Expert-Level Intracranial Electroencephalogram Ictal Pattern Detection by a Deep Learning Neural Network

Artificial intelligence in epilepsy - applications and pathways to the clinic

Artificial intelligence (AI) is rapidly transforming health care, and its applications in epilepsy have increased exponentially over the past decade. Integration of AI into epilepsy management promises to revolutionize the diagnosis and treatment of this complex disorder. However, translation of AI into neurology clinical practice has not yet been successful, emphasizing the need to consider progress to date and assess challenges and limitations of AI. In this Review, we provide an overview of AI applications that have been developed in epilepsy using a variety of data modalities: neuroimaging, electroencephalography, electronic health records, medical devices and multimodal data integration. For each, we consider potential applications, including seizure detection and prediction, seizure lateralization, localization of the seizure-onset zone and assessment for surgical or neurostimulation interventions, and review the performance of AI tools developed to date. We also discuss methodological considerations and challenges that must be addressed to successfully integrate AI into clinical practice. Our goal is to provide an overview of the current state of the field and provide guidance for leveraging AI in future to improve management of epilepsy.

Utility of Chronic Intracranial Electroencephalography in Responsive Neurostimulation Therapy

Responsive neurostimulation (RNS) therapy is an effective treatment for reducing seizures in some patients with focal epilepsy. Utilizing a chronically implanted device, RNS involves monitoring brain activity signals for user-defined patterns of seizure activity and delivering electrical stimulation in response. Devices store chronic data including counts of detected activity patterns and brief recordings of intracranial electroencephalography signals. Data platforms for reviewing stored chronic data retrospectively may be used to evaluate therapy performance and to fine-tune detection and stimulation settings. New frontiers in RNS research can leverage raw chronic data to reverse engineer neurostimulation mechanisms and improve therapy effectiveness.

Differentiation of subclinical and clinical electrographic events in long-term electroencephalographic recordings

OBJECTIVE

With the advent of ultra-long-term recordings for monitoring of epilepsies, the interpretation of results of isolated electroencephalographic (EEG) recordings covering only selected brain regions attracts considerable interest. In this context, the question arises of whether detected ictal EEG patterns correspond to clinically manifest seizures or rather to purely electrographic events, that is, subclinical events.

METHODS

EEG patterns from 268 clinical seizures and 252 subclinical electrographic events from 50 patients undergoing video-EEG monitoring were analyzed. Features extracted included predominant frequency band, duration, association with rhythmic muscle artifacts, spatial extent, and propagation patterns. Classification using logistic regression was performed based on data from the whole dataset of 10-20 system EEG recordings and from a subset of two temporal electrode contacts.

RESULTS

Correct separation of clinically manifest and purely electrographic events based on 10-20 system EEG recordings was possible in up to 83.8% of events, depending on the combination of features included. Correct classification based on two-channel recordings was only slightly inferior, achieving 78.6% accuracy; 74.4% and 74.8%, respectively, of events could be correctly classified when using duration alone with either electrode set, although classification accuracies were lower for some subgroups of seizures, particularly focal aware seizures and epileptic arousals.

SIGNIFICANCE

A correct classification of subclinical versus clinical EEG events was possible in 74%-83% of events based on full EEG recordings, and in 74%-78% when considering only a subset of two electrodes, matching the channel number available from new implantable diagnostic devices. This is a promising outcome, suggesting that ultra-long-term low-channel EEG recordings may provide sufficient information for objective seizure diaries. Intraindividual optimization using high numbers of ictal events may further improve separation, provided that supervised learning with external validation is feasible.

At home EEG monitoring technologies for people with epilepsy and intellectual disabilities: A scoping review

BACKGROUND

Conducting electroencephalography in people with intellectual disabilities (PwID) can be challenging, but the high proportion of PwID who experience seizures make it an essential part of their care. To reduce hospital-based monitoring, interventions are being developed to enable high-quality EEG data to be collected at home. This scoping review aims to summarise the current state of remote EEG monitoring research, potential benefits and limitations of the interventions, and inclusion of PwID in this research.

METHODS

The review was structured using the PRISMA extension for Scoping Reviews and the PICOS framework. Studies that evaluated a remote EEG monitoring intervention in adults with epilepsy were retrieved from the PubMed, MEDLINE, Embase, CINAHL, Web of Science, and ClinicalTrials.gov databases. A descriptive analysis provided an overview of the study and intervention characteristics, key results, strengths, and limitations.

RESULTS

34,127 studies were retrieved and 23 were included. Five types of remote EEG monitoring were identified. Common benefits included producing useful results of comparable quality to inpatient monitoring and patient experience. A common limitation was the challenge of capturing all seizures with a small number of localised electrodes. No randomised controlled trials were included, few studies reported sensitivity and specificity, and only three considered PwID.

CONCLUSIONS

Overall, the studies demonstrated the feasibility of remote EEG interventions for out-of-hospital monitoring and their potential to improve data collection and quality of care for patients. Further research is needed on the effectiveness, benefits, and limitations of remote EEG monitoring compared to in-patient monitoring, especially for PwID.

