METHOD: This study is aimed to identify differentiating features of EPR between physiological and pathological population.

RESULTS: A total of 43 patients with pyramidal lesions and 113 normal controls were recruited for this study. The pathological EPRs were more reproducible, with 89.4% having at least two positive Babinski responses and 91.5% having two positive Chaddock responses (vs. 14.3% and 4.8% in controls, P < 0.001). The pathological EPR was more sensitive to stimulation, in which 89.1% were elicited when the stimulation reached mid-lateral sole (vs. 11.9% in controls, P < 0.001). Most (93.6%) pathological cases had sustained big toe extension throughout stimulation (vs. 73.8% in controls, P < 0.001). As compared to those with brain lesion, the plantar responses in those with spinal lesion are less likely to have ankle dorsiflexion (5.3% vs. 25%, P < 0.05) more likely to have sustained extensor response with Babinski (94.7% vs. 71.4%, P < 0.05), Chaddock (89.5% vs. 64.3%, P < 0.05), and Schaefer (26.3% vs. 3.6%, P < 0.05) methods. A scoring system was computed using four variables, i.e., two consecutive positive Babinski or Chaddock responses, extensor response at mid-lateral sole, and sustained extension throughout stimulation. A score ≥3 is predictive of pathological origin, with sensitivity and specificity of 78.7% and 95.2%, respectively.