Magnetic stimulation was applied over the motor cortex in forty-five normal human subjects and peristimulus time histograms (PSTHs) of the discharges of single motor units were used to record changes in the firing probability of individual spinal motoneurones of contralateral upper limb muscles. Recordings were obtained from 153 motor units from fourteen upper limb muscles. 2. For the majority of motor units the initial effect was a short latency facilitation. The estimated central conduction velocities and the rise times of the underlying excitatory postsynaptic potentials (EPSPs) were compatible with monosynaptic facilitation by a fast corticospinal pathway. In some motor units the initial effect was a short latency inhibition. Other units showed no statistically significant changes in firing probability. The proportion of the tested motor units in each of these categories depended on the muscle. All of the sampled units of first dorsal interosseous (1DI) showed short latency facilitation, as did the majority of units in the forearm and the biceps brachii. More than half of the sampled motor units of triceps brachii and deltoid showed either no effect or were inhibited. 3. To compare the net short latency actions of the neurones activated by magnetic stimulation on various motoneurone pools, the magnitude of the short latency facilitation or inhibition in a given motor unit was normalized to the magnitude of the short latency facilitation in the 1DI motor unit of the same subject at the same stimulus intensity, and these data were pooled for a number of subjects. 4. 1DI motoneurones received strong net facilitation (estimated mean EPSP amplitude 2.9 +/- 0.2 mV), the motoneurones of forearm muscles and biceps brachii received weaker net facilitation and triceps brachii and deltoid received no net effect. 5. It is concluded that the short latency corticospinal projections to upper limb motoneurones in humans have a distinct pattern which is similar to that in other primates.
