No full-text available
To read the full-text of this research,
you can request a copy directly from the author.
Epileptic seizure prediction and control
Abstract
Epileptic seizures are manifestations of epilepsy, a serious brain dynamical disorder second only to strokes. Of the world's ∼50 million people with epilepsy, fully 1/3 have seizures that are not controlled by anti-convulsant medication. The field of seizure prediction, in which engineering technologies are used to decode brain signals and search for precursors of impending epileptic seizures, holds great promise to elucidate the dynamical mechanisms underlying the disorder, as well as to enable implantable devices to intervene in time to treat epilepsy. There is currently an explosion of interest in this field in academic centers and medical industry with clinical trials underway to test potential prediction and intervention methodology and devices for Food and Drug Administration (FDA) approval. This invited paper presents an overview of the application of signal processing methodologies based upon the theory of nonlinear dynamics to the problem of seizure prediction. Broader application of these developments to a variety of systems requiring monitoring, forecasting and control is a natural outgrowth of this field.
... B.), Conselho Nacional de Desenvolvimento of research. From the neurophysiological point of view, it is associated to the progressive hypersynchrony of neural firing over time, generated by endogenous or exogenous agents [1,6]. However, the precise triggering mechanisms that disrupt a healthy brain leading to a seizure are not sufficiently understood yet [7]. ...
... Epilepsy affects millions of people worldwide [1,2], leading to several implications that reduce patients' quality of life [3][4][5], which defines this disorder as a very sensitive and relevant topic reconstructing the signal over time and then applying inputs to attenuate the undesired activity, respectively. These tools are commonly found in control engineering. ...
... Note that this model avoids terms of higher order, thus reducing the magnitude of the elements in the continuous-time matrix. It can be further adapted by considering a scaling factor such that A fcg ¼ 1 g 1 � A fcg , where γ 1 6 ¼ 0. This is particularly interesting for some reasons: (1) local or quasi-non-stationary behavior of the signals leads to higher AR coefficients in the discrete model [30], which in turn is reflected on the continuous one; thus, scaling down the models is proposed in this work to mitigate this problem; (2) this problem is endorsed by the sampling frequency f s of the signals, which is considerably high even after downsampling (1kHz) due to the time scale of neuron firing being relatively fast; (3) since the proposed approach focus on AR models of low order, local fluctuations may not be captured by them, which means that the difference between the dynamics of the recording being read and the one reconstructed by the identified AR model may not be compensated for by the gains of the observer. ...
Epilepsy affects millions of people worldwide every year and remains an open subject for research. Current development on this field has focused on obtaining computational models to better understand its triggering mechanisms, attain realistic descriptions and study seizure suppression. Controllers have been successfully applied to mitigate epileptiform activity in dynamic models written in state-space notation, whose applicability is, however, restricted to signatures that are accurately described by them. Alternatively, autoregressive modeling (AR), a typical data-driven tool related to system identification (SI), can be directly applied to signals to generate more realistic models, and since it is inherently convertible into state-space representation, it can thus be used for the artificial reconstruction and attenuation of seizures as well. Considering this, the first objective of this work is to propose an SI approach using AR models to describe real epileptiform activity. The second objective is to provide a strategy for reconstructing and mitigating such activity artificially, considering non-hybrid and hybrid controllers − designed from ictal and interictal events, respectively. The results show that AR models of relatively low order represent epileptiform activities fairly well and both controllers are effective in attenuating the undesired activity while simultaneously driving the signal to an interictal condition. These findings may lead to customized models based on each signal, brain region or patient, from which it is possible to better define shape, frequency and duration of external stimuli that are necessary to attenuate seizures.
... Literature shows evidence of the preictal interval firstly by observing changes in the electroencephalography (EEG) recordings seconds to hours before the seizure onset and, additionally, by the reported predictability of seizures 4,9 . Starting in the early 1990s, with the application of the mathematical theory of nonlinear dynamics, the preictal interval became associated with the state during which the brain activity evolves deterministically towards the seizure 7,10 . In other words, once the brain enters this state, a "point of no return" has been passed, meaning that the seizure will occur 6,7,11 . ...
... We performed hyperparameter tuning for UMAP (see Fig. 2). Namely, we applied UMAP considering different values of nearest neighbours (ten values in the range of [10,100]) and minimum distance (nine values in the range of [0.1, 0.9]). In the next section, we elaborate on finding the best parameters. ...
... The results obtained with nonlinear feature reduction methods such as UMAP may be more suitable to reveal the nonlinear dynamical functioning of the brain. Accordingly, most studies propose nonlinear systems for epilepsy EEG modelling 10,12 . Nevertheless, we are aware that using this nonlinear method and the consequent parameter tuning could significantly impact further data interpretations 58 . ...
Typical seizure prediction models aim at discriminating interictal brain activity from pre-seizure electrographic patterns. Given the lack of a preictal clinical definition, a fixed interval is widely used to develop these models. Recent studies reporting preictal interval selection among a range of fixed intervals show inter- and intra-patient preictal interval variability, possibly reflecting the heterogeneity of the seizure generation process. Obtaining accurate labels of the preictal interval can be used to train supervised prediction models and, hence, avoid setting a fixed preictal interval for all seizures within the same patient. Unsupervised learning methods hold great promise for exploring preictal alterations on a seizure-specific scale. Multivariate and univariate linear and nonlinear features were extracted from scalp electroencephalography (EEG) signals collected from 41 patients with drug-resistant epilepsy undergoing presurgical monitoring. Nonlinear dimensionality reduction was performed for each group of features and each of the 226 seizures. We applied different clustering methods in searching for preictal clusters located until 2 h before the seizure onset. We identified preictal patterns in 90% of patients and 51% of the visually inspected seizures. The preictal clusters manifested a seizure-specific profile with varying duration (22.9 ± 21.0 min) and starting time before seizure onset (47.6 ± 27.3 min). Searching for preictal patterns on the EEG trace using unsupervised methods showed that it is possible to identify seizure-specific preictal signatures for some patients and some seizures within the same patient.
... Epilepsy is one of the most common brain disorders worldwide, affecting millions of people every year (Iasemidis, 2003;Dua, De Boer, Prilipko, & Saxena, 2006). It is usually associated with a hypersynchronization of neurons (Iasemidis, 2003;Fisher, Harding, Erba, Barkley, & Wilkins, 2005) caused by endogenous agents, such as the result of genetic abnormalities, traumas, or nervous system infection (Iasemidis, 2003), for example, or exogeneous ones, such as visual stimuli modulated in specific frequencies (Zhu, Bieger, Molina, & Aarts, 2010). ...
... Epilepsy is one of the most common brain disorders worldwide, affecting millions of people every year (Iasemidis, 2003;Dua, De Boer, Prilipko, & Saxena, 2006). It is usually associated with a hypersynchronization of neurons (Iasemidis, 2003;Fisher, Harding, Erba, Barkley, & Wilkins, 2005) caused by endogenous agents, such as the result of genetic abnormalities, traumas, or nervous system infection (Iasemidis, 2003), for example, or exogeneous ones, such as visual stimuli modulated in specific frequencies (Zhu, Bieger, Molina, & Aarts, 2010). Because it may lead to several negative physical, psychological, and social implications (Fisher & Schachter, 2010), thus considerably decreasing patients' quality of life, possible therapies that aim at seizure mitigation or control are available now, ranging from pharmacological, surgical, or electrical stimulation procedures, to specific types of diet (Iasemidis, 2003;D'Andrea Meira et al., 2019). ...
