We used a paired-pulse magnetic stimulation technique to study ipsilateral cortico-cortical inhibition of the motor cortex in 48 patients with various neurological disorders and in 20 normal volunteers. In the normal subjects, the first subthreshold conditioning stimulus suppressed responses to the second suprathreshold test stimulus at interstimulus intervals (ISIs) of 1-5 ms (inhibition at short intervals), and facilitated them at ISIs of 8-15 ms (facilitation at long intervals). Patients with motor neuron disease, except those in whom brain stimulation produced control responses that were generated by direct activation of corticospinal neurons (D-waves), had normal inhibition at short intervals. Facilitation at long intervals was not elicited in some patients with amyotrophic lateral sclerosis. Less inhibition at short intervals and normal facilitation at long intervals was found for all the patients with progressive myoclonic epilepsy, a condition in which the excitability of cortical inhibitory interneurons is thought to be affected. Inhibition at short intervals was disturbed, but facilitation at long intervals was intact in the patients with movement disorders (Parkinson's disease, corticobasal degeneration, and Wilson's disease). In these patients, positron emission tomography (PET) studies showed decreased regional cerebral blood flow (rCBF) in the basal ganglia in the relaxed state. However, normal suppression was elicited in the patients with Parkinson's disease with normal rCBF. In four patients with chorea, the time-course of inhibition and facilitation was normal, even though PET studies showed decreased rCBF in the basal ganglia in two of them. Normal inhibition could not be elicited in patients who had a small lesion in the basal ganglia or in the pathway from basal ganglia to the primary motor cortex; the putamen, globus pallidus, and supplementary motor cortex. In contrast, patients who had a lesion in a sensory system (sensory cortex or sensory thalamus) or in the pontine nucleus had normal suppression. We conclude that the results of ipsilateral cortico-cortical inhibition with paired magnetic stimulation reflect the excitability of inhibitory interneurons in the motor cortex and that outputs from the basal ganglia markedly affect this inhibition, but outputs from somato-sensory systems or cerebellum do not. Moreover, dysfunction of the corticospinal tract or spinal motoneurons does not affect results obtained by the paired magnetic stimulation technique when the control responses are generated by I-waves (i.e. descending volleys are produced by transsynaptic activation of the corticospinal tract neurons.