Tag Archives: EEG

Early Detection of Seizures

Current Research Advances, , , ,

For decades,  the early detection of a seizure was a serious research effort.  It was obvious that if a seizure could be detected before its initiation, the patient would be better prepared to seek help to suppress the seizure.  However, early detection of a seizure with a lead time of hours proved to be unattainable.  Over the past decade, early detection of a seizure within minutes not hours of its initiation seemed a more reasonable objective.  This blog will focus on the state of research on early detection of seizures with advances developed by the epilepsy laboratory at Johns Hopkins University.

Introduction

As reviewed by Jouny et al., (2011), “The window of opportunity is the time of the earliest detectable changes in the EEG and the onset of disabling clinical symptoms”.  Thus, this provides a definition of early detection.  It may sound straightforward but there are numerous challenges to early detection of seizures. Jouny et al., reviews some of them.

1. Early detection of seizures –  Sensitivity and specificity

Early detection is just that.  Firstly, it must detect the earliest change in the electrical (neuronal) activity that leads to a seizure (called a true positive) and not an electrical event that would not lead to a seizure (called false positive).  Stated another way, the seizure is termed the ictal event and non seizure activity is the interictal activity.  Secondly, early detection must distinguish between lots of interictal activity and a smaller amount of ictal activity.  Many times initially there is little neuronal difference in the two.  Therefore, both sensitivity and specificity are required elements of any measurement targeting early detection of a seizure.

2.  Early detection of seizures – Capacity of microelectrodes.  

Microelectrodes in high density arrays are useful in obtaining detailed data of abnormal neuronal activity (termed pathological high frequency oscillations) indicative of  localized seizures. Specifically, researchers attach microelectrodes to the EEG or intracellular electrodes to capture as much information as possible.  Seizures that spread over extensive areas, unfortunately, surpass the present day recording capabilities of the microelectrode arrays.

3.  Early detection of seizures – Sampling rate.

Higher sampling rates are desirable but this produces more data for analysis.  This requires more powerful computers and better electronics.  Improvements here are needed to detect unique potential changes that distinguish the initiation of the seizure and the seizure itself from all the electrical activity in between.

Current Advances in Early Detection of Seizures.

Early detection devices depend on what is called support vector machines (SVM) which are a type of machine learning algorithms.  “Several algorithms have been developed for early-seizure-onset detection with real-time capabilities for scalp-EEG” (Ehrens et al., 2022).  In particular, they include early detection algorithms worldwide from labs in the Netherlands (Bogaarts et al., 2016); in India (Sridevi et al., 2019); in Italy (Chisci et al., 2010); in Austria/Netherlands (Fürbass et al., 2015); in Germany (Meier et al., 2008; Manzouri et al., 2021); in Canada (Saab and Gotman, 2005) and in the US, (Minasyan et al., 2010, Yidiz et al., 2022, Ehrens et al., 2022). 

Application of early detection algorithms

Early detection devices are appropriately evaluated in an epilepsy monitoring unit (EMU).  The EMU needs to gather reams of data for precise surgical intervention in patients with epilepsy that are resistant to drugs.  The Johns Hopkins Epilepsy Laboratory developed an “dynamical and self-adapting algorithm” (Ehrenss et al., 2022) which detects in real time the seizure phase of early onset as well as the seizure itself.  It is patient specific and does not require prior EEG training.  Basically, the ” algorithm uses a dynamic training set, based on a 20-min dataset that is updating every second” and ” requires no prior training because the algorithm compares the past 20 minutes of activity with the current activity and looks for novelty” (Ehrens et al., 2022)  A processing time of 0.5 seconds allows  algorithm processing, removal of artifacts, and extraction of essential features in the 20 minute learning session.

One-class SVM in early detection of seizures.

This algorithm, referred to as a one-class SVM, was tested in 35 patients with epilepsy (male: female, 62%:38%, 34 years average age) in the Johns Hopkins Hospital EMU.  “Patients were implanted intracranially with a combination of subdural grids, strips and/or depth electrodes to record cortical EEG and stereo EEG signals respectively ” (Ehrens et la., 2022).   For other information on intracranial recordings, check out Blog 10.

As presented above, the algorithm collected EEG data, eliminated non-neural artifacts and identified features essential to seizure identification  This preprocessing trains the one-class SVM and determines novelty.  The output is “post processed” such that it will then identify and signal the detection of an epileptiform event in real time. 

Although multiple SVM configurations were evaluated, the single SVM configuration yielded the best results.  The John Hopkins Laboratory SVM achieved a rate of 87% for true positives and a rate of 1.25% for false positive/hour.  That is, the SVM algorithm detected a seizure correctly 87% of the time and was wrong 1.25% of the time.  The true positive rate increased another 6% points and the false positives declined by half with removal of artifacts post analysis.  The mean detection latency (early detection) was 10.4 seconds. 