... Epilepsy is one of the most common brain disorders worldwide, affecting millions of people every year (Iasemidis, 2003;Dua, De Boer, Prilipko, & Saxena, 2006). It is usually associated with a hypersynchronization of neurons (Iasemidis, 2003;Fisher, Harding, Erba, Barkley, & Wilkins, 2005) caused by endogenous agents, such as the result of genetic abnormalities, traumas, or nervous system infection (Iasemidis, 2003), for example, or exogeneous ones, such as visual stimuli modulated in specific frequencies (Zhu, Bieger, Molina, & Aarts, 2010). Because it may lead to several negative physical, psychological, and social implications (Fisher & Schachter, 2010), thus considerably decreasing patients' quality of life, possible therapies that aim at seizure mitigation or control are available now, ranging from pharmacological, surgical, or electrical stimulation procedures, to specific types of diet (Iasemidis, 2003;D'Andrea Meira et al., 2019). ...
Epilepsy is one of the most common brain disorders worldwide, affecting millions of people every year. Although significant effort has been put into better understanding it and mitigating its effects, the conventional treatments are not fully effective. Advances in computational neuroscience, using mathematical dynamic models that represent brain activities at different scales, have enabled addressing epilepsy from a more theoretical standpoint. In particular, the recently proposed Epileptor model stands out among these models, because it represents well the main features of seizures, and the results from its simulations have been consistent with experimental observations. In addition, there has been an increasing interest in designing control techniques for Epileptor that might lead to possible realistic feedback controllers in the future. However, such approaches rely on knowing all of the states of the model, which is not the case in practice. The work explored in this letter aims to develop a state observer to estimate Epileptor's unmeasurable variables, as well as reconstruct the respective so-called bursters. Furthermore, an alternative modeling is presented for enhancing the convergence speed of an observer. The results show that the proposed approach is efficient under two main conditions: when the brain is undergoing a seizure and when a transition from the healthy to the epileptiform activity occurs.
... Epileptic seizure in recurrent is a neurological disorder of epilepsy patients [1], [2]. Seizures occur due to excessive electrical impulses inside the brain. ...
... The recording of these impulses is called an electroencephalogram (EEG). The seizure is usually detected by analysis of EEG signals [1], [2], [3]. The occurrences of seizure events are violent shaking, loss of control, and loss of consciousness. ...
Epileptic seizures are unpredictable events due to sudden abnormal electrical activities in the brain of epilepsy patients. A seizure can be predicted by analyzing the EEG signals to prevent unwanted life risks. The goal of this paper is to implement a method that will apply to design a lightweight, wearable, and efficient seizure prediction device. The proposed method will satisfy two objectives. The first objective is relevant feature extraction for the classification of EEG signals with excellent accuracy. The second objective is the use of fewer EEG channels. In this paper, one 1D-CNN is applied for feature extraction and classification of raw EEG signals for early prediction of seizure events. The 1D-CNN is faster compared to 2D-CNN, which uses fewer trainable parameters. Hence, it is suitable to implement a low-power energy-efficient seizure prediction device. In this paper, the NSGA-II algorithm is applied to get the optimum set of EEG channels for seizure prediction. The NSGA-II algorithm identifies a set of three EEG channels from twenty-two channels as the optimum channel set. The proposed method optimizes the EEG channels from 22 to 3, i.e., 86.36% channel reduction. It provides the classification accuracy, sensitivity, and specificity of 0.9651, 0.9655, and 0.9647, respectively. The proposed method is better than the state-of-the-art works under the condition of using three channels. The proposed method provides excellent performance using only three EEG channels, which will be applicable to design a lightweight, low-power, and wearable seizure prediction device.
... This first seizure prediction algorithm was automated and adaptive, the precursor of the current so-called event-based models, needed to detect the occurrence of the first seizure for initialization of its parameters per patient and did not need any predetermined (i.e., arbitrary chosen) preictal (and hence neither SOP nor SPH) period. Seizure prediction became a whole new field in brain research thereafter, and a hot topic due to its many potential applications in the diagnosis, prognosis and treatment of epilepsy and potentially of other brain dynamical disorders (3,(19)(20)(21). With the development and application of Artificial Intelligence (AI) technology for clinical and diagnostic purposes, the use of Machine Learning (ML) and Deep Learning (DL) in constructing models for seizure prediction based on EEG features has become a popular method in epilepsy management ( Figure 1) (16-18, 22, 23). ...
... Kuhlmann et al. (26) provided a detailed overview of the field of seizure prediction, indicating its future directions. Others have outlined particular areas from previous studies, such as features, methods (21,(27)(28)(29), model selection, and Brain-Computer Interfaces (BCI) (3,20,30,31), etc. However, there are barely any review studies specifically focusing on the AIguided construction of "event-based" models for predicting seizures from EEG findings, and the importance of postprocessing techniques, SOP and SPH. ...
The recurrent and unpredictable nature of seizures can lead to unintentional injuries and even death. The rapid development of electroencephalogram (EEG) and Artificial Intelligence (AI) technologies has made it possible to predict seizures in real-time through brain-machine interfaces (BCI), allowing advanced intervention. To date, there is still much room for improvement in predictive seizure models constructed by EEG using machine learning (ML) and deep learning (DL). But, the most critical issue is how to improve the performance and generalization of the model, which involves some confusing conceptual and methodological issues. This review focuses on analyzing several factors affecting the performance of seizure prediction models, focusing on the aspects of post-processing, seizure occurrence period (SOP), seizure prediction horizon (SPH), and algorithms. Furthermore, this study presents some new directions and suggestions for building high-performance prediction models in the future. We aimed to clarify the concept for future research in related fields and improve the performance of prediction models to provide a theoretical basis for future applications of wearable seizure detection devices.
... Scalp EEG (sEEG) has been widely used in epilepsy prediction due to its ease of acquisition, low cost, and high temporal resolution [6], [7]. Most of the differences in seizure prediction algorithms based on sEEG are visible in two main steps, namely feature extraction and pre-ictal classification against interictal categories. ...
... The suddenness and individual variability make it difficult to perform accurate seizure prediction. sEEG has been widely implemented as the most powerful diagnostic and analytical tool for epilepsy [7]. There generally classify EEG data from epileptic patients into four phases: pre-ictal, seizure, post-ictal, and interictal (referring to the interictal period in addition to the previously mentioned states). ...
Infantile spasms (IS) is a typical childhood epileptic disorder with generalized seizures. The sudden, frequent and complex characteristics of infantile spasms are the main causes of sudden death, severe comorbidities and other adverse consequences. Effective prediction is highly critical to infantile spasms subjects, but few related studies have been done in the past. To address this, this study proposes a seizure prediction framework for infantile spasms by combining the statistical analysis and deep learning model. The analysis is conducted on dividing the continuous scalp electroencephalograms (sEEG) into 5 phases: Interictal, Preictal, Seizure Prediction Horizon (SPH), Seizure, and Postictal. The brain network of Phase-Locking Value (PLV) of 5 typical brain rhythms is constructed, and the mechanism of epileptic changes is analyzed by statistical methods. It is found that 1) the connections between the prefrontal, occipital, and central regions show a large variability at each stage of seizure transition, and 2) 4 sub-bands of brain rhythms (θ, α, β, γ) are predominant. Group and individual variabilities are validated by using the Resnet18 deep model on data from 25 patients with infantile spasms, where the consistent results to statistical analyses can be observed. The optimized model achieves an average of 79.78 %, 94.46%, 75.46% accuracy, specificity, and recall rate, respectively. The method accomplishes the analysis of the synergy between infantile spasms mechanism, model, data and algorithm, providing a guideline to build an intelligent and systematic model for comprehensive IS seizure prediction.
