By Peter B. Forgacs, MD
Assistant Professor of Neuroscience and Neurology, Feil Family Brain and Mind Research Institute and Department of Neurology, Weill Cornell Medical College, New York; Adjunct Assistant Professor of Clinical Investigation, The Rockefeller University, New York
Dr. Forgacs reports no financial relationships relevant to this field of study.
SYNOPSIS: In this cohort study of 104 patients with disorders of consciousness, the authors demonstrated a strong correlation between EEG-based metrics and clinical diagnosis using quantitative behavioral scales, brain metabolism as measured by PET, and clinical outcomes at one year.
SOURCE: Srivas C, et al. Brain networks predict metabolism, diagnosis and prognosis at the bedside in disorders of consciousness. Brain 2017;140:2120-2132.
In the last 15-20 years, there have been substantial efforts to accurately assess the potential preservation of organized higher-level neuronal functions in patients with limited or absent bedside evidence of consciousness. Such behavioral states typically arise after a severe brain injury and patients may have prolonged disorders of consciousness (DOC). Categories within DOC include coma (state of unarousable unresponsiveness), vegetative state (VS; distinguished from coma by intermittent eye opening despite unresponsiveness), and minimally conscious state (MCS; characterized by intermittent or inconsistent responses to external stimuli). The current gold standard to diagnose DOC is a standardized bedside clinical scale, the Coma Recovery Scale–Revised (CRS-R). However, obtaining CRS-R requires experienced professionals, and the rate of misdiagnosis in DOC using clinical assessments is high. This study seeks to validate objective measurements that can complement clinical evaluations in patients with DOC.
Recent studies focused on advanced functional neuroimaging (such as functional MRI or 18FDG-PET) to assess residual neuronal functions have provided useful information about structural and functional integrity of brain activity in DOC patients. However, such imaging studies are hindered by logistical constraints that limit their application to studies in research laboratories. Therefore, in recent years, considerable efforts have been devoted to applications of electroencephalogram (EEG) to assess functional brain integrity in patients with DOC. EEG is widely available in most clinical environments, it is easily affordable, and it can be repeated many times without significant risks. In addition, EEG is a direct measure of neuronal electrical activity and allows assessments of functional brain integrity without active participation of patients.
A total of 104 patients were involved in this study — 89 had DOC (23 in VS, 66 in MCS), 11 emerged from MCS, and four patients were in a locked-in state. In addition, 26 control subjects were included in the analysis. A 256-channel high-density EEG was used in all subjects assessed. Resting EEG analyses included relative spectral power estimates in three conventional frequency bands (delta [0-4 Hz], theta [4-8 Hz], and alpha [8-13 Hz]) and an assessment of synchronicity between corresponding EEG electrode pairs. These latter measures also were used to create connectivity matrices that were further applied to calculate seven different graph-theoretical measures of local and global network connectivity (a total of 21 metrics for the three frequency bands analyzed). Each metric was used to estimate the ability to discriminate between evidence of consciousness compared to standardized bedside exam (CRS-R), PET-based criteria for metabolic patterns as related to behavioral diagnosis, and clinical outcome at one-year after study assessments. In addition, these measures were used to train a classifier to predict evidence of consciousness in an individual patient in relation to each of the above-mentioned measures. To further validate their accuracy, the classifiers also were tested in out-of-sample subjects who were not used for training of the classifier.
The authors successfully demonstrated that their EEG metrics of network integrity were associated both with bedside behavioral diagnoses and metabolic activity patterns of the DOC patients, as well as long-term outcomes. Specifically, they showed that the most robust predictors of behavioral level were connectivity measures in the alpha band, especially between lateral and medial frontoparietal areas, consistent with previous electrographic and neuroimaging studies with highly comparable accuracy. Furthermore, they also found a significant correlation between delta-frequency measures and clinical outcomes at one year; stronger delta band connectivity in the central and parietal networks predicted worse outcomes.
This study represents an important step toward development of EEG metrics that can reliably predict level of consciousness in patients with DOC and are suitable to complement bedside clinical assessments. These results provide further evidence that EEG-based assessments of ongoing network-level cerebral activity could be used in everyday clinical practice in the future. The accuracy of the measures applied here is comparable with other, more complex EEG-based (e.g., TMS-EEG) or neuroimaging-based methodologies, but offer a much simpler, near-automated method for EEG analysis. Eventually, these methods will be deployed easily at the patients’ bedside and may provide valuable information about level of consciousness and complement clinical examinations.
However, the methods applied here used high-density (256-channel) EEG recordings that typically are not available in general clinical practice. As suggested, further work is needed to validate these approaches using a reduced number of electrodes, possibly using customized placement over the connectivity hubs demonstrated in this study. In addition, while the methods applied here are almost completely automated, they still require EEG pre-processing by experts using visual inspection of raw EEG recordings to remove artifacts, such as noisy data from movements, muscle activity, or environmental noise. Furthermore, they require complex analyses currently only available in research settings.
- Underwood E. An easy consciousness test? Science 2014;346:531-532.
- Forgacs PB, et al. A proposed role for routine EEGs in patients with consciousness disorders. Ann Neurol 2015;77:185-186.