In Parkinson's disease, striatal dopamine depletion produces profound alterations in the neural activity of the cortico-basal ganglia motor loop, leading to dysfunctional motor output and parkinsonism. A key regulator of motor output is the balance between excitation and inhibition in the primary motor cortex, which can be assessed in humans with transcranial magnetic stimulation techniques. Despite decades of research, the functional state of cortical inhibition in Parkinson's disease remains uncertain. Towards resolving this issue, we applied paired-pulse transcranial magnetic stimulation protocols in 166 patients with Parkinson's disease (57 levodopa-naïve, 50 non-dyskinetic, 59 dyskinetic) and 40 healthy controls (age-matched with the levodopa-naïve group). All patients were studied OFF medication. All analyses were performed with fully automatic procedures to avoid confirmation bias, and we systematically considered and excluded several potential confounding factors such as age, gender, resting motor threshold, EMG background activity and amplitude of the motor evoked potential elicited by the single-pulse test stimuli. Our results show that short-interval intracortical inhibition is decreased in Parkinson's disease compared to controls. This reduction of intracortical inhibition was obtained with relatively low-intensity conditioning stimuli (80% of the resting motor threshold) and was not associated with any significant increase in short-interval intracortical facilitation or intracortical facilitation with the same low-intensity conditioning stimuli, supporting the involvement of cortical inhibitory circuits. Short-interval intracortical inhibition was similarly reduced in levodopa-naïve, non-dyskinetic and dyskinetic patients. Importantly, intracortical inhibition was reduced compared to control subjects also on the less affected side (n = 145), even in de novo drug-naïve patients in whom the less affected side was minimally symptomatic (lateralized Unified Parkinson's Disease Rating Scale part III = 0 or 1, n = 23). These results suggest that cortical disinhibition is a very early, possibly prodromal feature of Parkinson's disease.