... The significantly high TIR and AIR of seizure EEGs are in line with epileptic pathological brain states. Pathologically, epilepsy manifests as recurrent epileptic seizures caused by sudden development of synchronous neuronal firing in cerebral cortex [56,57,58,59,60]. During the onset of a seizure, a group of brain cells exhibits ictal activity, i.e., abnormally excessive electrical discharges. ...
... Therefore, joint permutation TIR and AIR correctly and effectively detects the significantly abnormal nonequilibrium features that reflect epileptic seizure activities. After onset of an epileptic discharge, partial seizures might remain localized [59,60]. Although exact time intervals between the ictal recordings in set E and postictal EEGs in sets C and D are not clear, the characteristics of EEGs in sets C and D indicate the pathological seizure impacts [9,50,58]. ...
Although time irreversibility (TIR) and amplitude irreversibility (AIR) are relevant concepts for nonequilibrium analysis, their association has received little attention. This paper conducts a systematic comparative analysis of the relationship between TIR and AIR based on statistical descriptions and numerical simulations. To simplify the quantification of TIR and AIR, the amplitude permutation and global information of the associated vector are combined to produce a joint probability estimation. Chaotic logistic, Henon, and Lorenz series are generated to evaluate TIR and AIR according to surrogate theory, and the distributions of joint permutations for these model series are measured to compare the degree of TIR and AIR. The joint permutation TIR and AIR are then used to investigate nonequilibrium pathological features in epileptic electroencephalography data. Test results suggest that epileptic brain electrical activities, particular those during the onset of seizure, manifest higher nonequilibrium characteristics. According to the statistical definitions and targeted pairs of joint permutations in the chaotic model data, TIR and AIR are fundamentally different nonequilibrium descriptors from time- and amplitude-reversibility, respectively, and thus require different forms of numerical analysis. At the same time, TIR and AIR both provide measures for fluctuation theorems associated with nonequilibrium processes, and have similar probabilistic differences in the pairs of joint permutations and consistent outcomes when used to analyze both the model series and real-world signals. Overall, comparative analysis of TIR and AIR contributes to our understanding of nonequilibrium features and broadens the scope of quantitative nonequilibrium measures. Additionally, the construction of joint permutations contributes to the development of symbolic time series analysis.
... The significantly high TIR and AIR of seizure EEGs are in line with epileptic pathological brain states. Pathologically, epilepsy manifests as recurrent epileptic seizures caused by sudden development of synchronous neuronal firing in cerebral cortex [56][57][58][59][60]. During the onset of a seizure, a group of brain cells exhibits ictal activity, i.e., abnormally excessive electrical discharges. ...
... After onset of an epileptic discharge, partial seizures might remain localized [59,60]. Although exact time intervals between the ictal recordings in set E and postictal EEGs in sets C and D are not clear, the characteristics of EEGs in sets C and D indicate the pathological seizure impacts [9,50,58]. ...
Although both time irreversibility (TIR) and amplitude irreversibility (AIR) are relevant concepts for nonequilibrium analysis, their association has received little attention. This paper describes a systematic comparative analysis of the relationship between TIR and AIR based on a statistical description and numerical simulations. To simplify the quantification of TIR and AIR, the amplitude permutation and global information of the associated vector are combined to produce a joint probability estimation. Chaotic logistic, Henon, and Lorenz series are generated to evaluate TIR and AIR according to surrogate theory, and the distributions of the joint permutations for these model series are measured to compare the degree of TIR and AIR. The joint permutation TIR and AIR are then used to investigate nonequilibrium pathological features in epileptic electroencephalography data. The results suggest that the epileptic brain electrical activities, particular those during the onset of a seizure, manifest higher nonequilibrium characteristics. According to the statistical definitions and targeted pairs of joint permutations in the chaotic model data, TIR and AIR are fundamentally different nonequilibrium descriptors from time- and amplitude-reversibility, respectively, and thus require different forms of numerical analysis. Both TIR and AIR provide measures for fluctuation theorems associated with nonequilibrium processes, and have very similar probabilistic differences in the pairs of joint permutations and consistent outcomes when used to analyze both the model series and real-world signals. Overall, the comparative analysis of TIR and AIR contributes to our understanding of nonequilibrium features and broadens the scope of quantitative nonequilibrium measures. Additionally, the construction of joint permutations contributes to the development of symbolic time series analysis.
... Epilepsy is one of the most severe, intriguing brain disorders worldwide, whose causes, despite possibly linked to genetic abnormalities, centrous nervous system infection, traumas Iasemidis (2003) or even external stimuli Zhu et al. (2010), to name a few, are not fully understood yet Vezzani et al. (2011). Some of its consequences are: physical implications (increase/decrease of muscle tone, or contraction) Browne and Holmes (2008), physiological implications (headaches, mental confusion or unconsciousness), psychological or psychiatric disorders (as higher incidence of depression and psychosis), and social issues (social dislocation, embarrassment, and isolation) Fisher and Schachter (2000). ...
... Although there are many possible treatments for epilepsy nowadays, such as pharmacological, surgical, based on electrical stimulation Iasemidis (2003), neurofeedback Walker and Kozlowski (2005) and ketogenic diets D' Andrea et al. (2019), they, in general, achieve only partially successful results. Understanding the possible epileptogenic mechanisms may, thus, help to clarify the triggering factors of seizures and target the best available treatments. ...
Epilepsy is one of the most common brain disorders worldwide, afecting millions of people every year. Given the partially
successful existing treatments for epileptiform activity suppression, dynamic mathematical models have been proposed with
the purpose of better understanding the factors that might trigger an epileptic seizure and how to mitigate it, among which
Epileptor stands out, due to its relative simplicity and consistency with experimental observations. Recent studies using this
model have provided evidence that establishing a feedback-based control approach is possible. However, for this strategy to
work properly, Epileptor’s parameters, which describe the dynamic characteristics of a seizure, must be known beforehand.
Therefore, this work proposes a methodology for estimating such parameters based on a successive optimization technique.
The results show that it is feasible to approximate their values as they converge to reference values based on diferent initial
conditions, which are modeled by an uncertainty factor or noise addition. Also, interictal (healthy) and ictal (ongoing seizure)
conditions, as well as time resolution, must be taken into account for an appropriate estimation. At last, integrating such a
parameter estimation approach with observers and controllers for purposes of seizure suppression is carried out, which might
provide an interesting alternative for seizure suppression in practice in the future.
... One of the nonepileptic seizure is pseudo epileptic seizure. Thus, in the following section, it is summarized the pseudo and epileptic seizure characteristics, and their classifications methods [9,10]. ...
... This causes the delay in the treatment to the patients. Thus, researcher deals with decision making systems by using different feature extraction methods to give support to the clinical cases [9][10][11]. ...
... Fig. 1 shows multiple states of epileptic seizures in scalp EEG signal. Researchers (Iasemidis, 2003;Yu et al., 2020;Nasseri et al., 2020;Das et al., 2020;Ozcan and Erturk, 2019;Tsipouras, 2019;Usman and Hassan, 2018;Sudalaimani et al., 2019;Quintero-Rincón et al., 2018;Usman et al., 2019;Zhou et al., 2018;Ibrahim et al., 2018;Usman et al., 2021a;Zhang et al., 2018;Yavuz et al., 2018;Li et al., 2017;Usman et al., 2020;Swami et al., 2016;Usman et al., 2019;Winterhalder et al., 2003;Usman et al., 2021b;Carney et al., 2011;Acharya et al., 2018a;Ramgopal et al., 2014;DuBois et al., 2010;Bandarabadi et al., 2015;Usman et al., 2017) have devised different techniques for classification of preictal and interictal states of EEG signals. All these methods include some common steps like preprocessing for noise removal from EEG signals and conversion of time to frequency domain, feature extraction/selection and classification using machine learning or deep learning methods. ...