Conclusions

There is definitely a need for devices capable of early detection of seizures.  In the EMU, early detection alerts the staff and enables immediate treatment of the patient with epilepsy.  Outside the EMU, an early detection device would dramatically enhance the quality of life of patients with epilepsy.  The algorithm (one class SVM) is moving closer to achieving these goals.  Presently, 100% sensitivity of early detection is achievable in 74% of patients in the EMU setting with the unique algorithm developed at Johns Hopkins Laboratory (Ehrens et al., 2022). 

References (pubmed)

Bogaarts JG et al.,. Improved epileptic seizure detection combining dynamic feature normalization with EEG novelty detection. Med Biol Eng Comput  54:1883–92, 2016.

Chisci L. et al., Real-time epileptic seizure prediction using AR models and support vector machines. IEEE Trans Biomed Eng. 57(5):1124-32, 2010.

Ehrens et al., Dynamic training of a novelty classifier algorithm for real-time detection of early seizure onset.  Clin Neurophysiol.  135: 85–95, 2022.

Furbass F. et al., Prospective multi-center study of an automatic online seizure detection system for epilepsy monitoring units Clin Neurophysiol. 126(6):1124-1131, 2015.

Jouny CC. et al., Improving Early Seizure Detection Epilepsy Behav.  22(Suppl 1): S44–S48, 2011.

Manzouri F et al., A Comparison of Energy-Efficient Seizure Detectors for Implantable Neurostimulation Devices Front Neurol 12:703797, 2022.

Meier R et al., Detecting epileptic seizures in long-term human EEG: a new approach to automatic online and real-time detection and classification of polymorphic seizure patterns J Clin Neurophysiol. 25(3):119-31, 2008.

Minasyan GR et al., -specific early seizure detection from scalp electroencephalogram J Clin Neurophysiol. 27(3):163-78, 2010.

Saab ME, Gotman J. A system to detect the onset of epileptic seizures in scalp EEG Clin Neurophysiol. 116(2):427-42, 2005.

Sridevi V, et al.,. Improved Patient-Independent System for Detection of Electrical Onset of Seizures. J Clin Neurophysiol 36:14–24, 2019.

Yidiz I et al., Unsupervised seizure identification on EEG Comput Methods Programs Biomed. 215:106604, 2022.

Assessment 8 – Nature of Epilepsy

Pathophysiology of epilepsy, ,

Introduction

The electroencephalogram known as the EEG is one of the main clinical tools used by neurologists to capture a seizure so as to make a  diagnosis of epilepsy (Assessment 7).  The EEG is a 20 minute procedure in which electrodes are placed on the scalp and brain wave activity is recorded and analyzed.  It has many flaws.

EEG limitations

The EEG  has serious limitations, chief of which is its short duration. Therefore, the EEG is unable to detect most seizures since most seizures, occur randomly.  In lieu of an EEG seizure, physicians rely on the patient and/or family observations of seizure activity for a diagnosis.  Thus, physicians make a selection of anti-seizure medication without hard evidence of seizure type and brain location.  Although this empirical approach may eventually lead to seizure suppression for the patient, albeit with side effects, it does not advance the field of epilepsy.  Treating the symptoms of epilepsy does not add to understanding the disease in order to find a cure.

Insights from intracranial EEG recordings

Cyclic Nature

Brain wave recordings gathered over a period of years in patients with epilepsy have revealed new and important insights into the nature of epilepsy.  Basically, these findings show that epilepsy creates a brain where abnormal brain waves (interictal epileptiform activity) occur in cycles of 7, 15, 20-30 days or more. 

What this means is that abnormal brain wave activity flows over the brain like ocean waves, rising to a peak and then falling to a baseline.  Cycle length is unique to each patient with epilepsy.  Importantly, these cycles are predictive of a seizure. This is because it is the rising number of abnormal brain waves of a cycle (the upswing of the cycle) that is associated with seizure onset.  Therefore, as the number of abnormal brain waves in each cycle increases to its peak so does the risk of seizures.  The waning number of abnormal brain waves of each cycle is not related to seizure risk. The section below describes the supportive experiments.

These are incredibly significant revelations. Firstly, cyclic brain wave activity exists in epilepsy. Secondly, the upward swing to the peak of each cycle is the time of greatest risk for seizures.  Therefore, cycle intervals are unique for each patient with epilepsy. Each cycle predicts the next round of seizures (since it follows the rise in wave number per cycle).  The knowledge of this relation between cycle and seizure could allow for adequate intervention and prevention.

