-controlled analyses, and explicit adjustment for motor severity are required before symptom-specific mechanisms or PD subtypes can be inferred.