... Analysis of EEG signals has been performed by many researchers for detection and prediction of epileptic seizures for more than two decades. Seizure detection (Kumar et al., 2014) involves automated detection of onset of a seizure, whereas, seizure prediction (Iasemidis, 2003) involves the detection of start of preictal class so that subsequent seizure may be prevented before it occurs. It is evident from EEG signal recordings of an epilepsy affected patient that there is a clear difference between ictal state and other states. ...
Objective
Epilepsy affected patient experiences more than one frequency seizures which can not be treated with medication or surgical procedures in 30% of the cases. Therefore, an early prediction of these seizures is inevitable for these cases to control them with therapeutic interventions.
Methods
In recent years, researchers have proposed multiple deep learning based methods for detection of preictal state in electroencephalogram (EEG) signals, however, accurate detection of start of preictal state remains a challenge. We propose a novel ensemble classifier based method that gets the comprehensive feature set as input and combines three different classifiers to detect the preictal state.
Results
We have applied the proposed method on the publicly available scalp EEG dataset CHBMIT of 22 subjects. An average accuracy of 94.31% with sensitivity and specificity of 94.73% and 93.72% respectively has been achieved with the method proposed in this study.
Conclusions
Proposed study utilizes the preprocessing techniques for noise removal, combines deep learning based and handcrafted features and an ensemble classifier for detection of start of preictal state. Proposed method gives better results in terms of accuracy, sensitivity, and specificity.
... Despite decades of research on automatic seizure detection and forecasting [12][13][14], the latter task turns out to be extremely challenging [15]. Nevertheless, inspired by the successes of artificial intelligence (AI) in clinical diagnosis [16] and disease forecasting [17], consistent research efforts are being made to tackle the seizure prediction problem using advanced deep learning techniques [18][19][20][21]. ...
The recent scientific literature abounds in proposals of seizure forecasting methods that exploit machine learning to automatically analyze electroencephalogram (EEG) signals. Deep learning algorithms seem to achieve a particularly remarkable performance, suggesting that the implementation of clinical devices for seizure prediction might be within reach. However, most of the research evaluated the robustness of automatic forecasting methods through randomized cross-validation techniques, while clinical applications require much more stringent validation based on patient-independent testing. In this study, we show that automatic seizure forecasting can be performed, to some extent, even on independent patients who have never been seen during the training phase, thanks to the implementation of a simple calibration pipeline that can fine-tune deep learning models, even on a single epileptic event recorded from a new patient. We evaluate our calibration procedure using two datasets containing EEG signals recorded from a large cohort of epileptic subjects, demonstrating that the forecast accuracy of deep learning methods can increase on average by more than 20%, and that performance improves systematically in all independent patients. We further show that our calibration procedure works best for deep learning models, but can also be successfully applied to machine learning algorithms based on engineered signal features. Although our method still requires at least one epileptic event per patient to calibrate the forecasting model, we conclude that focusing on realistic validation methods allows to more reliably compare different machine learning approaches for seizure prediction, enabling the implementation of robust and effective forecasting systems that can be used in daily healthcare practice.
... According to an estimation of the World Health Organization, more than 50 million of population are affected by epilepsy [2,3]. Approximately, almost 1 % population have the neurological disorders [4][5][6]. It leads to numerous research works to identify epilepsy and related treatments. ...
... Many recent studies have been published using ML methods with EEG signals. The relevant studies are as follows: In the study by Iasemidis, a continuous and long-term adaptable procedure was identified to analyze EEG records only when the first seizure occurred [5]. EEG signals are processed using AR analysis methods and applied to the ANN [6]. ...
Epilepsy is a neurological disorder in which involuntary contractions, sensory abnormalities, and changes occur as a result of abrupt and uncontrolled discharges in the neurons in the brain, which disrupt the systems regulated by the brain. In epilepsy, abnormal electrical impulses from cells in various brain areas are noticed. The accurate interpretation of these electrical impulses is critical in the illness diagnosis. This study aims to use different machine-learning algorithms to diagnose epileptic seizures. The frequency components of EEG data were extracted using parametric approaches. This feature extraction approach was fed into machine learning classification algorithms, including Artificial Neural Network (ANN), Gradient Boosting, and Random Forest. The ANN classifier was shown to have the most significant test performance in this investigation, with roughly 97% accuracy and a 91% F1 score in recognizing epileptic episodes. The Gradient Boosting classifier, on the other hand, performed similarly to the ANN, with 96% accuracy and a 93% F1 score.
... before the occurrence of a seizure the preictal phase starts and there are several interpretations of this interval [8][9][10]. The application of mathematical theories of non-linear dynamics initially allowed the beginning of preictal interval to be identified as the point beyond which brain activity develops deterministically, leading up to a seizure [11,12]. In other words, it is a point of no return, which once passed, a seizure will occur. ...
Epilepsy is a common neurological condition, typically diagnosed using Electroencephalogram (EEG). Large scale EEG datasets have recently been made publicly available, allowing the use of advanced Machine Learning (ML) algorithms to analyze EEG patterns associated with epilepsy. While most existing studies focus on identifying seizures in the EEG, few have tried to identify preictal EEG segments. Identifying preictal EEG segments are not only useful in developing early warning systems but also helps inform the course of treatment for the patient. In this study, we propose to represent EEG segments as images, instead of time-series data, and identify preictal EEG segments using a preexisting ML algorithm (YOLOv8) designed for image processing. Multiplexed images (containing the original EEG signal represented on a 2D grid, kurtosis, and spectral entropy) achieve the best accuracy of 95.75% on the dataset while images just containing the EEG signal result in an accuracy of 91.25%. Using only spectrograms, generated from the original EEG signal, results in an accuracy of 90.15%.
... In the last years, around 50 million of the inhabitance has been affected by epilepsy depending on the records of the World Health Organization [1,2]. Therefore, many research works have been done to identify the potential epilepsy and treatments which related with it [3][4][5][6]. In the literature, EEG records have played a main role in understanding and diagnosing of neurological diseases i.e epilepsy. ...
... From the machine learning (ML) viewpoint, a programmed forecast strategy diminishes the dangers of the abrupt eventuality of seizures. A standard seizure prediction model [8,9] comprises three different stages. Preprocessing is the first and foremost step in the prediction model to increase the Signal to Noise Ratio (SNR) to remove noise effectively. ...
Seizure prediction from electroencephalogram (EEG) time series data and a sequential deep learning (DL) predictor substantially boosts epileptic patients’ quality of life. However, a significant challenge is a variation in seizure characteristics with time and individuals along with a need for more data. Also, considerable dissimilarity is noticed in the duration between various seizure stages. Thus, a patient-generic approach is required to mitigate the problem. As a result, multiple feature augmentation procedures are used to create a hybrid feature space to capture the non-linearity of epileptic seizures. This elaborate feature space helps the predictor learn better to enhance the seizure prediction. Additionally, the predictor is optimized using a novel hybrid Forensic-based-Search-and-Rescue Optimization (FB-SARO) to improve the seizure prediction. In addition, an optimal seizure prediction horizon (SPH) is also determined through the classifier’s learning. The SPH helps attain early prediction while preserving accuracy and achieving a minimum False Prediction Rate (FPR). It also helps raise the alarm to provide the patients with ample preparation time for medical assistance. The proposed approach is testified through publicly available datasets and compared with existing state-of-the-art techniques
... The repeated unprovoked seizures are evidence of the most common neurological disease known as epilepsy. It is a long-lasting, foremost disorder of the brain, and it is affecting about 70 million individuals around the world [1]. Analysis of EEG is one of the frequently used techniques to diagnose epilepsy due to its adequate time-based resolution. ...
