ObjectiveFreezing of gait (FoG) is a debilitating problem in patients with PD. The multifactorial pathogenesis of FoG remains poorly understood. We aimed to find which factors are most strongly associated with the occurrence of FoG.MethodsThree hundred five PD patients were enrolled and subdivided according to the presence (FoG +, n = 128) or absence (FoG-, n = 177) of FoG. Several clinical, functional, and neuropsychological data were collected and compared between groups. The association between the probability of presence of FoG and possible explanatory variables was assessed by logistic regression analysis.ResultsFoG + patients were younger at the diagnosis (p = 0.04), and their mean daily dose of dopaminergic drugs (p < 0.0001) was higher in comparison with FoG- patients. FoG + patients get worse in Frontal Assessment Battery (p = 0.005), had higher scores in Apathy Evaluation Scale (p = 0.03), and were much more impaired on Wisconsin Card Sorting Test (WCST) (p = 0.018), Trail Making Test A (p = 0.0013), and Ray Auditory Verbal Learning Test (p = 0.012). Levodopa equivalent dose, age (direct), age at disease onset (inverse), and WCST were significant predictors of FoG (p = 0.01, p = 0.0025, p = 0.0016, and p = 0.029, respectively).ConclusionFoG + patients show more deficits in executive functions and in motivation. The main explanatory variables of FoG occurrence are levodopa equivalent dose, age, age at disease onset, and WCST. These data suggest that a specific involvement of frontal cortical circuits in PD is responsible for certain cognitive–behavioral alterations related to the occurrence of FoG.