COVID-19-Associated Neurological Manifestations: An Emerging Electroencephalographic Literature

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has spread worldwide since the end of year 2019 and is currently responsive for coronavirus infectious disease 2019 (COVID-19). The first reports considered COVID-19 as a respiratory tract disease responsible for pneumonia, but numerous studies rapidly emerged to warn the medical community of COVID-19-associated neurological manifestations, including encephalopathy at the acute phase and other postinfectious manifestations. Using standard visual analysis or spectral analysis, recent studies reported electroencephalographic (EEG) findings of COVID-19 patients with various neurological symptoms. Most EEG recordings were normal or revealed non-specific abnormalities, such as focal or generalized slowing, interictal epileptic figures, seizures, or status epilepticus. Interestingly, novel EEG abnormalities over frontal areas were also described at the acute phase. Underlying mechanisms leading to brain injury in COVID-19 are still unknown and matters of debate. These frontal EEG abnormalities could emphasize the hypothesis whereby SARS-CoV-2 enters the central nervous system (CNS) through olfactory structures and then spreads in CNS via frontal lobes. This hypothesis is reinforced by the presence of anosmia in a significant proportion of COVID-19 patients and by neuroimaging studies confirming orbitofrontal abnormalities. COVID-19 represents a new viral disease characterized by not only respiratory symptoms but also a systemic invasion associated with extra-respiratory signs. Neurological symptoms must be the focus of our attention, and functional brain evaluation with EEG is crucial, in combination with anatomical and functional brain imaging, to better understand its pathophysiology. Evolution of symptoms together with EEG patterns at the distance of the acute episode should also be scrutinized.


INTRODUCTION
The coronavirus infectious disease 2019 , caused by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection, was initially recognized as a respiratory tract disease which could lead to an acute respiratory distress syndrome. However, there is growing evidence of a multi-organ involvement (Gupta et al., 2020). Several authors

Seizures and SE
Seizures and/or SE were recorded in 10 patients out of 111 included in the series of Pellinen et al. (2020), in 2 out of 22 in the series of Louis et al. (2020), in 1 out of 37 in the series of Ayub et al. (2020), in 1 out of 15 in the series of Pasini et al. (2020), in 1 out of 27 in the series of Scullen et al. (2020), and in an unknown precise number of patients out of the 73 included in the series of Chougar et al. (2020) (Table 1).
Seizures and/or SE were recorded in reports of patients without any acute or chronic cortical lesions on brain imaging nor cerebrospinal fluid (CSF) abnormalities. The EEG of the patient reported by Balloy et al. (2020) revealed two widespread, but predominantly in frontal localizations, seizures that were interrupted by a moderate interictal frontal activity. Sohal and Mansur (2020) reported a patient whose 24-h EEG revealed six left temporal seizures and left temporal sharp waves. One of the two patients reported by Somani et al. (2020) displayed, on a continuous EEG (cEEG) monitoring, multiple seizures emanating from the midline and left frontocentral regions ( Figure 1B). Hepburn et al. (2020) reported the cases of two patients whose cEEG monitoring showed, for the first one, three focal seizures arising from the right frontocentral region and, Central pontine myelinolysis, n=3

PRES, n=2
Meningeal enhancement, n=2 Neuritis, n=2 Deep venous thrombosis, n=1 Corticospinal tracts FLAIR hyperintensity, n=1 >Perfusion MRI, n=46 Seizure-related perfusion abnormalities, n=9 Recent or old vascular lesions-related perfusion abnormalities, n=4 Perfusion abnormalities unrelated to seizures or ischemia, n=10 Subset of post-anoxic comas, n=2 Severely suppressed activity, n=1 Discontinued activity compatible with post-anoxic SE, n=1 Unreactive, n=2 Frontiers in Physiology | www.frontiersin.org for the second one, left more than right frontotemporal seizures which progressed to focal SE. The EEG of the patient reported by Le Guennec et al. (2020) revealed a non-convulsive SE (NCSE) over the right frontal region. The brain MRI of this patient only showed peri-ictal diffusion abnormalities over the right orbital and mesial prefrontal cortex and right caudate nucleus. Flamand et al. (2020) reported the case of a patient who benefited from several EEG. The first two EEG findings were consistent with a bilateral frontal SE. One EEG in the series of five patients reported by Chen et al. (2020) showed a bifrontal SE, and another one revealed a generalized NCSE. Finally, the EEG of the patient reported by Rodrigo-Armenteros et al. (2020) showed a bilateral frontotemporal NCSE. Seizures and/or SE were recorded more rarely in patients with acute CNS lesions on brain imaging and/or significant CSF abnormalities, of either vascular or inflammatory origin. Among the four patients with a PRES reported by Parauda et al. (2020), two had seizures or SE emanating from posterior regions: for the first one, a focal NCSE arising from the left posterior quadrant and, for the second one, focal seizures arising from the right posterior quadrant. The history of a 2-month-old boy was published by Schupper et al. (2020). His brain imaging revealed multiple infarctions with hemorrhagic transformations, and his cEEG showed NCSE. Zanin et al. (2020) published the case of a patient with diffuse CNS demyelinating lesions on brain and spine imaging whose EEG revealed two seizures starting from the right frontotemporal region and diffusing in the homologous contralateral hemisphere. Hussein et al. (2020) reported the case of a patient with an ADEM whose EEG revealed left hemispheric seizures and, 3 days later, brief focal right posterior seizures. Finally, Bernard-Valnet et al. (2020) reported the history of a patient with a lymphocytic meningitis on CSF analysis with normal brain MRI whose EEG showed a focal anterior NCSE.
Seizures and/or SE were recorded in patients with a prior neurological history and radiological sequelae but without any acute lesions. The EEG of the second patient, who had a prior history of skull base surgery, reported by Somani et al. (2020) showed recurrent seizures emanating from either right or left frontocentroparietal regions. Vollono et al. (2020) reported the case of a left frontocentrotemporal SE in a patient with a remote herpes simplex virus 1 encephalitis.
Seizures were reported on cEEG in the series of 33 patients published by Radmard et al. (2020), as frontotemporal and parasagittal seizures in two patients, but without precise imaging or CSF results available for these two patients.

