EEG Training Research Articles

Comparison Between Methods of Choosing EEG Training Data in Epileptic Seizures Prediction Systems

Considering the large number of people that suffer with epilepsy, there is great interest in developing a system capable of predicting the occurrence of the seizures. Warning patients about the proximity of a seizure can help them to avoid dangerous situations. Many efforts in the development of this system have been placed using, basically, machine learning techniques. In most of these works, regardless of the technique employed, it is assumed that the system is able to learn when the patient’s brain signals, obtained from electroencephalogram (EEG), change from interictal to pre-ictal state. Some published works do not explicitly clarify the method of choosing training and test data. This means that good results may be due to the use of an inadequate method. That is, the EEG segments may have been temporarily mixed up during training, which would help the network correctly predict the seizure. However, this would not work from a real point of view. The objective of this work is to investigate, through experiments, the effect of the way of choosing the training and test data on the accuracy of a seizure prediction system. Experimental results showed that the accuracy of the systems increases when training is performed with windows close to the seizure used for testing.

User-wise perturbations for user identity protection in EEG-based BCIs.

An electroencephalogram (EEG)-based brain-computer interface (BCI) is a direct communication pathway between the human brain and a computer. Most research so far studied more accurate BCIs, but much less attention has been paid to the ethics of BCIs. Aside from task-specific information, EEG signals also contain rich private information, e.g., user identity, emotion, disorders, etc., which should be protected. We show for the first time that adding user-wise perturbations can make identity information in EEG unlearnable. We propose four types of user-wise privacy-preserving perturbations, i.e., random noise, synthetic noise, error minimization noise, and error maximization noise. After adding the proposed perturbations to EEG training data, the user identity information in the data becomes unlearnable, while the BCI task information remains unaffected. Experiments on six EEG datasets using three neural network classifiers and various traditional machine learning models demonstrated the robustness and practicability of the proposed perturbations. Our research shows the feasibility of hiding user identity information in EEG data without impacting the primary BCI task information.

EEG training in Neurology Residency at Egyptian Universities

EEG training in Neurology Residency at Egyptian Universities

Session-independent subject-adaptive mental imagery BCI using selective filter-bank adaptive Riemannian features.

The brain-computer interfaces (BCIs) facilitate the users to exploit information encoded in neural signals, specifically electroencephalogram (EEG), to control devices and for neural rehabilitation. Mental imagery (MI)-driven BCI predicts the user's pre-meditated mental objectives, which could be deployed as command signals. This paper presents a novel learning-based framework for classifying MI tasks using EEG-based BCI. In particular, our work focuses on the variation in inter-session data and the extraction of multi-spectral user-tailored features for robust performance. Thus, the goal is to create a calibration-free subject-adaptive learning framework for various mental imagery tasks not restricted to motor imagery alone. In this regard, critical spectral bands and the best temporal window are first selected from the EEG training trials of the subject based on the Riemannian user learning distance metric (Dscore) that checks for distinct and stable patterns. The filtered covariance matrices of the EEG trials in each spectral band are then transformed towards a reference covariance matrix using the Riemannian transfer learning, enabling the different sessions to be comparable. The evaluation of our proposed Selective Time-window and Multi-scale Filter-Bank with Adaptive Riemannian (STFB-AR) features on four public datasets, including disabled subjects, showed around 15% and 8% improvement in mean accuracy over baseline and fixed filter-bank models, respectively.

The online educational tool "Roadmap to EEGs" significantly improved trainee performance in recognizing EEG patterns.

