Poststimulus time histograms (PSTHs) of repetitively discharging motor units can be used to detect postsynaptic potentials (PSPs) in single human motoneurons. The present theoretical study explores the possibility of extending this technique to detect projections onto pools of interneurons upstream from motoneurons in humans. Convergence onto interneurons in the central nervous system of experimental animals can be tested by stimulating two afferent fiber systems separately and then together while recording intracellularly from an output neuron. Nonalgebraic summation of the postsynaptic potentials recorded in the output neuron implies convergence onto intervening interneurons. Can PSTHs of an output neuron be used for the same purpose? We modeled PSPs, their combination, and the PSTHs they would produce in a repetitively discharging neuron, in an attempt to predict the potential errors of such experiments. The model predicts that where the PSPs are exactly superimposed, peaks of increased firing probability in PSTHs summate algebraically. Under these circumstances, PSTHs could be used to test for convergence onto upstream interneurons. Where the PSPs are not exactly aligned, the PSTH peak from the combination of excitatory postsynaptic potentials (EPSPs) is smaller than expected. This could lead to a false impression that occlusion was occurring in a population of interneurons. This error is least with EPSPs having a short rise time. Where dissimilar EPSPs are superimposed, the error is less if the EPSP with the shortest rise time occurs later. EPSP-IPSP combinations can produce increases in firing probability that are greater than expected. This could lead to a false impression that facilitatory convergence was occurring in a population of interneurons. We conclude that, with an understanding of these potential errors, PSTHs can be used to detect nonalgebraic summation of PSPs and thus convergence onto interneurons in humans.