Source of Insights

These findings appear in a publication by Baud MO et al. Under-sampling in epilepsy: Limitations of conventional EEG. Clinical Neurophysiology Practice 6 : 41–49, 2021.  Since copyright laws prevent reproduction of Baud’s figures, it is recommended that the reader obtain the free publication through PubMed (https://pubmed.ncbi.nlm.nih.gov/).  The representative EEGs with superimposed seizure events provide visual confirmation of these findings.

These findings are important. Two hundred patients with generalized epilepsy of unknown cause (idiopathic) resistant to antiseizure medication participated in Baud’s clinical study.  To suppress their uncontrolled seizures they chose an FDA approved intracranial neurostimulator (RNS_System.NeuroPace, Inc).  The device captures the abnormal brain waves (interictal epileptiform activity) described and can respond with a counter stimulation to neutralize the seizure.  

Sixty percent of the patients in Baud (2021)’s study exhibit cyclic rhythms as described above.  This is certainly a significant number of patients to propose cyclic brain waves rhythms as a basic characteristic of idiopathic generalized epilepsy.  However, in the future, it is important to evaluate patients with different types of epilepsy to determine if these cycles also occur across all epilepsies.

Future Challenges 

The hurdle to more wide spread research is how to obtain years long data without  implantation of electrodes directly into the brain.  Intracranial electrode implantation is major invasive surgery and hence, a choice of last resort.  Since these observations would never have been obtained in a 20 minute diagnostic EEG, there is clearly a need to develop external devices with the same sensitivity and accuracy as intracranial devices that would be suitable for long term use. 

Summary

The exciting findings of Baud et al (2021) uncovered significant novel characteristics in the brains of patients with epilepsy.  Results of years long brain recordings of patients with idiopathic generalized epilepsy show abnormal brain wave activity that continuously cycles over intervals of weeks or months.  As the number of abnormal waves increase so does the risk of seizures.  This is an incredible insight and if confirmed in other types of epilepsy, provides the first step toward understanding the disease of epilepsy.

Assessment 7 – Value of Diagnostics for Epilepsy

Diagnostics, , , ,

Epilepsy is a disease defined by the presence of two or more seizures, separated by more than 24 hours.  However, in order to identify the appropriate anti-seizure drug for seizure suppression, a diagnosis of epilepsy is more complicated than the definition implies.  There are 4 sources of potentially valid information that together should yield the most accurate diagnosis.  The diagnostics for epilepsy are:

a) seizure history provided by the patient;

b) digital data from an electroencephalogram (EEG) which records seizure activity or, in the absence of a seizure, the presence of abnormal discharges;   

c) analysis from magnetic resonance imaging (MRI) showing presence of a tumor, scar tissue, or abnormal anatomical lesions related to site of seizure origin;

d) genetic assessment matching one or more of the known genetic errors associated with specific types of epilepsy.

What is the value of the information gathered from these sources?

1. Seizure History

Seizure history is important but can be lacking in specifics.  This is because many patients have little to no recollection of their seizures.  This starting point for a diagnosis is not particularly helpful in identifying the epilepsy type and thus, it is not helpful at all in selection of the appropriate antiseizure medication.

2.  EEG

The EEG was developed over a century ago.  Electrodes, placed on the scalp, record summated electrical activity primarily from the outer portion of the brain (the cortex).  Electrodes are precisely positioned to receive defined recognizable brain wave activity.  The EEG recording itself generally takes 20 minutes but considerable time is required to position the electrodes and also to remove them.  During the actual recording, procedures such as light stimulation or hyperventilation are used to “evoke” a seizure.

Advantages of diagnostics for epilepsy: 

Whereas the EEG usually does not record a seizure in progress, its ability to detect abnormal discharges, (called interictal epileptogenic discharges or IEDs), is significant because the presence of IEDs have been found to signal a high risk of having a future seizure.  The EEG many record local IEDs (expressed in one hemisphere) or general IEDs (expressed in the entire brain). This information contributes to a more accurate diagnosis and hence is one clear benefit of the EEG.  Additionally, the EEG is helpful in monitoring the progress of seizure suppression with drugs or surgery. Anti-seizure drugs has previously been discussed (Assessment 6 – Epilepsy Medication – Need to Know Information)

Limitations of Diagnostics for Epilepsy: 

As revealed by years of intracranial recording (surgical implantation of electrodes) in patients with epilepsy, it is apparent that the EEG has limitations.  Its short duration, lack of sensitivity and tendency for over interpretation are some of its limitations.  It has also failed to identify the multiday cyclic nature of epileptic activity. 

