High resolution automated labeling of the hippocampus and amygdala using a 3D convolutional neural network trained on whole brain 700 µm isotropic 7T MP2RAGE MRI.

Image labeling using convolutional neural networks (CNNs) are a template-free alternative to traditional morphometric techniques. We trained a 3D deep CNN to label the hippocampus and amygdala on whole brain 700 µm isotropic 3D MP2RAGE MRI acquired at 7T. Manual labels of the hippocampus and amygdala were used to (i) train the predictive model and (ii) evaluate performance of the model when applied to new scans. Healthy controls and individuals with epilepsy were included in our analyses. Twenty-one healthy controls and sixteen individuals with epilepsy were included in the study. We utilized the recently developed DeepMedic software to train a CNN to label the hippocampus and amygdala based on manual labels. Performance was evaluated by measuring the dice similarity coefficient (DSC) between CNN-based and manual labels. A leave-one-out cross validation scheme was used. CNN-based and manual volume estimates were compared for the left and right hippocampus and amygdala in healthy controls and epilepsy cases. The CNN-based technique successfully labeled the hippocampus and amygdala in all cases. Mean DSC = 0.88 ± 0.03 for the hippocampus and 0.8 ± 0.06 for the amygdala. CNN-based labeling was independent of epilepsy diagnosis in our sample (p = .91). CNN-based volume estimates were highly correlated with manual volume estimates in epilepsy cases and controls. CNNs can label the hippocampus and amygdala on native sub-mm resolution MP2RAGE 7T MRI. Our findings suggest deep learning techniques can advance development of morphometric analysis techniques for high field strength, high spatial resolution brain MRI.

Systematic review of EEG findings in 617 patients diagnosed with COVID-19.

OBJECTIVE: We performed a systematic review of the literature to synthesize the data on EEG findings in COVID-19. Frontal EEG patterns are reported to be a characteristic finding in COVID-19 encephalopathy. Although several reports of EEG abnormalities are available, there is lack of clarity about typical findings. METHODS: Research databases were queried with the terms "COVID" OR "coronavirus" OR "SARS" AND "EEG". Available data was analyzed from 617 patients with EEG findings reported in 84 studies. RESULTS: The median age was 61.3 years (IQR 45-69, 33.3 % female). Common EEG indications were altered mental status (61.7 %), seizure-like events (31.2 %), and cardiac arrest (3.5 %). Abnormal EEG findings (n = 543, 88.0 %) were sub-classified into three groups: (1) Background abnormalities: diffuse slowing (n = 423, 68.6 %), focal slowing (n = 105, 17.0 %), and absent posterior dominant rhythm (n = 63, 10.2 %). (2) Periodic and rhythmic EEG patterns: generalized periodic discharges (n = 35, 5.7 %), lateralized/multifocal periodic discharges (n = 24, 3.9 %), generalized rhythmic activity (n = 32, 5.2 %). (3) Epileptiform changes: focal (n = 35, 5.7 %), generalized (n = 27, 4.4 %), seizures/status epilepticus (n = 34, 5.5 %). Frontal EEG patterns comprised of approximately a third of all findings. In studies that utilized continuous EEG, 96.8 % (n = 243) of the 251 patients were reported to have abnormalities compared to 85.0 % (n = 311) patients who did not undergo continuous EEG monitoring (χ2 = 22.8, p =< 0.001). SIGNIFICANCE: EEG abnormalities are common in COVID-19 related encephalopathy and correlates with disease severity, preexisting neurological conditions including epilepsy and prolonged EEG monitoring. Frontal findings are frequent and have been proposed as a biomarker for COVID-19 encephalopathy.

Simultaneous scalp EEG improves seizure lateralization during unilateral intracranial EEG evaluation in temporal lobe epilepsy.

PURPOSE: To determine if simultaneous bilateral scalp EEG (scEEG) can accurately detect a contralateral seizure onset in patients with unilateral intracranial EEG (IEEG) implantation. METHODS: We evaluated 39 seizures from 9 patients with bitemporal epilepsy who underwent simultaneous scEEG and IEEG (SSIEEG). To simulate conditions of unilateral IEEG implantation with a missed contralateral seizure onset, we analyzed the IEEG recording contralateral to the seizure onset (CL- IEEG), in conjunction with simultaneous scEEG. The following criteria were evaluated between scEEG and CL- IEEG (1) latency: the time to onset of EEG seizure (2) location: concordance of ictal onset zones and (3) pattern: congruence of EEG morphology and frequency. RESULTS: SSIEEG correctly lateralized 36/39 (92.3%) seizures compared to 13/39 (33.3%) seizures using CL- IEEG alone (OR = 24.0, p < 0.01), 33 (84.6%) seizures using scEEG alone (OR = 2.2, p = 0.29) and 26 (66.9%) seizures using time of clinical onset alone (OR = 6.0, p = 0.01). For the three criteria evaluated, (1) 22/39 (56.4%) seizures had an earlier onset on the scEEG, compared to CL- IEEG; (2) lack of congruence of location of seizure onset was noted in 33/39 (84.6%) of the seizures; and (3) 22/39 (56.4%) seizures did not have a congruent ictal pattern. CONCLUSIONS: The chronological, topographic and morphologic features of SSIEEG can accurately detect the hemisphere of seizure onset in most cases with unilateral IEEG implantation. SSIEEG is significantly better than, IEEG, scEEG or clinical onset alone in this scenario. We propose that SSIEEG should be considered in all cases of intractable focal epilepsy undergoing unilateral IEEG evaluation.

Randomized study of intravenous valproate and phenytoin in status epilepticus.

INTRODUCTION: The evidence based data to guide management in patients of benzodiazepine refractory status epilepticus (SE) is still lacking. We conducted a randomized study to evaluate the comparative effect of intravenous (IV) phenytoin and intravenous valproate (IV VA) in patients of benzodiazepine refractory SE. BACKGROUND AND METHODS: Hundred, age and sex matched, patients of benzodiazepine refractory SE were randomly divided into Group A (50 patients), treated with IV VA and Group B (50 patients) treated with IV phenytoin. Twelve patients, in whom SE was not controlled with a single drug, were switched over to the other group. Treatment was considered successful when all motor or EEG seizure activity ceased within 20 min after the beginning of the drug infusion and no return of seizure activity during the next 12h. Secondary study end points were adverse events to treatment, in-hospital complications and the neurological outcome at discharge. RESULTS: In this study, IV VA was successful in 88% and IV phenytoin in 84% (p>0.05) of patients of SE with a significantly better response in patients of SE <2h (p<0.05). The total number of adverse events did not differ significantly between the two groups (p>0.05). There were no differences among the treatments with respect to recurrence after 12-h study period or the outcome at 7 days. CONCLUSION: IV VA is as effective as IV phenytoin. It is easy to use, better tolerated and can be used as an alternative to IV phenytoin in patients of benzodiazepine refractory SE, especially in patients of cardio-respiratory disease. The better outcome in patients having shorter duration of SE (<2h) suggests need of immediate treatment.