The seizure is an unusual event of the brain, which leads to the second most common disease of the brain called epilepsy. Electroencephalography (EEG) has the potential to provide insight into the diagnosis of seizure. Our objective is to explore the practical efficiency of the convolutional neural network (CNN) in detecting seizures using EEG signal. A novel CNN-FCM architecture is proposed to classify the seizure signals. The conventional CNN is modified with Fuzzy C-means (FCM) clustering algorithm. The competency of the clustering method is confirmed with the cluster validity index (CVI) parameters such as partition entropy (PE), partition coefficient (PC) and the Xie-Beni index (XB). The efficiency of proposed CNN-FCM architecture is validated and confirmed by considering the standard classification parameters such as accuracy, sensitivity, specificity and F-measure. The two different seizure EEG datasets are utilized to examine the proposed system. The proposed CNN-FCM architecture achieved the classification accuracy of 98.33% and outperformed with other existing deep learning methods, achieving a less computational time of 0.3286 s in classification. The performance outcomes exhibit the efficiency of the proposed one-dimensional CNN-FCM in the diagnosis of epilepsy. In contrast to other existing automatic seizure detection techniques, the CNN-FCM architecture can perform real-time seizure detection and classification.
... Epilepsy is the second most common neurological disorder after stroke, according to a report from World Health Organization [1,2]. People with epilepsy account for about 1% of the world population. ...
Background
Epilepsy is a neurological disorder that is usually detected by electroencephalogram (EEG) signals. Since manual examination of epilepsy seizures is a laborious and time-consuming process, lots of automatic epilepsy detection algorithms have been proposed. However, most of the available classification algorithms for epilepsy EEG signals adopted a single feature extraction, in turn to result in low classification accuracy. Although a small account of studies have carried out feature fusion, the computational efficiency is reduced due to too many features, because there are also some poor features that interfere with the classification results.
Methods
In order to solve the above problems, an automatic recognition method of epilepsy EEG signals based on feature fusion and selection is proposed in this paper. Firstly, the Approximate Entropy (ApEn), Fuzzy Entropy (FuzzyEn), Sample Entropy (SampEn), and Standard Deviation (STD) mixed features of the subband obtained by the Discrete Wavelet Transform (DWT) decomposition of EEG signals are extracted. Secondly, the random forest algorithm is used for feature selection. Finally, the Convolutional Neural Network (CNN) is used to classify epilepsy EEG signals.
Results
The empirical evaluation of the presented algorithm is performed on the benchmark Bonn EEG datasets and New Delhi datasets. In the interictal and ictal classification tasks of Bonn datasets, the proposed model achieves an accuracy of 99.9%, a sensitivity of 100%, a precision of 99.81%, and a specificity of 99.8%. For the interictal-ictal case of New Delhi datasets, the proposed model achieves a classification accuracy of 100%, a sensitivity of 100%, a specificity of 100%, and a precision of 100%.
Conclusion
The proposed model can effectively realize the high-precision automatic detection and classification of epilepsy EEG signals. This model can provide high-precision automatic detection capability for clinical epilepsy EEG detection. We hope to provide positive implications for the prediction of seizure EEG.
... The major causes of seizure are due to the mental depression and failure of inter-neural communication. The process of seizure prediction and diagnosis under modern technological tools such as machine learning is discussed under the review in [3]. The study holds grounds on the control measures and the precautions for the avoiding of seizure occurrence in active patients. ...
Epileptic seizure (ES) is caused due to the unpredictable and imbalanced discharge of electric signals causing a muscle ruptures. The instance is critical if unattended medically. In the proposed paper, a feature optimization and classification technique is discussed. The technique is based on the dynamic feature set extraction and producing cluster based on categorization labels. The technique is structured on grey-wolf optimization algorithm in identifying the highlighted feature–attribute co-relationship. The technique has processed attribute inter-connectivity coordinates in creating a virtual mapping and labeling of cluster-heads to provide seizure severity. The technique has successfully adopted multi-dimensional datasets for improved performance and calibration under inter-dependent attribute-feature mapping. The technique has achieved 96.76% accuracy on trained datasets with 98.76% sensitivity and 97.86% in precision on epileptic seizure classification for decision-making.
... However, evidence suggests that specific alterations in brain dynamics can be observed before epileptic attacks [5]. This discovery spurred the interest of academic centers and medical companies in building devices to anticipate seizures, primarily by analyzing the electroencephalogram (EEG) [6,7]. Monitoring devices would allow patients to avoid dangerous situations and plan the administration of preventive treatments, such as electrical stimulation or targeted drug delivery, with much greater precision. ...
There is an increasing interest in applying artificial intelligence techniques to forecast epileptic seizures. In particular, machine learning algorithms could extract nonlinear statistical regularities from electroencephalographic (EEG) time series that can anticipate abnormal brain activity. The recent literature reports promising results in seizure detection and prediction tasks using machine and deep learning methods. However, performance evaluation is often based on questionable randomized cross-validation schemes, which can introduce correlated signals (e.g., EEG data recorded from the same patient during nearby periods of the day) into the partitioning of training and test sets. The present study demonstrates that the use of more stringent evaluation strategies, such as those based on leave-one-patient-out partitioning, leads to a drop in accuracy from about 80% to 50% for a standard eXtreme Gradient Boosting (XGBoost) classifier on two different data sets. Our findings suggest that the definition of rigorous evaluation protocols is crucial to ensure the generalizability of predictive models before proceeding to clinical trials.
... As an analogy, epileptic seizures occur in what originally seemed to be an unpredictable manner. Researchers have nevertheless started using signals from electroencephalogram (EEG) to detect seizures before they happen, using classical signal processing methodologies (47) and more recently using deep learning techniques (48): if a seizure is detected before it happens, implantable devices such as VNS (49) can then sometimes be used to intervene in time to prevent the seizure from occurring. Similarly, if the models presented in this paper are improved enough, it may be possible to use exercise levels and other signals to predict when substantial pain increases may occur in order to intervene in time to prevent them from happening. ...
Overexertion can be the origin of chronic pain and exercise has been shown to increase pain in the short term in chronic pain patients. However, exercise has also been shown to cause hypoalgesia (a decreased sensitivity to painful stimuli) and is considered a treatment option for nearly all forms of chronic pain. In order to further investigate this currently unclear impact of physical strain on pain in chronic pain patients, we used consumer fitness wearables and other data gathering tools to track the exercise and pain levels of three chronic knee pain patients on a daily basis over several years: the big datasets we've collected (at the patient level) allowed us to gain a much more in depth vision of the complex interactions between physical strain and pain than what is usually possible in more classical clinical settings. We found that the timing of occurrences of maximum peaks in rolling averaged physical strain relative to peaks in rolling averaged pain points to physical strain causing pain both in the short term (on the same or the next day) and in the long term (days or weeks later). We also show preliminary evidence indicating that periods of build-up in strain may explain how physical strain is having a long term impact on pain. Our results thus also suggest that patients can't always rely exclusively on their current pain to know if they are putting too much strain on a body region, since as shown in our results, physical strain can have a long term, delayed (thus originally imperceivable) impact on pain. Therefore, going forward, building a model/AI to more accurately predict future pain based on stressors could lead to new treatment options: an AI coach could for instance transmit information and warnings to patients on a daily basis to avoid having them put too much (or too little) strain on a body region. The data gathering and analysis methods presented in this paper could also be the basis of future big personal data longitudinal studies in which biological information would be integrated.