The short duration of the EEG test fails to capture the majority of  in-progress seizures and many patients show no evidence of IEDs.  Although in a hospital setting, the EEG may be used continually for several days, this is not routinely practiced.  To avoid obtaining incorrect data, technical expertise is required for electrode placement.  Despite this, sensitivity is diminished due to distance of the electrodes (on the scalp) from the origin of the brain activity (somewhere in the brain) and the inability of electrodes to record all brain wave activity but only a sampling of those reaching the outer layer of the brain.

3.  MRI

The MRI is one of several imaging techniques with success in identifying abnormalities in the brain in persons with epilepsy (see https://medlineplus.gov/mriscans.html ) for details on MRI imaging.  In cases of epilepsies resistant to drug, surgical protocols require prior imaging procedures.

Imaging techniques include the MRI, the positron emission tomography (PET) and MRI spectrometry, to name a few.  The MRI uses magnetic energy to define the location of structures in the brain.  The stronger the magnet, the better the resolution of the individual structures in the brain.  The PET scan adds to the MRI with the use of a radioactive tracer.  This tracer localizes to areas of high metabolism as in tumor or indicates areas of low metabolism as in epileptic lesions.  The MRI spectrometry is a supplemental test only.  It measures brain compounds known to indicate the presence of a seizure lesion.  Abnormal levels of specific compounds add to MRI data of structural changes, pinpointing a epileptic lesion.

Advantages of Diagnostics for Epilepsy: 

The MRI in association with other imaging tools (PET, Spectrometry) provide essential information for epilepsy patients resistant to drug therapy.  These patients may be candidates for surgical therapy.  High quality data from imaging techniques assist with accurate surgery but only facilities with extensive experience in imaging and treating patients whose epilepsy is not suppressed with antiseizure drugs, produce these desired outcomes.

Limitations of Diagnostics for Epilepsy: 

Outpatient use of the MRI technology generally contributes little to the initial diagnosis of epilepsy.  Most clinics lack expertise on recognition of epileptogenic lesions and additionally do not employ the optimal MRI protocols to capture these lesions.  Technologists may also lack the clinical information to alert them to the reason for the MRI and hence fail to seek expert help.  

4.  Genetic Analysis

Epilepsy is considered to have a heritable component.  Identification of the genes involved in epilepsy should provide greater understanding of this disease and lead to better therapy.  The International League Against Epilepsy Consortium on Complex Epilepsies initiated a large analysis (meta-analysis) of the genome (DNA content) of persons with epilepsy compared to controls.  The most recent work (2018) included over 15,000 epilepsy cases and over 29,000 controls.  This study identified 16 different chromosome locations that relate with confidence to epilepsy.  These chromosome positions reveal many genes already associated with epilepsy such as the ion channels on nerve cells but also importantly, many new genes not previously identified as a cause of epilepsy.

Advantages of Diagnostics for Epilepsy: 

Routine genomic testing in epilepsy could be a step closer to precision medicine. Genomic testing would allow for a precise diagnosis, and a more rational selection of anti-seizure medication.  This type of testing opens the door for discovery of novel drugs and has potential to determine exactly which genes play a role in specific types of epilepsy.

Limitations of Diagnostics for Epilepsy: 

Analysis of genomic-wide testing is complicated and the variety of different forms of epilepsy are equally complex.  To date, most of the identified genes are “associated with genetic generalized epilepsy” (ILAE Consortium).  Considerable research commitment will be needed to unravel the genetic influence on epilepsy and then to reveal how the environment interacts with these genes.

Summary

There exists 4 different types of information (patient assessment, EEG, MRI, genetic analysis) that should provide an accurate diagnosis of epilepsy.  Unfortunately, their limitations outweigh their assets.  Thus, a diagnosis of epilepsy may be absent, delayed or incorrect and consequently, antiseizure therapy may be inappropriate.

Myths and Misconceptions

Myths/Misconceptions, , ,

Assessment 2 – Objective seizure detection with biosensors

My first assessment for epilepsy evaluation concluded that persons with epilepsy (PWE) unfortunately have little to no ability to reproducibly predict the onset of their own seizures.  Therefore, subjective impressions are of marginal value.  Given that, there is a vital need for devices or instruments with ability to objectively predict a seizure. 

