In 1982, we developed a new coma scale, the Glasgow-Liege scale, which combines the quantified analysis of five brain stem reflexes with the Glasgow methodology. The present study was undertaken to determine to what extent agreement exists among different raters assessing brain stem reflexes (Parameter R) and to compare the results with those observed from motor responses (Parameter M). We show the good agreement achieved by different examiners in the evaluation of brain stem reflexes. Brain stem reflexes offer a slightly higher agreement (kappa = 0.69) than that of the study of motor response (kappa = 0.65). Within Parameters M and R, we observed less agreement in the evaluation of flexion responses and in the interpretation of oculocephalic reactions. The reliability of the evaluation of M and R parameters justifies the use of the Glasgow-Liege scale as a means for evaluating disturbances of consciousness.