We created a framework to assess the competency-based EEG curriculum, outlined by the International League Against Epilepsy (ILAE) through a video-based online educational resource ("Roadmap to EEGs") and assessed its effectiveness and feasibility in improving trainees' knowledge. Ten video-based e-learning modules addressed seven key topics in EEG and epileptology (normal EEG, normal variants, EEG artifacts, interictal epileptiform discharges (IED), focal seizures, idiopathic generalized epilepsy (IGE), and developmental and epileptic encephalopathies (DEE)). We posted the educational videos on YouTube for free access. Pre- and post-tests, each comprising 20 multiple-choice questions, were distributed to institution leadership and advertised on social media platforms to reach a global audience. The tests were administered online to assess the participants' knowledge. Pre- and post-test questions showed different EEG samples to avoid memorization and immediate recall. After completing the post-test, participants were asked to respond to 7 additional questions assessing their confidence levels and recommendations for improvement. A total of 52 complete and matched pre- and post-test responses were collected. The probability of a correct response was 73% before teaching (95% CI: 70%-77%) and 81% after teaching (95% CI: 78%-84%). The odds of a correct response increased significantly by 59% (95% CI: 28%-98%, p < .001). For participants having >4 weeks of EEG training, the probability of a correct response was 76% (95% CI: .72-.79) and 81% after teaching (95% CI: .78-.84). The odds of answering correctly increased by 44% (95% CI: 15%-80%, p = .001). Participants felt completely confident in independently interpreting and identifying EEG findings after completing the teaching modules (17.1% before vs. 37.8% after, p-value < .0001). 86.5% of participants expressed a high likelihood of recommending the module to other trainees. The video-based online educational resource allows participants to acquire foundational knowledge in EEG/epilepsy, and participants to review previously learned EEG/epilepsy information.

Education Research: Competency-Based EEG Education: An Online Routine EEG Examination for Adult and Child Neurology Residents.

We recently published expert consensus-based curricular objectives for routine EEG (rEEG) interpretation for adult and child neurology residents. In this study, we used this curriculum framework to develop and validate an online, competency-based, formative and summative rEEG examination for neurology residents. We developed an online rEEG examination consisting of a brief survey and 30 multiple-choice questions covering EEG learning objectives for neurology residents in 4 domains: normal, abnormal, normal variants, and artifacts. Each question contained a deidentified EEG image, displayed in 2 montages (bipolar and average), reviewed and optimized by the authors to address the learning objectives. Respondents reported their level of confidence (LOC, 5-point Likert scale) with identifying 4 categories of EEG findings independently: states of wakefulness/sleep, sleep structures, normal variants, and artifacts. Accuracy and item discrimination were calculated for each question and LOC for each category. The test was disseminated by the International League Against Epilepsy and shared on social media. Of 2,080 responses, 922 were complete. Respondents comprised clinical neurophysiologists/experts (n = 41), EEG/epilepsy clinical fellows (n = 211), EEG technologists (n = 128), attending neurologists (n = 111), adult neurology residents (n = 227), child neurology residents (n = 108), medical students (n = 24), attending non-neurologists (n = 18), and others (n = 54). Mean overall scores (95% CI) were 82% (77-86) (clinical neurophysiologists), 81% (79-83) (clinical fellows), and 72% (70-73) (adult and child neurology residents). Experts were more confident than clinical fellows in all categories but sleep structures. Experts and clinical fellows were more confident than residents in all 4 categories. Among residents, accuracy and LOC increased as a function of prior EEG weeks of training. Accuracy improved from 67% (baseline/no prior EEG training) to 77% (>12 prior EEG weeks). More than 8 weeks of EEG training was needed to reach accuracy comparable with clinical neurophysiologists on this rEEG examination. Increase in LOC was slower and less robust than increase in accuracy. All but 3 questions had a high discrimination index (>0.25). This online, competency-based rEEG examination, mapped to a published EEG curriculum, has excellent psychometrics and differentiates experienced EEG readers from adult and child neurology residents. This online tool has the potential to improve resident EEG education worldwide.

The Effects of EEG Biofeedback Training on Visual Reaction Time in Judo Athletes.

The aim of the study was to expand the current knowledge on the effects of EEG biofeedback training on the reaction time of judo athletes, as well as to develop an optimal EEG training protocol in terms of the number of sessions and their duration that would significantly improve the reaction time of athletes. The study included 24 male athletes from the national team of the Polish Judo Association. The selected group was randomly divided into two subgroups: experimental and control. The study was conducted in four cycles varying in terms of frequency and duration of neurofeedback (NFB) sessions, both in the control and experimental groups. In the experimental group, each training cycle consisted of 15 sessions, followed by a four-week break. The effects of NFB training on the visual reaction time of judo athletes were evaluated using computerized simple and complex reaction time tests along with selected trials of the Vienna Test System (VTS). Following NFB training according to the theta/beta1 protocol, while maintaining appropriate duration and frequency of individual training sessions, statistically significant improvements in reaction times to visual stimuli of athletes, both in simple and complex tasks, were observed in the experimental group. No such changes were found in the control group. The greatest improvement in reaction times was observed in complex tasks, indicating the high effectiveness of EEG biofeedback training in enhancing this ability.

