Welcome to the WCN 2023 Interactive Program

Background and Aims:

Cognitive impairment from epilepsy is a well-recognized. Neuropsychological assessment can detect cognitive impairments through clinical criteria and quantitative measures. In this study, we test the validity of robotic assessment to measure cognitive ability beyond motor function in people with epilepsy by comparing it to a brief neurocognitive assessment.

Methods:

Participants with temporal lobe epilepsy (TLE, n=33) and genetic generalized epilepsy (GGE, n=25) underwent a brief neurocognitive assessment and robotic assessment. Five cognitive and sensorimotor tasks were added to the Repeatable Battery for the Assessment of Neuropsychology Status (RBANS) to form a brief neuropsychological assessment across a range of cognitive domains. The Kinarm Endpoint robot was used for robotic assessment with 9 standardized tasks assessing domains integrating motor, cognitive, and sensory function. Robotic assessment measurements were converted to composite task scores which are adjusted for age, sex, and handedness.

Results:

In the cognitive domains of complex attention (3/6 tests, p<0.05), executive function (3/7 tests, p<0.05), memory (2/4 tests, p<0.05), visual-motor coordination (5/12 tests, p<0.05), and visuospatial skill (1/7 tests, p<0.05) moderate (r ~.30) to strong (r ~.50) associations exist between our brief neurocognitive assessments and robotic assessments.

Conclusions:

These results demonstrate that robotic assessment can measure cognitive ability beyond sensorimotor function, similar to a brief neurocognitive screening for people with both general and focal epilepsies.

Background and Aims:

To investigate the feasibility of using radiofrequency thermocoagulation (RFTC) to selectively destroy key nodes of the epileptic high-frequency oscillation (HFO) network to control seizures, which can provide insights into the microcircuit level of the epileptogenic network.

Methods:

We enrolled 40 patients undergoing RFTC at three comprehensive epilepsy centers. In each patient, we randomly selected 5-minute intracranial electroencephalogram (iEEG) segments of interictal slow-wave sleep for analysis. Using time-delayed network analysis method, we mapped the HFO network. We classified each electrode point located in the HFO network into 3 grade key nodes based on the strength of proactive and passive HFO propagation. We compared the ablation ratios of key nodes at different grades and analyzed changes in HFO characteristics and network connection strengths before and after RFTC in relation to patient prognosis.

Results:

Our results showed that the distribution regions of ripple network were significantly larger than those of the fast ripple network. The rates of HFOs, the area of network distribution, and connection strength were significantly decreased in patients with good outcomes. Complete ablation of grade 1 and 2 key nodes was highly correlated with seizure-free outcome, especially the key nodes (1+2) of ripples. The complete disappearance of ripples are a predicted seizure-free, while newly emerged or increasing FRs inside the peri-RFTC area predicted poor outcomes. Different key nodes showed different theta-HFO coupling features.

Conclusions:

The dynamics of HFO network correspond to the epileptogenic network, complete destruction of HFO network is necessary for good outcome. Cross-frequency coupling mediate the spread of HFOs and may be one of the mechanism of HFO network formation.

Background and Aims:

Current automatic solutions for electroencephalogram (EEG) analyses embed machine-learning approaches that report cases quickly and make robust operator independent measurement of epileptic episodes. Identification of optimal settings for acquiring brain signals can be described as a trade-off dilemma between the high spatial or temporal resolution and the growing complexity of diagnostic models. Finding top-performing EEG parameters has the potential to advance automatic disease detection and severity assessment.

We aim to compare the accuracy of machine learning-based seizure detection and classification by the EEG sampling rate and reveal the optimal temporal resolution of EEG recordings.

Methods:

We used the Temple University Hospital EEG Seizure Corpus dataset to extract data from 33 patients. Next, we built accurate machine learning models for automatic seizure detection and classification. Binary and multigroup classification models were run on the dataset with the acquisition frequency downsampled from 250 to 50 Hz.

