Disorders of consciousness are a heterogeneous mixture of different diseases or injuries. Although some indicators and models have been proposed for prognostication, any single method when used alone carries a high risk of false prediction. This study aimed to develop a multidomain prognostic model that combines resting state functional MRI with three clinical characteristics to predict one year-outcomes at the single-subject level. The model discriminated between patients who would later recover consciousness and those who would not with an accuracy of around 88% on three datasets from two medical centers. It was also able to identify the prognostic importance of different predictors, including brain functions and clinical characteristics. To our knowledge, this is the first reported implementation of a multidomain prognostic model that is based on resting state functional MRI and clinical characteristics in chronic disorders of consciousness, which we suggest is accurate, robust, and interpretable.

Effective treatment and accurate long-term prognostication of patients with disorders of consciousness (DOC) remain pivotal clinical issues and challenges in neuroscience. Previous studies have shown that zolpidem produces paradoxical recovery and induces similar change patterns in specific electrophysiological features in some DOC (∼6%). However, whether these specific features are neural markers of responders, and how neural features evolve over time remain unclear. Here, we capitalized on static and dynamic EEG analysis techniques to fully uncover zolpidem-induced alterations in eight patients with DOC and constructed machine-learning models to predict long-term outcomes at the single-subject level. We observed consistent patterns of change across all patients in several static features (e.g., decreased relative theta power and weakened alpha-band functional connectivity) after zolpidem administration, albeit none zolpidem responders. Based on the current evidence, previously published electrophysiological features are not neural markers for zolpidem responders. Moreover, we found that the temporal dynamics of the brain slowed down after zolpidem intake. Brain states before and after zolpidem administration could be completely characterized by the EEG features. Furthermore, long-term outcomes were accurately predicted using connectivity features. Our findings suggest that EEG neural signatures have huge potential to assess consciousness states and predict fine-grained outcomes. In summary, our results extend the understanding of the effects of zolpidem on the brain and open avenues for the application prospect of zolpidem and EEG in patients with DOC.

Background: Brain assessment shows great values in prognosis, treatment, resource allocation, and decision-making for patients with disorders of consciousness (DOC). However, less research focused on cortical conditions of patients with unresponsive wakefulness syndrome (UWS).Methods: We recorded resting-state EEG and TMS-EEG from post-anoxic patients with UWS, diagnosed by repeated Coma Recovery Scale-Revised (CRS-R). Measurements of functional connectivity and networks were performed by phase lock value (PLV) and network parameters of graph theory (average path length, clustering coefficient, and small-world). Global cortical reactivity values (GCRV) were used to assess cortical excitability.Results: The coefficient of variation (CV) presented marked inter-individual variations of PLV (CV = 0.285), network parameters (CV > 0.2), and GCRV (CV = 0.929) within these patients. The patients’ PLV and network parameters at theta and alpha bands significantly correlated with their GCRV values. Patients with higher PLV (r = 0.560, 0.406), as well as better preserved network (lower average path length (r = −0.522, −0.483), higher clustering coefficient (r = 0.522, 0.445), and small-world (r = 0.522, 0.445) at theta and alpha bands, presented higher GCRV. The functional connectivity, which is significantly correlated with frontal GCRV, is also mainly located in the frontal region. These correlations were not significant at other frequency bands: Delta, beta, and gamma bands.Conclusion: These findings suggested that the CRS-R-diagnosed post-anoxic patients with UWS had very different cortical conditions. Functional networks and cortical excitability measured by TMS-EEG could complement behavioral assessment to assess these patients’ cortical conditions.Significance: It provides a deeper understanding of neurophysiological dysfunction in patients with UWS and hints to the clinics that neural-electrophysiological assessment for such patients may be necessary to acquire their brain conditions, which may benefit stratified management for them.