Rhythmic and Periodic Discharges
Rhythmic discharges were mentioned in series, as generalized rhythmic delta activity (GRDA) (Ayub et al., 2020;Galanopoulou et al., 2020;Louis et al., 2020;Pellinen et al., 2020;Petrescu et al., 2020), lateralized rhythmic delta activity (LRDA) (Ayub et al., 2020;Galanopoulou et al., 2020;Louis et al., 2020;Pellinen et al., 2020), and frontal intermittent rhythmic delta activity (Canham et al., 2020;Pasini et al., 2020) (Table 1). Rhythmic patterns were also reported in a few case reports. Vandervorst et al. (2020) published the EEG of a patient with a clinical and radiological picture of encephalitis with temporal bilateral more left than right imaging abnormalities. The EEG showed short-lasting left temporal LRDA. In the series of Beach et al. (2020), one patient, with a previous history of dementia with Lewy bodies and remote traumatic brain injury, displayed GRDA with sharp contouring and bifrontal predominance, without any acute lesions on brain imaging. The EEG of the three other patients reported in the series of Chen et al. (2020) previously mentioned revealed GRDA, with unremarkable CSF analysis for the three and no acute lesions on brain imaging for one of them (unavailable for the two others). One EEG recorded among the seven patients reported by Anand et al. (2020) showed GRDA in a patient with extensive leukoencephalopathy on brain MRI and normal CSF sample.
Periodic discharges were noted in series, as generalized periodic discharges (GPDs) (Ayub et al., 2020;Galanopoulou et al., 2020;Louis et al., 2020;Pellinen et al., 2020;Petrescu et al., 2020;Vespignani et al., 2020) and lateralized periodic discharges (LPDs) (Pellinen et al., 2020;Petrescu et al., 2020;Vespignani et al., 2020) (Table 1). Especially, in the series of Vespignani et al. (2020), five EEGs out of 26 showed periodic discharges. Four of these five patients were under mechanical ventilation (MV), and three were sedated. One patient suffered from a cardiac arrest. EEG showed periodic (with a < 4 s interval), monomorphic biphasic, delta activity, which was diffuse with frontal predominance for four and lateralized over right frontal area for one. The second patient reported in the work of Beach al. previously mentioned presented with a left-sided acute-onchronic subdural hematoma (SDH) due to a fall with head trauma. The EEG showed frequent runs of epileptiform GPDs (Beach et al., 2020). Young et al. (2020) reported 1-1.5 Hz LPDs and diffuse delta-theta slowing in a patient who displayed Creutzfeldt-Jakob disease in tandem with symptomatic onset of COVID-19. Conte et al. (2020) published the history of a patient who presented a severe COVID-19 pneumonia and then a PRES-like encephalopathy. She displayed focal seizures, and after seizure treatment, EEG revealed LPDs in the right posterior regions. Vellieux et al. (2020) published the EEG of two critically ill patients who displayed a severe COVID-19 pneumonia requiring MV. For the first one, the brain MRI was consistent with a hypoxic encephalopathy, and the EEG was recorded while he was sedated and under extracorporeal membrane oxygenation. For the second one, the EEG was recorded 24 h after sedation discontinuation. EEG revealed continuous, symmetric, non-reactive, generalized but mainly bifrontal, monomorphic diphasic or even triphasic, periodic (with a short interval of 1-2 s) delta slow waves ( Figure 1C). One patient, without any acute abnormalities on brain MRI and with normal CSF analysis, reported by Delorme et al. (2020) showed GPDs. In the previously mentioned case reported by Le Guennec et al. (2020), a control follow-up EEG was recorded the day after the first EEG. It showed persistent right frontal LPDs with a short interval (0.7-1.2 s). The brain MRI performed 1 month later was normal. Finally, the previously mentioned patient reported by Flamand et al. (2020) who benefited from iterative EEG showed, on the last two recordings, a generalized periodic triphasic activity with short periods (1-1.5 s) over a worsened background activity, without concomitant metabolic disorders.

