Brain Research, 114 (1976) 505-510 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands 505 Corticomotoneuronal connections of precentral cells detected by post- spike averages of EMG activity in behaving monkeys E. E. FETZ, P. D. CHENEY and D. C. GERMAN Department of Physiology and Biophysics, and Regional Primate Research Center, University of Wash- ington, Seattle, Wash. 98195 (U.S.A.) (Accepted June 1lth, 1976) The observation that activity of certain precentral motor cortex cells covaries consistently with contralateral muscles under different behavioral conditions may suggest a functional relationship, but cannot establish the existence of an anatomic connectiona, 4. Even those precentral cells with monosynaptic connections to moto- neurons would be expected to generate unitary EPSPs too small to fire the motoneuron reliably e,8,9. However, the effects of such unitary EPSPs on the firing probability of an active motoneuron should, in principle, be statistically detectable by cross-correla- tion techniques. A convenient approximation to a true cross-correlation is the post-spike average, which is triggered from action potentials of the presynaptic neuron and pref- erentially summates those postsynaptic events correlated with that cell. By summing motoneuron membrane potentials following action potentials in Ia afferent fibers, Mendell and Henneman 7 first demonstrated that each Ia afferent fiber produces monosynaptic EPSPs in virtually all homonymous motoneurons. Similarly, post-spike averages have revealed monosynaptic connections to motoneurons from secondary afferents 6 and from inhibitory spinal interneurons 5, as well as intracortical connections between motor cortex cells 1. While these studies summed intracellular membrane potentials, recorded in anesthetized animals, we have investigated the possibility that spike-triggered averages could detect the effects of monosynaptic corticomotoneuronal connections on the firing probability of motor units in alert, behaving monkeys. Two rhesus monkeys (Macaca mulatta) were trained to flex and extend the wrist alternately against a programmed load; wrist position was tonically maintained for 1-2 sec to provide long periods of coactivation of precentral cells and agonist muscles. One monkey moved the wrist between two stops, the other between two electronically detected hold zones. EMG activity of forelimb muscles was recorded with either surface or implanted electrodes, or both. For 201 motor cortex cells whose activity covaried with wrist flexion or extension, post-spike averages were compiled with a Lab 8E computer. The full-wave rectified EMG activity was summed over a 30-msec analysis interval from 5 msec before to 25 msec after the cortical spike. Time bins were usually 250/~sec and at least 2000 events were averaged for each cell. A second type of average, the response average, was also compiled to document the covariation of cell and muscle