Small-world characteristics of EEG patterns in post-anoxic encephalopathy

Diagnostic and prognostic EEG analysis of critically ill patients: A deep learning study

Visual interpretation of electroencephalography (EEG) is time consuming, may lack objectivity, and is restricted to features detectable by a human. Computer-based approaches, especially deep learning, could potentially overcome these limitations. However, most deep learning studies focus on a specific question or a single pathology. Here we explore the potential of deep learning for EEG-based diagnostic and prognostic assessment of patients with acute consciousness impairment (ACI) of various etiologies. EEGs from 358 adults from a randomized controlled trial (CERTA, NCT03129438) were retrospectively analyzed. A convolutional neural network was used to predict the clinical outcome (based either on survival or on best cerebral performance category) and to determine the etiology (four diagnostic categories). The largest probability output served as marker for the confidence of the network in its prediction ("certainty factor"); we also systematically compared the predictions with raw EEG data, and used a visualization algorithm (Grad-CAM) to highlight discriminative patterns. When all patients were considered, the area under the receiver operating characteristic curve (AUC) was 0.721 for predicting survival and 0.703 for predicting the outcome based on best CPC; for patients with certainty factor ≥ 60 % the AUCs increased to 0.776 and 0.755 respectively; and for certainty factor ≥ 75 % to 0.852 and 0.879. The accuracy for predicting the etiology was 54.5 %; the accuracy increased to 67.7 %, 70.3 % and 84.1 % for patients with certainty factor of 50 %, 60 % and 75 % respectively. Visual analysis showed that the network learnt EEG patterns typically recognized by human experts, and suggested new criteria. This work demonstrates for the first time the potential of deep learning-based EEG analysis in critically ill patients with various etiologies of ACI. Certainty factor and post-hoc correlation of input data with prediction help to better characterize the method and pave the route for future implementations in clinical routine.

EEG functional connectivity contributes to outcome prediction of postanoxic coma

OBJECTIVE

To investigate the additional value of EEG functional connectivity features, in addition to non-coupling EEG features, for outcome prediction of comatose patients after cardiac arrest.

METHODS

Prospective, multicenter cohort study. Coherence, phase locking value, and mutual information were calculated in 19-channel EEGs at 12 h, 24 h and 48 h after cardiac arrest. Three sets of machine learning classification models were trained and validated with functional connectivity, EEG non-coupling features, and a combination of these. Neurological outcome was assessed at six months and categorized as "good" (Cerebral Performance Category [CPC] 1-2) or "poor" (CPC 3-5).

RESULTS

We included 594 patients (46% good outcome). A sensitivity of 51% (95% CI: 34-56%) at 100% specificity in predicting poor outcome was achieved by the best functional connectivity-based classifier at 12 h after cardiac arrest, while the best non-coupling-based model reached a sensitivity of 32% (0-54%) at 100% specificity using data at 12 h and 48 h. Combination of both sets of features achieved a sensitivity of 73% (50-77%) at 100% specificity.

CONCLUSION

Functional connectivity measures improve EEG based prediction models for poor outcome of postanoxic coma.

SIGNIFICANCE

Functional connectivity features derived from early EEG hold potential to improve outcome prediction of coma after cardiac arrest.

Brain functional connectivity during the first day of coma reflects long-term outcome

•

Coma patients show different connectivity patterns depending on long-term outcome.

•

Time-variance of functional connectivity is an early prognostic marker for coma patients.

•

Connectivity patterns observed in chronic patients may develop early after coma onset.

Objective

In patients with disorders of consciousness (DOC), properties of functional brain networks at rest are informative of the degree of consciousness impairment and of long-term outcome. Here we investigate whether connectivity differences between patients with favorable and unfavorable outcome are already present within 24 h of coma onset.

Methods

We prospectively recorded 63-channel electroencephalography (EEG) at rest during the first day of coma after cardiac arrest. We analyzed 98 adults, of whom 57 survived beyond unresponsive wakefulness. Functional connectivity was estimated by computing the ‘debiased weighted phase lag index’ over epochs of five seconds duration. We evaluated the network’s topological features, including clustering coefficient, path length, modularity and participation coefficient and computed their variance over time. Finally, we estimated the predictive value of these topological features for patients’ outcomes by splitting the patient sample in training and test datasets.

Results

Group-level analysis revealed lower clustering coefficient, higher modularity and path length variance in patients with favorable compared to those with unfavorable outcomes (p < 0.01). Within all features, the path length variance in the network provided the best positive predictive value (PPV) for favorable outcome and specificity for unfavorable outcome in the test dataset (PPV: 0.83, p < 0.01; specificity: 0.86, p < 0.01) with above-chance negative predictive value and accuracy. Of note, the exclusion of patients with epileptiform activity (20 in total) eliminates all false positive predictions (n = 6) for path length variance.