Many devices called biosensors are already available  (see reviews Nagaraj et al., 2015; Ulate-Campos et al., 2016).  The capability of existing biosensors to predict seizures varies with the specific device.  Unfortunately detection and prevention are generally not part of the same device.  Despite the fact that current biosensors (with possibly one exception) are only able to detect seizures, nevertheless, they represent a significant first step to development of something more sophisticated with therapeutic value.  Furthermore, present day biosensors supply information on seizure frequency, important information that PWE are unable to provide due to an inability to recall the seizure (Hoppe et al., 2015).  As a note of caution, biosensors to date are useful only in certain types of seizures and so to be of benefit, appropriate biosensor selection requires accurate seizure type identification (Ulate-Campos et al., 2016).  

How biosensors work

There are numerous biosensors in use that detect seizures but generally they do not as yet predict seizures within a reasonable time frame for adequate intervention.  Some biosensors record muscle movements e.g. surface electromyography (Conradsen et al., 2012), measure change in muscle acceleration called accelerometry  (Beniczky et al., 2013), or measure movement via mattress sensors (Poppel et al., 2013).  Other biosensors record effects produced by activation of the autonomic nervous system.  This includes measurement of sweating (electrodermal activity, EDA), heart rate (electrocardiogram, EKG), respiration, body temperature and combinations of these with sophisticated computer programs (Ulate-Campos et al., 2016). 

The aforementioned biosensors target seizures in which muscle and autonomic nervous system changes are evident and occur before a seizure.  Other biosensors use near-infrared to detect changes in brain blood flow prior to a seizure (Tewolde et al., 2015).  Ulate-Campos et al., 2016 lists the more than 40 available biosensors and their web sites. 

EEG and intracranial electrodes  – important contributions

However, for all seizures, the gold-standard for seizure detection is the electroencephalogram (EEG).  The EEG records the summation of brain wave activity via topically attached electrodes positioned around the scalp.  Unfortunately, everyday use of the EEG is incompatible with daily life. 

Representative EEG brain wave activity

Another approach, intracranial electrode implantation (iEEG), provides continuous long term measurement of brain activity in the seizure prone region.  Results of small studies and clinical trials with the iEEG (Iasemidis et al., 2003; Cook et al., 2013; Spencer et al., 2016; Karoly et al.,2018; Baud et al., 2018) have confirmed the  presence of

a) unique neurological activity termed epileptiform discharges that occur with a regularity of 24 hours (circadian) or over several days (multi-day), and

b) the relation of aspects of these rhythms (phases) with seizures in PWE with low to moderate seizure frequency rate (Baud et al., 2018).

NeuroVista Advisory System is an example of the iEEG that has undergone clinical evaluation. NeuroVista Advisory System employs an assessment system using sophisticated algorithms.  In clinical trials (Cook et al., 2013; Bergey et al., 2015) it performed with reasonable success.  In several cases where seizure prediction fell below pre-designated criteria, optimization of the computer system i.e. its algorithms, dramatically improved predictability of the seizure.  Thus it appears possible to tailor the algorithm to each patient to achieve the best result (Kuhlmann et al., 2018). 

Although the iEEG successfully analyzes prodrome data (see Assessment 1) to predict a seizure and hence supplies information in a sufficient time frame to avert a seizure, it is not without serious concerns.  An implantation of intracranial electrodes is an invasive technique requiring major surgery and carries the risk of infection and related brain problems.

Biofeedback Biosensors – hope for the future

The most desirable biosensor is one that employs a biofeedback loop with both

1) reliable seizure detection and

2) appropriate and effective therapeutic intervention that prevents the seizure. 

Neuropace Responsive Neurostimulation (RNS) (http://www.neuropace.com)  device is an example of a biosensor with a biofeedback loop.  The device is implanted in the brain and detects epileptiform discharges and delivers neurostimulation to suppress the seizure.  Results of a two year clinical trial with 191 medication- resistant PWE, showed significant decline in seizure frequency (Heck et al., 2014).  The PWE enrolled in this trial had intractable partial-onset epilepsy.  The development of the Neuropace RNS is a significant achievement for select PWE.  However, the biggest limitation is the requirement of major surgery for implantation of this device.

Characteristics of the ideal biosensor

The ideal biosensor would monitor seizure-related external biological phenomena such as EKG, EDA, muscle movement blood flow etc., as discussed above and accurately assess the data to predict a seizure.  Additionally, seizure detection would be linked with effective seizure prevention measures.  This could take the form of a) an alert  to a caretaker to give appropriate therapy, b)  neurostimulation or c) minipump infusion of appropriate mediation (Ulate-Campos et al., 2016).  Additional characteristics of the ideal biosensor include safety and efficacy without false alarms, user friendly instrumentation, ease of care, wireless transmission of data and full compatibility with daily life (Hoppe et al., 2015).

Assessment-2-referencesDownload

Questions for readers

Comments on this blog are welcome.  Experience with biosensors would be appreciated.