Background: Motor evoked potentials (MEPs) are commonly used during surgery for spinal cord tumor resection. However, it can be difficult to record reliable MEPs from the muscles of the lower extremities during surgery in patients with preoperative weakness due to spinal cord compression. In this study, motor function of patients' lower extremities and their association with intraoperative MEP recording were compared. Patients and Methods: Patients undergoing thoracic spinal cord tumor resection were studied. Patients' motor function was checked immediately before the surgical procedure. MEP responses were recorded from the tibialis anterior and foot muscles, and the hand muscles were used as control. Electrical current with train of eight pulses, 200 to 500V was delivered through 2 corkscrews placed at C3' and C4' sites. Anesthesia was maintained by total intravenous anesthesia using a combination of propofol and remifentanil after induction with intravenous propofol, remifentanil, and rocuronium. Rocuronium was not repeated. Bispectral Index was maintained between 40 to 50. Results: From 178 lower limbs of 89 patients, myogenic MEPs could be recorded from 100% (105/105) of the patients with 5 of 5 motor strength in lower extremity; 90% (36/40) from the patients with 4/5 motor strength; only 25% (5/20) with 3/5; and 12.5% (1/8) with 2/5 motor strength; none (0/5) were able to be recorded if the motor strength was 1/5. Summary: The ability to record myogenic MEPs is closely associated with the patient's motor function. They are difficult to obtain if motor function is 3/5 motor strength in the lower extremity. They are almost impossible to record if motor function is worse than 3/5.