Deep net detection and onset prediction of electrographic seizure patterns in responsive neurostimulation

OBJECTIVE

Managing the progress of drug-resistant epilepsy patients implanted with the Responsive Neurostimulation (RNS) System requires the manual evaluation of hundreds of hours of intracranial recordings. The generation of these large amounts of data and the scarcity of experts' time for evaluation necessitate the development of automatic tools to detect intracranial electroencephalographic (iEEG) seizure patterns (iESPs) with expert-level accuracy. We developed an intelligent system for identifying the presence and onset time of iESPs in iEEG recordings from the RNS device.

METHODS

An iEEG dataset from 24 patients (36 293 recordings) recorded by the RNS System was used for training and evaluating a neural network model (iESPnet). The model was trained to identify the probability of seizure onset at each sample point of the iEEG. The reliability of the net was assessed and compared to baseline methods, including detections made by the device. iESPnet performance was measured using balanced accuracy and the F1 score for iESP detection. The prediction time was assessed via both the error and the mean absolute error. The model was evaluated following a hold-one-out strategy, and then validated in a separate cohort of 26 patients from a different medical center.

RESULTS

iESPnet detected the presence of an iESP with a mean accuracy value of 90% and an onset time prediction error of approximately 3.4 s. There was no relationship between electrode location and prediction outcome. Model outputs were well calibrated and unbiased by the RNS detections. Validation on a separate cohort further supported iESPnet applicability in real clinical scenarios. Importantly, RNS device detections were found to be less accurate and delayed in nonresponders; therefore, tools to improve the accuracy of seizure detection are critical for increasing therapeutic efficacy.

SIGNIFICANCE

iESPnet is a reliable and accurate tool with the potential to alleviate the time-consuming manual inspection of iESPs and facilitate the evaluation of therapeutic response in RNS-implanted patients.

Landscape and future directions of machine learning applications in closed-loop brain stimulation

Brain stimulation (BStim) encompasses multiple modalities (e.g., deep brain stimulation, responsive neurostimulation) that utilize electrodes implanted in deep brain structures to treat neurological disorders. Currently, BStim is primarily used to treat movement disorders such as Parkinson's, though indications are expanding to include neuropsychiatric disorders like depression and schizophrenia. Traditional BStim systems are "open-loop" and deliver constant electrical stimulation based on manually-determined parameters. Advancements in BStim have enabled development of "closed-loop" systems that analyze neural biomarkers (e.g., local field potentials in the sub-thalamic nucleus) and adjust electrical modulation in a dynamic, patient-specific, and energy efficient manner. These closed-loop systems enable real-time, context-specific stimulation adjustment to reduce symptom burden. Machine learning (ML) has emerged as a vital component in designing these closed-loop systems as ML models can predict / identify presence of disease symptoms based on neural activity and adaptively learn to modulate stimulation. We queried the US National Library of Medicine PubMed database to understand the role of ML in developing closed-loop BStim systems to treat epilepsy, movement disorders, and neuropsychiatric disorders. Both neural and non-neural network ML algorithms have successfully been leveraged to create closed-loop systems that perform comparably to open-loop systems. For disorders in which the underlying neural pathophysiology is relatively well understood (e.g., Parkinson's, essential tremor), most work has involved refining ML models that can classify neural signals as aberrant or normal. The same is seen for epilepsy, where most current research has focused on identifying optimal ML model design and integrating closed-loop systems into existing devices. For neuropsychiatric disorders, where the underlying pathologic neural circuitry is still being investigated, research is focused on identifying biomarkers (e.g., local field potentials from brain nuclei) that ML models can use to identify onset of symptoms and stratify severity of disease.

Information Entropy Measures for Evaluation of Reliability of Deep Neural Network Results

Deep neural networks (DNN) try to analyze given data, to come up with decisions regarding the inputs. The decision-making process of the DNN model is not entirely transparent. The confidence of the model predictions on new data fed into the network can vary. We address the question of certainty of decision making and adequacy of information capturing by DNN models during this process of decision-making. We introduce a measure called certainty index, which is based on the outputs in the most penultimate layer of DNN. In this approach, we employed iEEG (intracranial electroencephalogram) data to train and test DNN. When arriving at model predictions, the contribution of the entire information content of the input may be important. We explored the relationship between the certainty of DNN predictions and information content of the signal by estimating the sample entropy and using a heatmap of the signal. While it can be assumed that the entire sample must be utilized for arriving at the most appropriate decisions, an evaluation of DNNs from this standpoint has not been reported. We demonstrate that the robustness of the relationship between certainty index with the sample entropy, demonstrated through sample entropy-heatmap correlation, is higher than that with the original signal, indicating that the DNN focuses on information rich regions of the signal to arrive at decisions. Therefore, it can be concluded that the certainty of a decision is related to the DNN's ability to capture the information in the original signal. Our results indicate that, within its limitations, the certainty index can be used as useful tool in estimating the confidence of predictions. The certainty index appears to be related to how effectively DNN heatmaps captured the information content in the signal.