Virtual and In-person Electroencephalography (EEG) Training Among Pediatric and Adult Neurology Residents During the COVID-19 Pandemic

Background: During the COVID-19 pandemic, educational programs have increasingly relied on virtual methods. Objectives: In this study, we compared the effects of in-person and virtual electroencephalography (EEG) training on the knowledge of pediatric and adult neurology residents. Methods: The study participants consisted of 30 pediatric and adult neurology residents who were recruited nationwide via a virtual network. They were randomly divided into two groups for in-person and virtual education, respectively. Prior to the first workshop session, participants completed a pre-test comprising 19 knowledge-related questions. Two interactive training sessions focusing on benign variants in EEG were conducted, including two in-person workshops for eight residents and two online workshops for 22 residents. Each session lasted one hour, with one held every two weeks. After the second session, participants were asked to complete a post-test consisting of 24 questions (identical to the pre-test) and five workshop satisfaction-related questions. Results: Overall, the mean exam score after both the in-person and virtual workshops was significantly higher than the pretest score across all educational levels. The mean score differences in knowledge between the in-person and virtual groups after the workshop were similar, with no statistically significant difference. Approximately 90% of the participants believed that the workshop would be beneficial for their daily practice. Conclusions: Given the absence of a significant difference between in-person and virtual training methods in improving participants' knowledge and satisfaction with both approaches, we recommend integrating virtual EEG training into the educational protocol for pediatric and adult neurology residents. Nonetheless, larger sample size studies are required to further validate these findings.

Neonatal EEG graded for severity of background abnormalities in hypoxic-ischaemic encephalopathy

This report describes a set of neonatal electroencephalogram (EEG) recordings graded according to the severity of abnormalities in the background pattern. The dataset consists of 169 hours of multichannel EEG from 53 neonates recorded in a neonatal intensive care unit. All neonates received a diagnosis of hypoxic-ischaemic encephalopathy (HIE), the most common cause of brain injury in full term infants. For each neonate, multiple 1-hour epochs of good quality EEG were selected and then graded for background abnormalities. The grading system assesses EEG attributes such as amplitude, continuity, sleep–wake cycling, symmetry and synchrony, and abnormal waveforms. Background severity was then categorised into 4 grades: normal or mildly abnormal EEG, moderately abnormal EEG, majorly abnormal EEG, and inactive EEG. The data can be used as a reference set of multi-channel EEG for neonates with HIE, for EEG training purposes, or for developing and evaluating automated grading algorithms.

Leveraging Multiple Distinct EEG Training Sessions for Improvement of Spectral-Based Biometric Verification Results.

Most studies on EEG-based biometry recognition report results based on signal databases, with a limited number of recorded EEG sessions using the same single EEG recording for both training and testing a proposed model. However, the EEG signal is highly vulnerable to interferences, electrode placement, and temporary conditions, which can lead to overestimated assessments of the considered methods. Our study examined how different numbers of distinct recording sessions used as training sessions would affect EEG-based verification. We analyzed the original data from 29 participants with 20 distinct recorded sessions each, as well as 23 additional impostors with only one session each. We applied raw coefficients of power spectral density estimate, and the coefficients of power spectral density estimate converted to the decibel scale, as the input to a shallow neural network. Our study showed that the variance introduced by multiple recording sessions affects sensitivity. We also showed that increasing the number of sessions above eight did not improve the results under our conditions. For 15 training sessions, the achieved accuracy was 96.7 ± 4.2%, and for eight training sessions and 12 test sessions, it was 94.9 ± 4.6%. For 15 training sessions, the rate of successful impostor attacks over all attack attempts was 3.1 ± 2.2%, but this number was not significantly different from using six recording sessions for training. Our findings indicate the need to include data from multiple recording sessions in EEG-based recognition for training, and that increasing the number of test sessions did not significantly affect the obtained results. Although the presented results are for the resting-state, they may serve as a baseline for other paradigms.