Results:

The model accuracy slightly improved when we used the data with the higher sampling rate. The accuracy of the multigroup classification model was 0.8876 with 50 Hz, and 0.9580 with 250Hz. The increase in EEG sampling rate also boosted model precision: that of seizure detection was raised by 5%, and of seizure differentiation - by 7%.

Conclusions:

Seizure detection and classification with machine-learning models can be improved by increasing EEG sampling frequency. Still, low sampling rate can be a reasonable solution for training reliable models. One should also consider the amount of space required to store the brain signal recordings.

Background and Aims:

Background and aims: Around 30% of people with focal epilepsy are medically refractory; for these patients, neurosurgery may be a viable treatment option. However, identifying suitable candidates and localizing the epileptogenic zone remain significant challenges. We introduce and evaluate the clinical utility of three advanced neuroimaging techniques developed at established epilepsy surgery centers in the pre-surgical assessment at the recently designated District Epilepsy Center, Kingston Health Sciences Center (KHSC).

Methods:

Methods: Patients in the pre-surgical pathway were discussed in multidisciplinary team (MDT) meetings (n=11). Patients who had already undergone resections(n=3) and those deemed unsuitable for surgery(n=4) were excluded. Patients meeting the criteria for surgical candidacy with inconclusive clinical data, such as a negative MRI or discordant data (n=4), were recruited for a comprehensive MRI evaluation. This evaluation included high-resolution 3D T1-weighted scans for hippocampal/amygdala (HA) quantitative volumetry, high-resolution 3D FLAIR to facilitate lesion detection, and functional magnetic resonance imaging (fMRI) for language lateralization. The benefits of this additional data were documented in a follow-up questionnaire during the reevaluation of the patients.

Results:

Results: Following our multimodal assessment, previously discussed surgery candidates were now deemed suitable or could proceed to next assessment steps. The protocol has improved localization of epileptic tissue, identified imaging abnormalities, and lateralized language (Table 1), as acknowledged by MDT clinicians.

Conclusions:

Discussion: Our findings demonstrate that these advanced neuroimaging modalities can be introduced into clinical practice at a newly established epilepsy surgery center and provide valuable information to benefit patient care.

Background and Aims:

Neuroinflammation is one of the main factors that mediates the development of seizures and epileptogenesis. CD40L-CD40, a familiar member of TNF, increases in chronic epilepsy, after seizures and post-status epilepticus.

The goal of this research was to determine whether the negative modulation of CD40 limits neuroinflammation in a model of epilepsy

Methods:

The pentylenetetrazole (PTZ) model of seizure or the pilocarpine model of status epilepticus (SE) were induced in adult male mice deficient of CD40 (CD40KO) and their respective age-gender controls Seizure were analyzed clinically using Racine’s score and electrically from local field potential recordings from a silicon probe implanted in the brain. In a group of mice, siRNACD40 or shRNA were injected into the brain before PTZ treatment. Following euthanization, brain samples were processed for histological and biochemical analysis. Secretory sCD40L and a group of cytokines and chemokines were analyzed using ELISA or Meso-Scale Discovery platform. ANOVA, Student’s t test and Z-scores were used for statistical analysis

Results:

Our preliminary data indicates that CD40 and sCD40L increased after seizures; intracerebral siRNACD40 limits seizure susceptibility and sCD40L concentration in the hippocampus. In addition, CD40KO showed: a) a decreased concentration of pro-inflammatory KC/GRO, IL6 and TNF alpha and increased IL-10 after SE; b) reduction of seizure-induced gamma oscillations in hippocampus and recurrent seizures; and c) reduced a chronic brain damage

Conclusions:

. These data suggest that CD40L-CD40 is an key component to initiate an inflammatory cascade in the development of epilepsy.

Background and Aims:

Epilepsy is characterized by hyperactivity of neurons and brain circuits, leading to excessive and synchronous electrical discharges¹ of fundamental early diagnosis. In this sense, the analysis of the prevalence of hospitalizations for epilepsy in children in Brazil in the last 3 years is necessary to assist in the diagnosis and clinical outcome of patients affected by this pathology.

