Transcutaneous spinal cord stimulation (TSS) is purported to improve motor function in people after spinal cord injury (SCI). However, several methodology aspects are yet to be explored. We investigated whether stimulation configuration affected the intensity needed to elicit spinally evoked motor responses (sEMR) in four lower limb muscles bilaterally. Also, since stimulation intensity for therapeutic TSS (i.e., trains of stimulation, typically delivered at 15-50 Hz) is sometimes based on the single-pulse threshold intensity, we compared these two stimulation types. In non-SCI participants (n = 9) and participants with a SCI (n = 9), three different electrode configurations (cathode-anode); L1-midline (below the umbilicus), T11-midline and L1-ASIS (anterior superior iliac spine; non-SCI only) were compared for the sEMR threshold intensity using single pulses or trains of stimulation which were recorded in the vastus medialis, medial hamstring, tibialis anterior, medial gastrocnemius muscles. In non-SCI participants, the L1-midline configuration showed lower sEMR thresholds compared to T11-midline (p = 0.002) and L1-ASIS (p < 0.001). There was no difference between T11-midline and L1-midline for participants with SCI (p = 0.245). Spinally evoked motor response thresholds were ~13% lower during trains of stimulation compared to single pulses in non-SCI participants (p < 0.001), but not in participants with SCI (p = 0.101). With trains of stimulation, threshold intensities were slightly lower and the incidence of sEMR was considerably lower. Overall, stimulation threshold intensities were generally lower with the L1-midline electrode configuration and is therefore preferred. While single-pulse threshold intensities may overestimate threshold intensities for therapeutic TSS, tolerance to trains of stimulation will be the limiting factor in most cases.