A general sample-weighted framework for epileptic seizure prediction

Objective:Effective epileptic seizure prediction can make the patients know the onset of the seizure in advance to take timely preventive measures. Many studies based on machine learning methods have been proposed to tackle this problem and achieve significant progress in recent years. However, most studies treat each EEG training sample’s contribution to the model as equal, while different samples have different predictive effects on epileptic seizures (e.g., preictal samples from different times). To this end, in this paper, we propose a general sample-weighted framework for patient-specific epileptic seizure prediction. Methods:Specifically, we define the mapping from the sample weights of training sets to the performance of the validation sets as the fitness function to be optimized. Then, the genetic algorithm is employed to optimize this fitness function and obtain the optimal sample weights. Finally, we obtain the final model by using the training sets with optimized sample weights. Results:To evaluate the effectiveness of our framework, we conduct extensive experiments on both traditional machine learning methods and prevalent deep learning methods. Our framework can significantly improve performance based on these methods. Among them, our framework based on Transformer achieves an average sensitivity of 94.6%, an average false prediction rate of 0.06/h, and an average AUC of 0.939 in 12 pediatric patients from the CHB-MIT database with the leave-one-out method, which outperforms the state-of-the-art methods. Conclusion:This study provides new insights into the field of epileptic seizure prediction by considering the discrepancies between EEG samples. Moreover, we develop a general sample-weighted framework, which applies to almost all classical classification methods and can significantly improve performance based on these methods.

Implementation of an electroencephalogram-guided propofol anesthesia education program in an academic pediatric anesthesia practice.

Propofol total intravenous anesthesia (TIVA) is increasingly popular in pediatric anesthesia, but education on its use is variable and over-dosage adverse events are not uncommon. Recent work suggests that electroencephalogram (EEG) parameters can guide propofol dosing in the pediatric population. This education quality improvement project aimed to implement a standardized EEG TIVA training program over 12 months in a large pediatric anesthesia division. The division consisted of 63 faculty, 11 clinical fellows, 32 residents, and 28 nurse anesthetists at the Children's Hospital of Philadelphia. The program was assessed for effectiveness (a significant improvement in EEG knowledge scores), scalability (training 50% of fellows and staff), and sustainability (recurring EEG lectures for 80% of rotating residents and 100% of new fellows and staff). The key drivers included educational content development (lectures, articles, and hand-outs), training a cohort of EEG TIVA trainers, intraoperative teaching (teaching points and dosing tables), decision support tools (algorithms and anesthesia electronic record pop-ups), and knowledge tests (written exam and verbal quiz during cases). Over 12 months, 78.5% of the division (62/79) completed EEG training and test scores improved (mean score 38% before training vs 59% after training, p < .001). Didactic lectures were given to 100% of the fellows, 100% (11/11) of new staff, and 80% (4/5 blocks) of rotating residents. This quality improvement education project successfully trained pediatric anesthesia faculty, staff, residents, and fellows in EEG-guided TIVA. The training program was effective, scalable, and sustainable over time for newly hired faculty staff and rotating fellows and residents.

Exploiting pretrained CNN models for the development of an EEG-based robust BCI framework

Identifying motor and mental imagery electroencephalography (EEG) signals is imperative to realizing automated, robust brain-computer interface (BCI) systems. In the present study, we proposed a pretrained convolutional neural network (CNN)-based new automated framework feasible for robust BCI systems with small and ample samples of motor and mental imagery EEG training data. The framework is explored by investigating the implications of different limiting factors, such as learning rates and optimizers, processed versus unprocessed scalograms, and features derived from untuned pretrained models in small, medium, and large pretrained CNN models. The experiments were performed on three public datasets obtained from BCI Competition III. The datasets were denoised with multiscale principal component analysis, and time-frequency scalograms were obtained by employing a continuous wavelet transform. The scalograms were fed into several variants of ten pretrained models for feature extraction and identification of different EEG tasks. The experimental results showed that ShuffleNet yielded the maximum average classification accuracy of 99.52% using an RMSProp optimizer with a learning rate of 0.000 1. It was observed that low learning rates converge to more optimal performances compared to high learning rates. Moreover, noisy scalograms and features extracted from untuned networks resulted in slightly lower performance than denoised scalograms and tuned networks, respectively. The overall results suggest that pretrained models are robust when identifying EEG signals because of their ability to preserve the time-frequency structure of EEG signals and promising classification outcomes.

EEG education in Brazil: a national survey of adult neurology residents.

