STUDY DESIGN: A comparison of electrical thresholds for biomechanical response in the ankle and for evoked electromyographic signals from specific leg muscles during intraoperative extradural direct stimulation of roots L4, L5, and S1. OBJECTIVE: To determine whether a biomechanical response in the ankle to direct root stimulation occurs before evoked electromyographic signals and to determine differences in electrical excitability of the roots circumferentially. SUMMARY OF BACKGROUND DATA: Stimulus intensities of 1.2-5.7 mA are reported to evoke electromyographic response in corresponding muscles to direct stimulation of normal roots. Stimulus intensities of 6-8 mA were suggested to detect bony pedicular compromise by stimulation of a hole or a screw during pedicle instrumentation. Electrical thresholds of three-dimensional torque response in the ankle to direct root stimulation have not yet been evaluated and compared with thresholds of evoked electromyogram. METHODS: Direct monopolar stimulation of the surgically exposed roots L4, L5, and S1 was performed from different sites around the root by a cuff multielectrode. Biomechanical response was measured as an isometric torque in the ankle at each of three orthogonal axes. Compound muscle action potentials (CMAPs) from root-specific muscles were detected by a pair of surface or wire electrodes. RESULTS: Mean threshold for biomechanical response in the ankle to stimulation of roots L4, L5, and S1 was 0.72 +/- 0.39 mA and for CMAP response was 1.09 mA +/- 0.36 (N = 13). Thresholds for biomechanical responses were significantly lower than for CMAP responses (P = 0.0004; paired t test). Nerve roots were electrically most excitable on their ventral aspects. CONCLUSION: The biomechanical response in the joint to root stimulation can be used to test all root-related muscles crossing that joint at their individual innervation pattern and their residual innervation and to detect electrical excitation of the root at electric thresholds lower than those for detecting CMAP from single standard root-specific muscle. However, this method does not provide sufficient root specificity. It will be valuable in conjunction with multimodality neurophysiologic monitoring of the roots for earlier and more reliable detection of pedicle bone breakthrough or integrity. Further clinical investigations are suggested.