The presence of either excessive tonic chin EMG activity during REM sleep, or excessive phasic submental or limb EMG twitching is required to diagnose REM sleep behavior disorder (RBD). The aim was to identify cut-off values and to assess the sensitivity and specificity of these values taken separately or combined to diagnose idiopathic RBD patients. Eighty patients presenting with a clinical diagnosis of idiopathic RBD and 80 age- and gender-matched normal controls were studied in the sleep laboratory. Receiver operating characteristic curves were drawn to find optimal cut-off values for three REM sleep EMG parameters. Tonic and phasic EMG activity were measured in the chin, but not in the limbs. Videos were examined during the recording but were not systematically reviewed by the authors. Total correct classification of 81.9% was found for tonic chin EMG density ≥30%; 83.8% for phasic chin EMG density ≥15% and 75.6% for ≥24 leg movements per hour of REM sleep. Five patients did not fulfill any of these three polysomnographic (PSG) criteria. Conversely, one subject of the control group met the PSG criteria for RBD. This study estimates the diagnostic value of a visual scoring method for the diagnosis of idiopathic RBD and establishes cut-off values to be used in clinical and research set-ups. For the five RBD patients who did not show chin EMG abnormalities, it cannot be excluded that they had increased phasic EMG activity in the upper limbs and presented visible motor activity.