... The dynamics of an epileptic signal are highly complex, involving different regions and networks of the brain, with signal frequency and amplitude varying over time between and even within patients (Iasemidis et al., 2003). In addition, the EEG tracing contains by nature signals that are noisy and non-stationary, which in a clinical context are also influenced by medications and vigilance states. ...
Electroencephalography (EEG) is one of the main pillars used for the diagnosis and study of epilepsy, readily employed after a possible first seizure has occurred. The most established biomarker of epilepsy, in case seizures are not recorded, are interictal epileptiform discharges (IEDs). In clinical practice, however, IEDs are not always present and the EEG may appear completely normal despite an underlying epileptic disorder, often leading to difficulties in the diagnosis of the disease. Thus, finding other biomarkers that reliably predict whether an individual suffers from epilepsy even in the absence of evident epileptic activity would be extremely helpful, since they could allow shortening the period of diagnostic uncertainty and consequently decreasing the risk of seizure. To date only a few EEG features other than IEDs seem to be promising candidates able to distinguish between epilepsy, i.e. > 60% risk of recurrent seizures, or other (pathological) conditions. The aim of this narrative review is to provide an overview of the EEG-based biomarker candidates for epilepsy and the techniques employed for their identification.
... However, the quantity of medical personnel is insufficient to manage all patients, and correct judgments cannot be made based solely on patient behavior monitoring. Therefore, several epilepsy-related studies are being conducted to ensure the stability of epilepsy patients' daily life 4,5 and to enable precise prevention and treatment with a limited medical workforce. ...
In this paper, we propose a method for predicting epileptic seizures using a pre-trained model utilizing supervised contrastive learning and a hybrid model combining Residual Networks (ResNet) and Long Short-Term Memory (LSTM). The proposed training method includes three phases: pre-processing, pre-training as a pretext task, and training as a downstream task. In the pre-processing phase, Short Time Fourier Transform (STFT) was used to convert the data into a spectrogram image with time and frequency information to compensate for the complexity and irregularity of the Electroencephalography (EEG) data, which made data analysis difficult. In the pre-training phase, band-stop filter and temporal cutout were applied to the original data to create augmented data, which were then pre-trained with a ResNet and supervised contrastive loss model to train the representation of the spectrogram image. During the training phase, image features and time information were extracted from a hybrid model consisting of ResNet initialized as weight values of a pre-trained model and LSTM. CHB-MIT and Seoul National University Hospital (SNUH) were used to validate the proposed method, and generalization performance was confirmed using Leave-one-out cross-validation. From the experimental results measuring accuracy, sensitivity, and False Positive Rate (FPR), CHB-MIT was 91.90%, 89.64%, 0.058 and SNUH was 83.37%, 79.89%, and 0.131. The experimental results demonstrate that the proposed method outperforms the conventional methods.
... Since identification of the pre-ictal state in the intracranial recording was the goal, at least three hours of recordings prior to a seizure event were taken for each analyzed seizure. EEG recordings were analyzed using a sliding window analysis technique [2]. The length of each window was 10 sec with 5 sec overlap between the adjacent windows. ...
... Epilepsy, characterised by recurrent seizures, is classified among the most commonly known chronic disorders afflicting about 70 million of the world's populace [1]. Seizure is a befalling calamity for a sufferer invoked by an abnormal and chaotic [2] yet rhythmic [3] discharging of the brain's neurons, bringing about an impermanent aberration from normal brain functioning lasting for a few seconds to a few minutes [4]. The term seizure originated from a belief inherited through the past three millennia [5], which means to be seized by an evil spirit. ...
Besides the real-time data acquired from
iEEG
, an algorithm that identifies key features is necessary for automated diagnoses of related diseases. The power of these algorithms plays a crucial role in the accuracy of medical devices. The present work reports a novel optimal feature extraction approach using wavelet transform, namely multi-depth wavelet packets, to accurately classify multi-labeled
iEEG
data acquired from epileptic patients using least training data. The paper also reports that the number of features employed by the algorithm is critical to the classification outcomes. In an attempt to select the optimal features, the employed algorithm obtains the multi-depth wavelet packets by excavating through wavelet tree down to seven levels, retaining packets at each level. Features of energy are computed and discrete cosine transform is applied across the channels for dimensionality reduction. All the extracted features are then ranked, following which an optimal number of them is determined. This optimal feature selection allows for drawing a clear line of demarcation among all the classes, which ensures perfect classification. Contrary to the state-of-the-art models, this work, in addition to providing perfect classification results in discriminating all the five classes, also takes a smaller fraction of training data to date. The
Monte-Carlo
scheme is employed to avoid any bias in the classification results.
... Epileptic seizures are disorders of the nervous system which manifest themselves in the form of hyperactivity within the cortex of the brain. It is estimated that 50~65 million people actively suffer epileptic seizures, and arebased on epidemiological statistics-highly prevalent in developing countries [1][2][3] . Epileptic seizures can be broken down into three primary types, namely: 1) generalized seizures, which are global across the brain, influencing the electrical activity of all neurons within the brain, and may result in impairment; 2) partial epilepsy, which is characterized by more localized manifestations where focal epilepsy is evidenced amongst a cluster of neurons within a particular hemisphere within the human brain [4][5][6] ; and finally 3) intermittent seizures, where the onset is unknown [1] . ...
Seizures are a widespread condition affecting 50~65 million people in the world, and newborns are also susceptible to them. EEG is used to monitor the brain activity of newborns with suspected brain injuries, followed by a qualitative waveform interpretation by a group of clinical experts, where the means towards detection of seizures include a set of distinct characteristics in the waveform. This means of seizure detection has been critiqued, particularly due to subjectivity where, at times, waveform reviewing clinicians fail to reach a consensus on the presence of seizure activity in the brain of a newborn. As a means towards dealing with this problem, the author investigated the use of Artificial Intelligence-driven prediction machines capable of an automated diagnosis of seizure, based on a newborn’s EEG waveform. This approach used a reduced selection of EEG electrodes, the Linear Series Decomposition Learner (LSDL), an ensemble of a group of features, and performance comparison across multiple classification models. Secondary work was also carried out, which leveraged the patient information available alongside the EEG dataset. This involved the use of EEG towards predicting the level of asphyxia within the neonatal brain. The results from the seizure prediction exercise showed an increment in prediction performance of the seizures when preprocessed with the LSDL. The results spanned a range of figures (depending on the classification model), with the highest accuracy of 88.1%, while a probabilistic approach towards predicting the extent of seizures provided a maximum accuracy of 93.5%. The results from the secondary analysis showed a maximum accuracy for asphyxia prediction of 89.1%. The obtained results have helped to demonstrate that a reduced selection of electrode segments, alongside the selected algorithms, can serve towards the prediction of seizures for newborns within a neonatal intensive care unit.
... This method is very subjective and needs a lot of time to inspect hours of multichannel EEG recordings. [7,8] To quantify the observation of EEG recordings, scientists develop a computer-aided diagnosis system to help neurologists detect or predict the seizure condition in the epileptic EEG signal recordings. [9][10][11] The EEG signals can provide useful information about the seizure condition. ...
