Subanaesthetic doses of the N-methyl-D-aspartate (NMDA) antagonist ketamine have been shown to determine a dual modulating effect on glutamatergic transmission in experimental animals, blocking NMDA receptor activity and enhancing non-NMDA transmission through an increase in the release of endogenous glutamate. Little is known about the effects of ketamine on the excitability of the human central nervous system. The effects of subanaesthetic, graded incremental doses of ketamine (0.01, 0.02 and 0.04 mg kg-1 min-1, I.V.) on the excitability of cortical networks of the human motor cortex were examined with a range of transcranial magnetic and electric stimulation protocols in seven normal subjects. Administration of ketamine at increasing doses produced a progressive reduction in the mean resting motor threshold (RMT) (F(3, 18) = 22.33, P < 0.001) and active motor threshold (AMT) (F(3, 18) = 12.17, P < 0.001). Before ketamine administration, mean RMT +/- S.D. was 49 +/- 3.3 % of maximum stimulator output and at the highest infusion level it was 42.6 +/- 2.6 % (P < 0.001). Before ketamine administration, AMT +/- S.D. was 38 +/- 3.3 % of maximum stimulator output and at the highest infusion level it was 33 +/- 4.4 % (P < 0.002). Ketamine also led to an increase in the amplitude of EMG responses evoked by magnetic stimulation at rest; this increase was a function of ketamine dosage (F(3, 18) = 5.29, P = 0.009). In contrast to responses evoked by magnetic stimulation, responses evoked by electric stimulation were not modified by ketamine. The differential effect of ketamine on responses evoked by magnetic and electric stimulation demonstrates that subanaesthetic doses of ketamine enhance the recruitment of excitatory cortical networks in motor cortex. Transcranial magnetic stimulation produces a high-frequency repetitive discharge of pyramidal neurones and for this reason probably depends mostly on short-lasting AMPA transmission. An increase in this transmission might facilitate the repetitive discharge of pyramidal cells after transcranial magnetic stimulation which, in turn, results in larger motor responses and lower thresholds. We suggest that the enhancement of human motor cortex excitability to transcranial magnetic stimulation is the effect of an increase in glutamatergic transmission at non-NMDA receptors similar to that described in experimental studies.