Pathological gambling in patients with Parkinson's disease is associated with fronto-striatal disconnection: A path modeling analysis

Abstract

Pathological gambling may occur in Parkinson's disease (PD) as a complication of dopaminergic therapy. Neuroimaging studies have suggested an abnormal dopamine transmission within the reward system, but the changes in the neural network characterizing PD patients with pathological gambling have never been investigated. Thirty PD patients (15 with active gambling and 15 matched controls, on-medication) and 15 healthy subjects underwent brain perfusion single photon emission tomography at rest. The severity of gambling was assessed using the South Oaks Gambling Scale. Covariance analysis was applied to identify brain regions whose activity was associated with gambling severity. These regions were used as volume-of-interest to identify functionally interconnected areas using voxel-wise covariance analysis. A path model was defined by means of effective connectivity analysis within the Structural Equation Modeling framework. Gambling severity in PD was associated with a dysfunction of the brain network implicated in decision making, risk processing, and response inhibition, including the ventrolateral prefrontal cortex, anterior (ACC) and posterior cingulate cortex, medial prefrontal cortex, insula and striatum. PD gamblers showed a disconnection between the ACC and the striatum, while this interaction was very robust in both control groups. ACC-striatal disconnection may underlie a specific impairment of shifting behaviors after negative outcomes, possibly explaining why PD gamblers use to perseverate into risktaking behaviors despite self-destructive consequences.