Rating by detection: an artifact detection protocol for rating EEG quality with average event duration

Objective. Quantitative evaluation protocols are critical for the development of algorithms that remove artifacts from real electroencephalography (EEG) optimally. However, visually inspecting the real EEG to select the top-performing artifact removal pipeline is infeasible while hand-crafted EEG data allow assessing artifact removal configurations only in a simulated environment. This study proposes a novel, principled approach for quantitatively evaluating algorithmically corrected EEG without access to ground truth in real-world conditions.Approach. Our offline evaluation protocol uses a detector to score the presence of artifacts. It computes their average duration, which measures the recovered EEG's deviation from the modeled background activity with a single score. As we expect the detector to make generalization errors, we employ a generic and configurable Wiener-based artifact removal method to validate the reliability of our detection protocol.Main results. Quantitative experiments extensively compare many Wiener filters and show their consistent rankings agree with their theoretical assumptions and expectations.Significance. The rating-by-detection protocol with the average event duration measure should be of value for EEG practitioners and developers. After removing artifacts from real EEG, the protocol experimentally shows that reliable comparisons between many artifact filtering configurations are possible despite the missing ground-truth neural signals.

Responsive Neurostimulation for Seizure Control: Current Status and Future Directions

Electrocorticography (ECoG) data are commonly obtained during drug-resistant epilepsy (DRE) workup, in which subdural grids and stereotaxic depth electrodes are placed on the cortex for weeks at a time, with the goal of elucidating seizure origination. ECoG data can also be recorded from neuromodulatory devices, such as responsive neurostimulation (RNS), which involves the placement of electrodes deep in the brain. Of the neuromodulatory devices, RNS is the first to use recorded ECoG data to direct the delivery of electrical stimulation in order to control seizures. In this review, we first introduced the clinical management for epilepsy, and discussed the steps from seizure onset to surgical intervention. We then reviewed studies discussing the emergence and therapeutic mechanism behind RNS, and discussed why RNS may be underperforming despite an improved seizure detection mechanism. We discussed the potential utility of incorporating machine learning techniques to improve seizure detection in RNS, and the necessity to change RNS targets for stimulation, in order to account for the network theory of epilepsy. We concluded by commenting on the current and future status of neuromodulation in managing epilepsy, and the role of predictive algorithms to improve outcomes.

A Data-Efficient Framework for the Identification of Vaginitis Based on Deep Learning

Vaginitis is a gynecological disease affecting the health of millions of women all over the world. The traditional diagnosis of vaginitis is based on manual microscopy, which is time-consuming and tedious. The deep learning method offers a fast and reliable solution for an automatic early diagnosis of vaginitis. However, deep neural networks require massive well-annotated data. Manual annotation of microscopic images is highly cost extensive because it not only is a time-consuming process but also needs highly trained people (doctors, pathologists, or technicians). Most existing active learning approaches are not applicable in microscopic images due to the nature of complex backgrounds and numerous formed elements. To address the problem of high cost of labeling microscopic images, we present a data-efficient framework for the identification of vaginitis based on transfer learning and active learning strategies. The proposed informative sample selection strategy selected the minimal training subset, and then the pretrained convolutional neural network (CNN) was fine-tuned on the selected subset. The experiment results show that the proposed pipeline can save 37.5% annotation cost while maintaining competitive performance. The proposed promising novel framework can significantly save the annotation cost and has the potential of extending widely to other microscopic imaging applications, such as blood microscopic image analysis.

Chronic electroencephalography in epilepsy with a responsive neurostimulation device: current status and future prospects

INTRODUCTION

Implanted neurostimulation devices are gaining traction as therapeutic options for people with certain forms of drug-resistant focal epilepsy. Some of these devices enable chronic electroencephalography (cEEG), which offers views of the dynamics of brain activity in epilepsy over unprecedented time horizons.

AREAS COVERED

This review focuses on clinical insights and basic neuroscience discoveries enabled by analyses of cEEG from an exemplar device, the NeuroPace RNS® System. Applications of RNS cEEG covered here include counting and lateralizing seizures, quantifying medication response, characterizing spells, forecasting seizures, and exploring mechanisms of cognition. Limitations of the RNS System are discussed in the context of next-generation devices in development.

EXPERT OPINION

The wide temporal lens of cEEG helps capture the dynamism of epilepsy, revealing phenomena that cannot be appreciated with short duration recordings. The RNS System is a vanguard device whose diagnostic utility rivals its therapeutic benefits, but emerging minimally invasive devices, including those with subscalp recording electrodes, promise to be more applicable within a broad population of people with epilepsy. Epileptology is on the precipice of a paradigm shift in which cEEG is a standard part of diagnostic evaluations and clinical management is predicated on quantitative observations integrated over long timescales.