Interpretation

Topological features of functional connectivity differ as a function of long-term outcome in patients on the first day of coma. These differences are not interpretable in terms of consciousness levels as all patients were in a deep unconscious state. The time variance of path length is informative of comatose patients’ outcome, as patients with favorable outcome exhibit a richer repertoire of path length than those with unfavorable outcomes.

Quantitative Electroencephalogram Trends Predict Recovery in Hypoxic-Ischemic Encephalopathy

OBJECTIVES

Electroencephalogram features predict neurologic recovery following cardiac arrest. Recent work has shown that prognostic implications of some key electroencephalogram features change over time. We explore whether time dependence exists for an expanded selection of quantitative electroencephalogram features and whether accounting for this time dependence enables better prognostic predictions.

DESIGN

Retrospective.

SETTING

ICUs at four academic medical centers in the United States.

PATIENTS

Comatose patients with acute hypoxic-ischemic encephalopathy.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We analyzed 12,397 hours of electroencephalogram from 438 subjects. From the electroencephalogram, we extracted 52 features that quantify signal complexity, category, and connectivity. We modeled associations between dichotomized neurologic outcome (good vs poor) and quantitative electroencephalogram features in 12-hour intervals using sequential logistic regression with Elastic Net regularization. We compared a predictive model using time-varying features to a model using time-invariant features and to models based on two prior published approaches. Models were evaluated for their ability to predict binary outcomes using area under the receiver operator curve, model calibration (how closely the predicted probability of good outcomes matches the observed proportion of good outcomes), and sensitivity at several common specificity thresholds of interest. A model using time-dependent features outperformed (area under the receiver operator curve, 0.83 ± 0.08) one trained with time-invariant features (0.79 ± 0.07; p < 0.05) and a random forest approach (0.74 ± 0.13; p < 0.05). The time-sensitive model was also the best-calibrated.

CONCLUSIONS

The statistical association between quantitative electroencephalogram features and neurologic outcome changed over time, and accounting for these changes improved prognostication performance.

Disrupted resting-state brain functional network in methamphetamine abusers: A brain source space study by EEG

This study aimed to examine the effects of chronic methamphetamine use on the topological organization of whole-brain functional connectivity network (FCN) by reconstruction of neural-activity time series at resting-state. The EEG of 36 individuals with methamphetamine use disorder (IWMUD) and 24 normal controls (NCs) were recorded, pre-processed and source-reconstructed using standardized low-resolution tomography (sLORETA). The brain FCNs of participants were constructed and between-group differences in network topological properties were investigated using graph theoretical analysis. IWMUD showed decreased characteristic path length, increased clustering coefficient and small-world index at delta and gamma frequency bands compared to NCs. Moreover, abnormal changes in inter-regional connectivity and network hubs were observed in all the frequency bands. The results suggest that the IWMUD and NCs have distinct FCNs at all the frequency bands, particularly at the delta and gamma bands, in which deviated small-world brain topology was found in IWMUD.

EEG-based outcome prediction after cardiac arrest with convolutional neural networks: Performance and visualization of discriminative features

Prognostication for comatose patients after cardiac arrest is a difficult but essential task. Currently, visual interpretation of electroencephalogram (EEG) is one of the main modality used in outcome prediction. There is a growing interest in computer-assisted EEG interpretation, either to overcome the possible subjectivity of visual interpretation, or to identify complex features of the EEG signal. We used a one-dimensional convolutional neural network (CNN) to predict functional outcome based on 19-channel-EEG recorded from 267 adult comatose patients during targeted temperature management after CA. The area under the receiver operating characteristic curve (AUC) on the test set was 0.885. Interestingly, model architecture and fine-tuning only played a marginal role in classification performance. We then used gradient-weighted class activation mapping (Grad-CAM) as visualization technique to identify which EEG features were used by the network to classify an EEG epoch as favorable or unfavorable outcome, and also to understand failures of the network. Grad-CAM showed that the network relied on similar features than classical visual analysis for predicting unfavorable outcome (suppressed background, epileptiform transients). This study confirms that CNNs are promising models for EEG-based prognostication in comatose patients, and that Grad-CAM can provide explanation for the models' decision-making, which is of utmost importance for future use of deep learning models in a clinical setting.

EEG synchronization measures are early outcome predictors in comatose patients after cardiac arrest

OBJECTIVE

Outcome prognostication in comatose patients after cardiac arrest (CA) remains a major challenge. Here we investigated the prognostic value of combinations of linear and non-linear bivariate EEG synchronization measures.