Methods:

It is characterized as transversal, descriptive, retrospective, and quantitative. Data provided by the Department of Informatics of the Brazilian Unified Health System (DATASUS) were used. The information related to hospitalizations for epilepsy that occurred in Brazil from 2020 to 2022 was selected according to the nature of the service, the place of hospitalization, and the gender, and ethnicity of the patients. For this study, patients aged less than 1 year up to 14 years of age were classified as children.

Results:

There were 63,295 hospitalizations in the analyzed period. The region with the highest number of hospitalizations was the Southeast (33.9%), followed by the Northeast (29.43%), South (19.28%), Midwest (10.23%), and North (7.14%). Urgent admissions accounted for 91.79% of the total. The male gender was more affected (53.82%) and patients with brown skin represented 47.19% of admissions.

Conclusions:

Hospitalizations for epilepsy in children in Brazil show a prevalence in brown, male, and emergency patients. The highest number of occurrences occurred in the southeast region and the lowest in the northern part, respectively. Thus, such data offer a better understanding of this neurological disorder's presentation in Brazil's pediatric population.

Background and Aims:

Surgery is very effective for drug-resistant epilepsy. The National Epilepsy Center (NEC) in Sri Lanka which offers an epilepsy surgical program was established in 2017.

Methods:

We report post-operative seizure outcomes, surgical complications, and impact on quality of life among 15 adult patients ≥16 years who underwent epilepsy surgery at NEC. We used Engel classification and Quality of Life in Epilepsy 31 (QOLIE-31) questionnaire to assess seizure outcome and quality of life (QOL) respectively.

Results:

Six female and 9 male patients underwent surgery (mean age 27.7 years, (range 16-44 years) from October 2017 to June 2022. The mean follow-up duration was 32 months (range 6-55 months). Ten patients underwent temporal lobe resections. At 6-months follow-up, 80% (8/10) had favorable seizure outcomes with Engel class I/II. At 1-year follow up 6/8 patients (75%) and at 2-year follow-up, 5/7 patients (71%) had a favorable outcome. Five patients had extra-temporal lobe surgeries, and one of them was defaulted. Seizure freedom was observed in four at 6 months, 3 at one-year, and two at 2-year follow-up.

Four patients (28.6%) experienced minor post-operative surgical site infection. Two (14.2%) had persistent quadrantanopia.

Meaningful improvement in QOL (change in QOLIE-31 score ≥11.8) was observed irrespective of seizure outcome or type of surgery (P value <0.001). QOL did not improve in one.

Conclusions:

Epilepsy surgery is effective for selected patients with drug-resistant epilepsy and is an option in developing countries. Seizure outcomes in our patients are comparable to those worldwide. Clinically important QOL improvement was observed in our series.

Background and Aims:

The alpha rhythm has been a subject of research for the past few decades. Right-left alpha amplitude asymmetry is a common phenomenon. Several explanations have been proposed to explain this asymmetry, including differences in skull thickness. Our research aims to improve our understanding of the relationship between alpha asymmetry and skull thickness measured by CT/MRI images.

Methods:

We reviewed EEGs from our database to examine characteristics of alpha rhythm. We calculated alpha amplitude by measuring the maximum peak to peak in the O1 and O2 reference channels. Alpha asymmetry was deemed significant when it was more than 20%. The skull thickness at locations corresponding to the O1 and O2 channels was assessed using CT/MRI brain and the difference between right and left was calculated. We investigated the relationship between alpha asymmetry and skull thickness asymmetry using the Kruskal-Wallis test and Spearman correlation.

Results:

We reviewed a total of 400 EEGs and images. Patients with alpha asymmetry less than 20 percent were classified as Group 1. Group 2 consisted of patients with higher alpha amplitudes on the right side, while Group 3 included patients with higher amplitudes on left side. We observed a significant difference between the groups of asymmetry and skull thickness difference (p<0.001). The Spearman correlation analysis revealed a statistically significant negative correlation (Rs = -0.259, p < 0.001)

Conclusions:

The present study delves into the factors influencing alpha asymmetry and demonstrates a significant relationship between alpha amplitude asymmetry and skull thickness difference. However, other factors causing asymmetry needs to be explored.