When an epileptic seizure occurs, the neuron's activity of the brain is dynamically changed, which affects the connectivity between brain regions. The connectivity of each brain region can be quantified by electroencephalography (EEG) coherence, which measures the statistical correlation between electrodes spatially separated on the scalp. Previous studies conducted a coherence analysis of all EEG electrodes covering all parts of the brain. However, in an epileptic condition, seizures occur in a specific region of the brain then spreading to other areas. Therefore, this study applies an energy-based channel selection process to determine the coherence analysis in the most active brain regions during the seizure. This paper presents a quantitative analysis of inter- and intrahemispheric coherence in epileptic EEG signals and the correlation with the channel activity to glean insights about brain area connectivity changes during epileptic seizures. The EEG signals are obtained from ten patients' data from the CHB-MIT dataset. Pair-wise electrode spectral coherence is calculated in the full band and five sub-bands of EEG signals. The channel activity level is determined by calculating the energy of each channel in all patients. The EEG coherence observation in the preictal (Cohpre ) and ictal (Cohictal ) conditions showed a significant decrease of Cohictal in the most active channel, especially in the lower EEG sub-bands. This finding indicates that there is a strong correlation between the decrease of mean spectral coherence and channel activity. The decrease of coherence in epileptic conditions (Cohictal <Cohpre ) indicates low neuronal connectivity. There are some exceptions in some channel pairs, but a constant pattern is found in the high activity channel. This shows a strong correlation between the decrease of coherence and the channel activity. The finding in this study demonstrates that the neuronal connectivity of epileptic EEG signals is suitable to be analyzed in the more active brain regions.
... We illustrate RING-CPD for the identification of epileptic seizures, which over two million Americans are suffering from (Iasemidis, 2003). As a promising therapy, responsive neurostimulation requires automated algorithms to detect seizures as early as possible. ...
Change-point detection (CPD) concerns detecting distributional changes in a sequence of independent observations. Among nonparametric methods, rank-based methods are attractive due to their robustness and efficiency and have been extensively studied for univariate data. However, they are not well explored for high-dimensional or non-Euclidean data. In this paper, we propose a new method, Rank INduced by Graph Change-Point Detection (RING-CPD), based on graph-induced ranks to handle high-dimensional and non-Euclidean data. The new method is asymptotically distribution-free under the null hypothesis with an analytic p-value approximation derived for fast type-I error control. Extensive simulation studies show that the RING-CPD method works well for a wide range of alternatives and is robust to heavy-tailed distribution or outliers. The new method is illustrated by the detection of seizures in a functional connectivity network dataset and travel pattern changes in the New York City Taxi dataset.
... A solution for uncontrolled seizures might come from prediction, 1-4 as its timely anticipation opens the way to several seizure control strategies, such as: (a) closedloop systems that trigger drug delivery or electrical brain stimulation; (b) warning devices that inform the patient to prevent accidents (eg, falling from stairs) and/or to selfadminister rescue medication. [5][6][7] Although seizure prediction research started in the 1970s through electroencephalogram (EEG) analysis, [8][9][10] few predictive devices 11 and closed-loop systems 12 have been clinically approved for trial. Additionally, these were based on "detection features alone" (line-length, bandpass, and energy-related), 13 which may be less robust than current state-of-the-art approaches. ...
Seizure prediction may be the solution for epileptic patients whose drugs and surgery do not control seizures. Despite 46 years of research, few devices/systems underwent clinical trials and/or are commercialized, where the most recent state‐of‐the‐art approaches, as neural networks models, are not used to their full potential. The latter demonstrates the existence of social barriers to new methodologies due to data bias, patient safety, and legislation compliance. In the form of literature review, we performed a qualitative study to analyze the seizure prediction ecosystem to find these social barriers. With the Grounded Theory, we draw hypotheses from data, while with the Actor‐Network Theory we considered that technology shapes social configurations and interests, being fundamental in healthcare. We obtained a social network that describes the ecosystem and propose research guidelines aiming at clinical acceptance. Our most relevant conclusion is the need for model explainability, but not necessarily intrinsically interpretable models, for the case of seizure prediction. Accordingly, we argue that it is possible to develop robust prediction models, including black‐box systems to some extent, while avoiding data bias, ensuring patient safety, and still complying with legislation, if they can deliver human‐ comprehensible explanations. Due to skepticism and patient safety reasons, many authors advocate the use of transparent models which may limit their performance and potential. Our study highlights a possible path, by using model explainability, on how to overcome these barriers while allowing the use of more computationally robust models.
... Approximately more than 47 million people have dementia and around 50 million are affected by epilepsy. Around one in every 100 persons will encounter a seizure at some point in their lifetime [1]. Most of these neurological disorders require long term electroencephalography (EEG) monitoring for diagnosis and treatment. ...
Motion artifacts contribute complexity in acquiring clean electroencephalography (EEG) data. It is one of the major challenges for ambulatory EEG. The performance of mobile health monitoring, neurological disorders diagnosis and surgeries can be significantly improved by reducing the motion artifacts. Although different papers have proposed various novel approaches for removing motion artifacts, the datasets used to validate those algorithms are questionable. In this paper, a unique EEG dataset was presented where ten different activities were performed. No such previous EEG recordings using EMOTIV EEG headset are available in research history that explicitly mentioned and considered a number of daily activities that induced motion artifacts in EEG recordings. Quantitative study shows that in comparison to correlation coefficient, the coherence analysis depicted a better similarity measure between motion artifacts and motion sensor data. Motion artifacts were characterized with very low frequency which overlapped with the Delta rhythm of the EEG. Also, a general wavelet transform based approach was presented to remove motion artifacts. Further experiment and analysis with more similarity metrics and longer recording duration for each activity is required to finalize the characteristics of motion artifacts and henceforth reliably identify and subsequently remove the motion artifacts in the contaminated EEG recordings.
... In the time domain, we extracted the first four statistical moments (mean, variance, skewness and kurtosis) and the three Hjorth parameters (activity, mobility and complexity). As for the frequency domain, we extracted the relative spectral power of the delta (0.5-4 Hz), theta (4-8 Hz), alpha (8)(9)(10)(11)(12), beta (13)(14)(15)(16)(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30), low-gamma , and high-gamma (79-128 Hz) bands, the spectral edge frequency at three different cut-off percentages (50%, 75%, and 90%) and the energy of each wavelet coefficient (D1 to A8, using the Daubechies 4 mother wavelet (db4)). As the frequency limit of gamma activity is not consensual among the scientific community, and its division into high-gamma and low-gamma is not uncommon 54 , we decided to divide it. ...
Seizure prediction might be the solution to tackle the apparent unpredictability of seizures in patients with drug-resistant epilepsy, which comprise about a third of all patients with epilepsy. Designing seizure prediction models involves defining the pre-ictal period, a transition stage between inter-ictal brain activity and the seizure discharge. This period is typically a fixed interval, with some recent studies reporting the evaluation of different patient-specific pre-ictal intervals. Recently, researchers have aimed to determine the pre-ictal period, a transition stage between regular brain activity and a seizure. Authors have been using deep learning models given the ability of such models to automatically perform pre-processing, feature extraction, classification, and handling temporal and spatial dependencies. As these approaches create black-box models, clinicians may not have sufficient trust to use them in high-stake decisions. By considering these problems, we developed an evolutionary seizure prediction model that identifies the best set of features while automatically searching for the pre-ictal period and accounting for patient comfort. This methodology provides patient-specific interpretable insights, which might contribute to a better understanding of seizure generation processes and explain the algorithm’s decisions. We tested our methodology on 238 seizures and 3687 h of continuous data, recorded on scalp recordings from 93 patients with several types of focal and generalised epilepsies. We compared the results with a seizure surrogate predictor and obtained a performance above chance for 32% patients. We also compared our results with a control method based on the standard machine learning pipeline (pre-processing, feature extraction, classifier training, and post-processing), where the control marginally outperformed our approach by validating 35% of the patients. In total, 54 patients performed above chance for at least one method: our methodology or the control one. Of these 54 patients, 21 (≈38%) were solely validated by our methodology, while 24 (≈44%) were only validated by the control method. These findings may evidence the need for different methodologies concerning different patients.