In light of the established challenges of resident EEG education worldwide, we sought to better understand the current state of neurology resident EEG education in Brazil. To define Brazilian EEG practices including in-residency requirements for EEG training and competency. We assessed the perspectives of adult residents (PGY1-3) on EEG education and their level of confidence interpreting EEG with a 24-question online survey. We analyzed 102 responses from 52 Brazilian neurology residency programs distributed in 14 states. There were 18 PGY1s, 45 PGY2s, and 39 PGY3s. Ninety-six percent of participants reported that learning how to read EEG during residency was very or extremely important. The most commonly reported barriers to EEG education were insufficient EEG exposure (70%) and ineffective didactics (46%). Residents believed that standard EEG lectures were the most efficient EEG teaching method followed by interpreting EEG with attendings' supervision. Roughly half of residents (45%) reported not being able to read EEG even with supervision, and approximately 70% of all participants did not feel confident writing an EEG report independently. Despite the well-established residency EEG education requirements recommended by the Brazilian Academy of Neurology (ABN), there seems to be a significant lack of comfort interpreting EEG among Brazilian adult neurology residents. We encourage Brazilian neurology residency leadership to re-evaluate the current EEG education system in order to ensure that residency programs are following EEG education requirements and to assess whether EEG benchmarks require modifications.

Neurology resident EEG training in Europe

ObjectiveTo detail current European EEG education practices and compare European and U.S. EEG teaching systems. MethodsA 19-question online survey focused on EEG clinical practices and residency training was emailed to all 47 European Academy of Neurology Societies. ResultsThirty-two (68 %) out of the 47 Societies completed the survey. In half of countries, general neurologists are either among the providers or the only providers who typically read EEGs. The number of weeks devoted to EEG learning required to graduate ranged from none to 26, and it was expected to be continuous in one country. In most countries (n = 17/32), trainees read >40 EEGs per EEG rotation, and the most commonly interpreted studies are routine and prolonged routine EEGs. Rotations involve clinic/outpatient (90 %), epilepsy monitoring unit/inpatient (60 %), or both (50 %). Roughly half of countries do not use objective measures to assess EEG competency. The most reported educational methods are teaching during EEG rotation and yearly didactics, and the most reported education barriers are insufficient didactics and insufficient EEG exposure. ConclusionsWe suggest neurology educators in Europe, especially in those countries where EEGs are read by general neurologists, consider ensuring that residency EEG learning is mandatory and establishing objective measures in teaching and evaluating competency. SignificanceSimilar to the U.S., neurology resident EEG training in Europe is highly variable.

Motor Imagery Classification for Brain Computer Interface Using Deep Convolutional Neural Networks and Mixup Augmentation.

Goal: Building a DL model that can be trained on small EEG training set of a single subject presents an interesting challenge that this work is trying to address. In particular, this study is trying to avoid the need for long EEG data collection sessions, and without combining multiple subjects training datasets, which has a detrimental effect on the classification performance due to the inter-individual variability among subjects. Methods: A customized Convolutional Neural Network with mixup augmentation was trained with [Formula: see text]120 EEG trials for only one subject per model. Results: Modified ResNet18 and DenseNet121 models with mixup augmentation achieved 0.920 (95% Confidence Interval: 0.908, 0.933) and 0.933 (95% Confidence Interval: 0.922, 0.945) classification accuracy, respectively. Conclusions: We show that the designed classifiers resulted in a higher classification performance in comparison to other DL classifiers of previous studies on the same dataset, despite the limited training dataset used in this work.

Understanding the landscape of electrophysiology services for children in sub-Saharan Africa.

Incidence of epilepsy is greatest in infancy and childhood; this is especially evident in sub-Saharan Africa (SSA). The aim of this study was to understand access to electrophysiology services in SSA including which health practitioner performs and interprets paediatric electroencephalogram (EEG) studies as well as their training in paediatric EEG. A web-based survey was sent to a cohort of health care practitioners who manage children with epilepsy in SSA. The questions addressed whether EEG was available to these health care practitioners, how the practitioners accessed EEG and who assisted interpretation of the study results. The survey was circulated (June-December 2019) to 305 participants from 32 African countries. A total of 73 (16 partial and 57 complete) surveys were returned from 18 countries. The respondents fell into two main categories: those with access to an EEG machine (44/73; 60%) and those without access to an EEG machine (29/73; 40%). For 32% (23/73), there was no dedicated technician and for 34% (25/73) no neurologist. Access to a neurologist resulted in the highest proportion of EEGs performed per annum. Of the respondents, 77% (56/73) agreed that there was a need for a paediatric apprenticeship in EEG skills. Qualitative data to justify need for paediatric EEG training was grouped into three themes: (1) "professional development"; (2) "better care"; and (3) "help paediatric patients and neurologists". There is a lack of paediatric EEG training amongst doctors and technicians working with epilepsy in SSA. Expanding training beyond current capacity in SSA, for technicians and practitioners involved in EEG, is necessary.

