Data_Sheet_1_The accuracy of different mismatch negativity amplitude representations in predicting the levels of consciousness in patients with disorders of consciousness.PDF

IntroductionThe mismatch negativity (MMN) index has been used to evaluate consciousness levels in patients with disorders of consciousness (DoC). Indeed, MMN has been validated for the diagnosis of vegetative state/unresponsive wakefulness syndrome (VS/UWS) and minimally conscious state (MCS). In this study, we evaluated the accuracy of different MMN amplitude representations in predicting levels of consciousness.
MethodsTask-state electroencephalography (EEG) data were obtained from 67 patients with DoC (35 VS and 32 MCS). We performed a microstate analysis of the task-state EEG and used four different representations (the peak amplitude of MMN at electrode Fz (Peak), the average amplitude within a time window −25– 25 ms entered on the latency of peak MMN component (Avg for peak ± 25 ms), the average amplitude of averaged difference wave for 100–250 ms (Avg for 100–250 ms), and the average amplitude difference between the standard stimulus (“S”) and the deviant stimulus (“D”) at the time corresponding to Microstate 1 (MS1) (Avg for MS1) of the MMN amplitude to predict the levels of consciousness.
ResultsThe results showed that among the four microstates clustered, MS1 showed statistical significance in terms of time proportion during the 100–250 ms period. Our results confirmed the activation patterns of MMN through functional connectivity analysis. Among the four MMN amplitude representations, the microstate-based representation showed the highest accuracy in distinguishing different levels of consciousness in patients with DoC (AUC = 0.89).
ConclusionWe discovered a prediction model based on microstate calculation of MMN amplitude can accurately distinguish between MCS and VS states. And the functional connection of the MS1 is consistent with the activation mode of MMN.

引言
失匹配负波（mismatch negativity, MMN）指标已被用于评估意识障碍（disorders of consciousness, DoC）患者的意识水平。事实上，MMN已被验证可用于诊断植物状态/无反应觉醒综合征（vegetative state/unresponsive wakefulness syndrome, VS/UWS）与最小意识状态（minimally conscious state, MCS）。本研究评估了不同MMN振幅表征方式对意识水平的预测准确性。

方法
本研究从67例意识障碍患者（其中35例为VS/UWS，32例为MCS）处获取了任务态脑电图（electroencephalography, EEG）数据。我们对任务态EEG进行微状态分析，并采用四种不同的MMN振幅表征方式预测意识水平：分别为Fz电极处MMN的峰值振幅（Peak）、以MMN成分峰值潜伏期为中心的-25~25ms时间窗内的平均振幅（峰值±25ms平均振幅）、100~250ms时段平均差异波的平均振幅（100~250ms平均振幅），以及MMN振幅中对应微状态1（Microstate 1, MS1）时刻标准刺激("S")与偏差刺激("D")之间的平均振幅差（MS1平均振幅）。

结果
结果显示，在聚类得到的四种微状态中，MS1在100~250ms时段的时间占比具有统计学意义。本研究通过功能连接分析验证了MMN的激活模式。在四种MMN振幅表征方式中，基于微状态的表征方式在区分意识障碍患者的不同意识水平方面表现出最高的准确率（曲线下面积Area Under Curve, AUC=0.89）。

结论
本研究发现，基于MMN振幅微状态计算的预测模型可准确区分MCS与VS/UWS状态。且MS1的功能连接模式与MMN的激活模式一致。