Cognitive impairment (CI) and dementia are frequent and debilitating features associated with Parkinson's disease (PD). Formal neuropsychological examination is required to ascertain the degree and pattern of CI over the course of the disease. The use of different tools may explain heterogeneous data obtained from studies to date. Normative data for extensively used scales [Mattis Dementia Rating Scale (MDRS), Mini-Mental State Examination (MMSE)] is incomplete in PD populations. According to sample characteristics, statistical analyses, and methodological quality, 33 studies using scales not specific to PD (MDRS, MMSE, Cambridge Cognitive Assessment, FAB) or PD-specific scales (Mini-Mental Parkinson, Scales for Outcomes of Parkinson's disease-Cognition, Parkinson's Disease-Cognitive Rating Scale, and Parkinson Neuropsychometric Dementia Assessment) were eligible for the critical analysis of their appropriateness to assess cognition in PD. Of the four scales specifically designed for PD, the SCOPA-COG and the PD-CRS have undergone extensive and rigorous validation processes. While the SCOPA-COG mainly assesses "frontal-subcortical" cognitive defects, the PD-CRS also assesses "instrumental-cortical" functions, allowing better characterization of the different patterns of CI that may be present in PD from the earliest stages. The MMP and PANDA scales were designed as brief screening tests for CI and have not yet been subjected to extensive clinimetric evaluations. Further research on PD-specific tools seems mandatory to help establish accurate cut-off scores for the diagnosis of mild PDD, detect cognitive profiles more prone to the future development of dementia, and allow comparisons between different descriptive or interventional studies.