MDD-TSVM: A novel semisupervised-based method for major depressive disorder detection using electroencephalogram signals

Major depressive disorder (MDD) is a common mental illness characterized by persistent feeling of depressed mood and loss of interest. It would cause, in a severe case, suicide behaviors. In clinical settings, automatic MDD detection is mainly based on electroencephalogram (EEG) signals with supervised learning techniques. However, supervised-based MDD detection methods encounter two ineviTable bottlenecks: firstly, such methods rely heavily on an EEG training dataset with MDD labels annotated by a physical therapist, leading to subjectivity and high cost; secondly, most of EEG signals are unlabeled in a real scenario. In this paper, a novel semisupervised-based MDD detection method named MDD-TSVM is presented. Specifically, the MDD-TSVM utilizes the semisupervised method of transductive support vector machine (TSVM) as its backbone, further dividing the unlabeled penalty item of the TSVM objective function into two pseudo-labeled penalty items with or without MDD. By such improvement, the MDD-SVM can make full use of labeled and unlabeled datasets as well as alleviate the class imbalance problem. Experiment results showed that our proposed MDD-TSVM achieved F1 score of 0.85 ± 0.05 and accuracy of 0.89 ± 0.03 on identifying MDD patients, which is superior to the state-of-the-art methods.

Reducing the Calibration Effort of EEG Emotion Recognition using Domain Adaptation with Soft Labels.

Electroencephalogram (EEG)-based emotion recognition has made great progress in recent years. The current pipelines collect EEG training data in a long-time calibration session for each new subject, which is time consuming and user unfriendly. To reduce the time required for the calibration session, there have been many studies using domain adaptation (DA) approaches to transfer knowledge from existing subjects (source domain) to the new subject (target domain) for reducing the dependence on the calibration session. Existing DA methods usually require substantial unlabeled EEG data of the new subject. However, the real scenario is that there are a small number of labeled samples in the calibration session of the target. Motivated by this, we introduce a novel domain adaptation architecture based on adversarial training to learn domain-invariant feature representations across subjects. To improve the performance when there are few labeled EEG data in the calibration session, we add a soft label loss to the architecture, which can ensure that the inter-class relationships learned from the source domain are transferred to target domain. We evaluate the method on the SEED dataset, and the experimental results show that our method uses only 15 examples per trial in the calibration session to achieve an average accuracy of 87.28%, indicating the effectiveness of our framework.

Evaluating the Efficacy of an Online Learning Tool for EEG Teaching: A Prospective Cohort Study.

To determine the effectiveness of a 6-month, interactive, multimodal, Web-based EEG teaching program (EEGonline) in improving EEG analysis and interpretation skills for neurologists, neurology residents, and technologists, particularly in resource-limited settings. Between June 2017 and November 2018, 179 learners originating from 20 African countries, Europe, and the United States were registered on the EEGonline course. Of these, 128 learners (91% African) participated in the study. Pre- and postcourse multiple choice question (MCQ) test results and EEGonline user logs were analyzed. Differences in pre- and posttest performance were correlated with quantified exposure to various EEGonline learning modalities. Participants' impressions of EEGonline efficacy and usefulness were assessed through pre- and postcourse satisfaction surveys. Ninety-one participants attempted both pre- and postcourse tests (71% response rate). Mean scores improved from 46.7% ± 17.6% to 64.1% ± 18%, respectively (p < 0.001, Cohen d 0.974). The largest improvement was in correct identification of normal features (43.2%-59.1%; p < 0.001, Cohen d 0.664) and artifacts (43.3%-61.6%; p < 0.001, Cohen d 0.836). Improvement in knowledge was associated with improved subjective confidence in EEG analysis. Overall confidence among postcourse survey respondents improved significantly from 35.9% to 81.9% (p < 0.001). Lecture notes, self-assessment quizzes, and discussion forums were the most utilized learning modalities. The majority of survey respondents (97.2%) concluded that EEGonline was a useful learning tool and 93% recommended that similar courses should be included in EEG training curricula. This study demonstrated that a multimodal, online EEG teaching tool was effective in improving EEG analysis and interpretation skills and may be useful in resource-poor settings.