METHODS

94 comatose patients with EEG within 24h after CA were included. Clinical outcome was assessed at 3months using the Cerebral Performance Categories (CPC). EEG synchronization between the left and right parasagittal, and between the frontal and parietal brain regions was assessed with 4 different quantitative measures (delta power asymmetry, cross-correlation, mutual information, and transfer entropy). 2/3 of patients were used to assess the predictive power of all possible combinations of these eight features (4 measures×2 directions) using cross-validation. The predictive power of the best combination was tested on the remaining 1/3 of patients.

RESULTS

The best combination for prognostication consisted of 4 of the 8 features, and contained linear and non-linear measures. Predictive power for poor outcome (CPC 3-5), measured with the area under the ROC curve, was 0.84 during cross-validation, and 0.81 on the test set. At specificity of 1.0 the sensitivity was 0.54, and the accuracy 0.81.

CONCLUSION

Combinations of EEG synchronization measures can contribute to early prognostication after CA. In particular, combining linear and non-linear measures is important for good predictive power.

SIGNIFICANCE

Quantitative methods might increase the prognostic yield of currently used multi-modal approaches.

EEG as an Indicator of Cerebral Functioning in Postanoxic Coma

Postanoxic coma after cardiac arrest is one of the most serious acute cerebral conditions and a frequent cause of admission to critical care units. Given substantial improvement of outcome over the recent years, a reliable and timely assessment of clinical evolution and prognosis is essential in this context, but may be challenging. In addition to the classic neurologic examination, EEG is increasingly emerging as an important tool to assess cerebral functions noninvasively. Although targeted temperature management and related sedation may delay clinical assessment, EEG provides accurate prognostic information in the early phase of coma. Here, the most frequently encountered EEG patterns in postanoxic coma are summarized and their relations with outcome prediction are discussed. This article also addresses the influence of targeted temperature management on brain signals and the implication of the evolution of EEG patterns over time. Finally, the article ends with a view of the future prospects for EEG in postanoxic management and prognostication.

Prognostic value of electroencephalography (EEG) for brain injury after cardiopulmonary resuscitation

Cardiac arrest (CA) patients can experience neurological sequelae or even death after successful cardiopulmonary resuscitation (CPR) due to cerebral hypoxia- and ischemia-reperfusion-mediated brain injury. Thus, it is important to perform early prognostic evaluations in CA patients. Electroencephalography (EEG) is an important tool for determining the prognosis of hypoxic-ischemic encephalopathy due to its real-time measurement of brain function. Based on EEG, burst suppression, a burst suppression ratio >0.239, periodic discharges, status epilepticus, stimulus-induced rhythmic, periodic or ictal discharges, non-reactive EEG, and the BIS value based on quantitative EEG may be associated with the prognosis of CA after successful CPR. As measures of neural network integrity, the values of small-world characteristics of the neural network derived from EEG patterns have potential applications.

Electroencephalographic Findings in Posthypoxic Myoclonus

The physical examination findings of early posthypoxic myoclonus (PHM) are associated with poor prognosis. Recent findings indicate that patients with multifocal PHM, assumed to have a cortical origin, have a comparable outcome to resuscitated patients without PHM. Generalized PHM, assumed to have a subcortical myoclonus origin, is still associated with a bad clinical outcome. It is not known whether the electroencephalographic (EEG) findings differ between the multifocal and generalized myoclonus groups nor is the clinical significance clearly defined. Forty-three patients with PHM were retrospectively derived from an EEG database. Patients were categorized as having multifocal (i), generalized (ii), or undetermined (iii) PHM. Outcome was expressed in cerebral performance category scores. The EEG background was categorized into isoelectric (I), low voltage (II), burst suppression (III), status epilepticus (SE; IV), diffuse slowing (V), and mild encephalopathic or normal (VI). 17 patients had generalized PHM and 23 had multifocal PHM (3 undetermined). The EEG showed more SE in generalized compared to multifocal PHM (64% vs 13%, P< .001). Diffuse slowing was more often present in multifocal PHM (52% vs 17%, P < .05). Early-onset myoclonus occurred significantly more often in generalized PHM, and early generalized PHM was invariantly associated with poor outcome. In conclusion, patients with generalized PHM showed more SE. These EEG findings might be either subcortical corollaries or primarily cortical phenomena. Our retrospective results conflict with currently used clinical criteria for myoclonus classification, and we suggest that more refined difference may be needed for accurate assessment of PHM. To better understand PHM, prospective research with standardized clinical assessment and quantitative EEG analysis is needed.