Spectral Analysis
Two studies reported quantitative analysis of EEG (qEEG) in COVID-19 patients. The study of Pastor et al. (2020) reported 20 patients with COVID-19 encephalopathy for whom standard visual analysis of EEG showed scarce abnormalities. However, compared to 31 infectious toxic encephalopathy patients and 21 post-cardiorespiratory arrest encephalopathy patients, some qEEG features were specific in COVID-19 patients, such as the distribution of EEG bands, the structure of Shannon's spectral entropy, and the hemispheric connectivity. Finally, the study of Pati et al. (2020) showed that some qEEG markers, especially an increase in both the theta power and its temporal variance during EEG reactivity, can predict a good neurological outcome in 10 critically ill COVID-19 patients.

DISCUSSION
The vast majority of these studies emphasized the absence of specificity of EEG abnormalities reported in COVID-19 patients, as generalized slowing of the background activity, focal slowing sometimes associated with sharp waves, seizures, SE, and predictable pattern of metabolic/toxic or postanoxic encephalopathy in ICU patients. Numerous EEGs in the context of COVID-19 were recorded in elderly patients and mainly in male patients, with multiple comorbidities especially chronic brain disorders, under various psychotropic drugs or in critically ill conditions. Confounding factors such as infections, metabolic disturbances, severe hypoxemia, hyperthermia, and psychotropic drugs (such as antiepileptic or sedative drugs) were frequent at the time of EEG recordings. All these confounding factors may contribute to the modification of brain activity and therefore EEG findings. Thus, based on the current literature, it seems not possible to identify a specific EEG pattern due to the suspected neuroinvasion of SARS-CoV-2 in patients who displayed neurological manifestations of COVID-19.
Most current studies with available EEG data are case reports or retrospective single-center series. All reported patients are very heterogeneous concerning prior neurological histories, illness severity, and use of psychotropic drugs. Moreover, some studies reported EEG recorded with limited montage and number of electrodes that may limit the detection of EEG abnormalities. EEG is not a systematic exam in the diagnostic workup of COVID-19 patients. All patients reported in the current literature had an EEG for an urgent clinical indication due to concerning neurological symptoms. A wider neurological multimodality screening, including EEG, of COVID-19 patients may be suggested to grow the body of knowledge on the SARS-CoV-2 infection. However, it will face many logistic difficulties and ethical and safety concerns regarding the availability of trained personnel to EEG recordings and the risk of contamination with the SARS-CoV-2.
It should be pointed out that many EEG abnormalities reported were recorded over anterior or frontal regions. Regardless of EEG montage used by clinicians and neurophysiologists, it thus seems essential to include frontal electrodes. Periodicity, morphology, and reactivity of these frontal abnormalities were not mentioned in all studies.

CONCLUSION
In the context of the SARS-CoV-2 infection, increasing EEG results were published along with clinical reports describing various neurological symptoms in patients with COVID-19. Due to the suspected neuroinvasion of SARS-CoV-2, the major issue when interpreting EEG is to determine whether the observed abnormalities reflect this viral neuroinvasion, a severe encephalopathy with systemic and brain inflammation, hypoxemia and hyperthermia, and/or many confounding factors especially due to critical illness. At this time, no study had described specific EEG abnormalities of the SARS-CoV-2 infection. The majority of currently reported EEGs showed generalized slowing, focal slowing, epileptiform discharges with seizures, and SE. However, frontal discharges, for some periodic, may integrate in the olfactory hypothesis of the CNS invasion of SARS-CoV-2. It reinforces the need to accumulate precise neurophysiological observations of COVID-19 patients worldwide and to aggregate multimodality screening of these patients also with clinical, radiological, biological, and neuropathological data.

AUTHOR CONTRIBUTIONS
GV collected the data and wrote the manuscript. RS, SV, PJ, and AR-T revised the manuscript. M-PO suggested and revised the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING STATEMENT
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.