Fig. 1: A schematic overview of the study design pipeline.

First, an ALE meta-analysis was conducted on functional activation data derived from the comparison between PD patients with and without fatigue to integrate findings across published rs-fMRI studies. Second, the significant convergence across these enrolled rs-fMRI findings yielded 12 significant clusters predominantly in the sensorimotor, frontal, and temporal regions. Third, the discovery dataset (referred to as “D1_Old Subjects”) and the two validation datasets (referred to as “D2_1000 Subjects” and “D3_Young Subjects”) were adopted for further ANM procedure. Fourth, the brain network underlying fatigue in PD was localized by the ANM method. Fifth, the identified fatigue-related network in PD was interpreted through transcriptomic patterns (i.e., gene functional features, including the functional enrichment and cell type-specific expression), chemoarchitectures (i.e., neurotransmitter distribution patterns), and behavioral relevance. ALE activation likelihood estimation, ANM activation network mapping, FC functional connectivity, PD Parkinson’s disease, rs-fMRI resting-state functional magnetic resonance imaging.