The aim of this study was to assess whether mild cognitive impairment (MCI) is associatedwith disruption in large-scale structural networks in newly diagnosed, drug-na€ıve patients with Parkin-son’s disease (PD). Graph theoretical analyses were applied to 3T MRI data from 123 PD patients and 56controls from the Parkinson’s progression markers initiative (PPMI). Thirty-three patients were classifiedas having Parkinson’s disease with mild cognitive impairment (PD-MCI) using the Movement DisordersSociety Task Force criteria, while the remaining 90 PD patients were classified as cognitively normal (PD-CN). Global measures (clustering coefficient, characteristic path length, global efficiency, small-world-ness) and regional measures (regional clustering coefficient, regional efficiency, hubs) were assessed inthe structural networks that were constructed based on cortical thickness and subcortical volume data.PD-MCI patients showed a marked reduction in the average correlation strength between cortical andsubcortical regions compared with controls. These patients had a larger characteristic path length andreduced global efficiency in addition to a lower regional efficiency in frontal and parietal regions com-pared with PD-CN patients and controls. A reorganization of the highly connected regions in the networkwas observed in both groups of patients. This study shows that the earliest stages of cognitive decline inPD are associated with a disruption in the large-scale coordination of the brain network and with adecrease of the efficiency of parallel information processing.