Sensory-Motor Systems Lab, Institute of Robotics and Intelligent Systems, Department of Health Science and Technologies, ETH Zurich, Tannenstrasse 1, Zurich 8092, Switzerland.
J Neuroeng Rehabil. 2012 May 30;9:30. doi: 10.1186/1743-0003-9-30.
Clinical scores represent the gold standard in characterizing the clinical condition of patients in vegetative or minimally conscious state. However, they suffer from problems of sensitivity, specificity, subjectivity and inter-rater reliability.In this feasibility study, objective measures including physiological and neurophysiological signals are used to quantify the clinical state of 13 low-responsive patients. A linear regression method was applied in nine patients to obtain fixed regression coefficients for the description of the clinical state. The statistical model was extended and evaluated with four patients of another hospital. A linear mixed models approach was introduced to handle the challenges of data sets obtained from different locations.Using linear backward regression 12 variables were sufficient to explain 74.4% of the variability in the change of the clinical scores. Variables based on event-related potentials and electrocardiogram account for most of the variability.These preliminary results are promising considering that this is the first attempt to describe the clinical state of low-responsive patients in such a global and quantitative way. This new model could complement the clinical scores based on objective measurements in order to increase diagnostic reliability. Nevertheless, more patients are necessary to prove the conclusions of a statistical model with 12 variables.
临床评分是描述处于植物人或最小意识状态患者临床状况的金标准。然而，它们存在敏感性、特异性、主观性和评分者间可靠性等问题。在这项可行性研究中，使用包括生理和神经生理信号在内的客观测量来量化 13 名低反应性患者的临床状态。在 9 名患者中应用线性回归方法，获得用于描述临床状态的固定回归系数。该统计模型通过另外 4 名来自另一家医院的患者进行扩展和评估。引入线性混合模型方法来处理来自不同地点的数据的挑战。使用线性向后回归，有 12 个变量足以解释临床评分变化的 74.4%的可变性。基于事件相关电位和心电图的变量占大部分可变性。考虑到这是首次尝试以这种全局和定量的方式描述低反应性患者的临床状态，这些初步结果很有希望。这种新模型可以补充基于客观测量的临床评分，以提高诊断的可靠性。然而，需要更多的患者来证明具有 12 个变量的统计模型的结论。