... Although the EEG has achieved reliable sensitivity and specificity in detecting electrical activity of the brain despite the various forms of seizures in epilepsy, it is invasive and often has an intolerance to motion artifacts, thereby providing false-positive readings [32]. Moreover, hospitals use EEG to diagnose seizures as a presurgical screening procedure and to monitor patients' seizure progression, thereafter prescribing them with anti-seizure medications [54], with no possibility of predicting the seizure occurrence in the patient after discharge and not knowing if the prescription made actively reduces seizures occurrence, other than relying on patients' seizure diaries. The preictal period remains as the most difficult period to be detected as it is not clinically annotated and has no presence of recurrent pattern [55]. ...
Background:
Epilepsy is a devastating neurological disorder that affects nearly 70 million people worldwide. Epilepsy causes uncontrollable, unprovoked and unpredictable seizures that reduce the quality of life of those afflicted, with 1-9 epileptic patient deaths per 1000 patients occurring annually due to sudden unexpected death in epilepsy (SUDEP). Predicting the onset of seizures and managing them may help patients from harming themselves and may improve their well-being. For a long time, electroencephalography (EEG) devices have been the mainstay for seizure detection and monitoring. This systematic review aimed to elucidate and critically evaluate the latest advancements in medical devices, besides EEG, that have been proposed for the management and prediction of epileptic seizures. A literature search was performed on three databases, PubMed, Scopus and EMBASE.
Methods:
Following title/abstract screening by two independent reviewers, 27 articles were selected for critical analysis in this review.
Results:
These articles revealed ambulatory, non-invasive and wearable medical devices, such as the in-ear EEG devices; the accelerometer-based devices and the subcutaneous implanted EEG devices might be more acceptable than traditional EEG systems. In addition, extracerebral signalbased devices may be more efficient than EEG-based systems, especially when combined with an intervention trigger. Although further studies may still be required to improve and validate these proposed systems before commercialization, these findings may give hope to epileptic patients, particularly those with refractory epilepsy, to predict and manage their seizures.
Conclusion:
The use of medical devices for epilepsy may improve patients' independence and quality of life and possibly prevent sudden unexpected death in epilepsy (SUDEP).
... Approximately more than 47 million people have dementia and around 50 million are affected by epilepsy. Around one in every 100 persons will encounter a seizure at some point in their lifetime [1]. Most of these neurological disorders require long term EEG monitoring for diagnosis and treatment. ...
Motion artifacts contribute complexity in acquiring clean EEG data. It is one of the major challenges for ambulatory EEG. The performance of mobile health monitoring, neurological disorders diagnosis and surgeries can be significantly improved by reducing the motion artifacts. Although different papers have proposed various novel approaches for removing motion artifacts, the datasets used to validate those algorithms are questionable. In this paper, a unique EEG dataset was presented where ten different activities were performed. For this purpose, an EMOTIV EEG headset alongside built-in motion sensors was used. No such previous EEG recordings are available in research history that explicitly mentioned and considered a number of daily activities that induced motion artifacts in EEG recordings. Quantitative study shows that in comparison to correlation coefficient, the coherence analysis depicted a better similarity measure between motion artifacts & motion sensor data. Motion artifacts were characterized with very low frequency which overlapped with the Delta rhythm of the EEG. Also, a general wavelet transform based approach was presented which can be used in future to remove motion artifacts. Further experiment and analysis with more similarity metrics and longer recording duration for each activity is required to finalize the characteristics of motion artifacts and henceforth reliably identify and subsequently remove the motion artifacts in the contaminated EEG recordings.
Epilepsy is a chronic brain disease caused by excessive discharge of brain neurons. Long-term recurrent seizures bring a lot of trouble to patients and their families. Prediction of different stages of epilepsy is of great significance. We extract pearson correlation coefficients (PCC) between channels in different frequency bands as features of EEG signals for epilepsy stages prediction. However, the features are of large feature dimension and serious multi-collinearity. To eliminate these adverse influence, the combination of traditional dimension reduction method principal component analysis (PCA) and logistic regression method with regularization term is proposed to avoid over-fitting and achieve the feature sparsity. The experiments are conducted on the widely used CHB-MIT dataset using different regularization terms L1 and L2, respectively. The proposed method identifies various stages of epilepsy quickly and efficiently, and it presents the best average accuracy of 94.86%, average precision of 96.71%, average recall of 93.48%, average kappa value of 0.90 and average Matthews correlation coefficient (MCC) value of 0.90 for all patients.
Knowing when seizures occur may help patients and can also provide insight into epileptogenesis mechanisms. We recorded seizures over periods of several days in the Genetic Absence Epileptic Rat from Strasbourg (GAERS) model of absence epilepsy, while we monitored behavioral activity with a combined head accelerometer (ACCEL), neck electromyogram (EMG), and electrooculogram (EOG). The three markers consistently discriminated between states of behavioral activity and rest. Both GAERS and control Wistar rats spent more time in rest (55–66%) than in activity (34–45%), yet GAERS showed prolonged continuous episodes of activity (23 vs. 18 min) and rest (34 vs. 30 min). On average, seizures lasted 13 s and were separated by 3.2 min. Isolated seizures were associated with a decrease in the power of the activity markers from steep for ACCEL to moderate for EMG and weak for EOG, with ACCEL and EMG power changes starting before seizure onset. Seizures tended to occur in bursts, with the probability of seizing significantly increasing around a seizure in a window of ±4 min. Furthermore, the seizure rate was strongly increased for several minutes when transitioning from activity to rest. These results point to mechanisms that control behavioral states as determining factors of seizure occurrence.
Delay Differential Analysis (DDA) is a nonlinear method for analyzing time series based on principles from nonlinear dynamical systems. DDA is extended here to incorporate network aspects to improve the dynamical characterization of complex systems. To demonstrate its effectiveness, DDA with network capabilities was first applied to the well-known Rössler system under different parameter regimes and noise conditions. Network-motif DDA, based on cortical regions, was then applied to invasive intracranial electroencephalographic data from drug-resistant epilepsy patients undergoing presurgical monitoring. The directional network motifs between brain areas that emerge from this analysis change dramatically before, during, and after seizures. Neural systems provide a rich source of complex data, arising from varying internal states generated by network interactions.
Data-driven and dynamic models have been proposed to better understand and describe epileptiform activities. However, there is still a need for more comprehensive ones suitable for individual seizures from patients but also general across them, taking into account common features. Thus, this work proposes an alternative nonlinear autoregressive model to address this issue. Essentially, the regression coefficients obtained from individual models are taken as random variables, and a new one is constructed based on their respective median values. The results show that the proposed approach is effective in describing all of the individual seizures with a single model, yielding low estimation errors. This finding is useful when large datasets are not available, and can be adopted in the context of seizure prediction and attenuation in a similar way to that of transfer learning applications.
Epilepsy is a common neurological disorder in which the activity of brain becomes abnormal, causing sensation, loss of awareness, and periods of unusual behavior or seizures, which is recurrent in nature. In localization, high-frequency oscillations (HFOs) are spontaneous EEG patterns that have been regarded as potential biomarkers of epileptic seizure on zones (SOZs). Accurately detected HFOs are used to localize SOZs, which is crucial for the presurgical assessment. Since the visual marking of HFOs is time-consuming, a method is desirable to detect HFOs for localizing SOZs in clinical practice automatically. However, the existing techniques cannot obtain satisfactory performance, which is not suitable for clinical application. To solve this problem, a new localization method for epileptic SOZs has been proposed based on Modified Remora Optimization based Matching Pursuit with Density Peak Clustering (MRO-DC). Initially, a Stockwell entropy based on the Hilbert transform is used to detect events of interest (EoIs). In addition, a time–frequency analysis method like Shannon-entropy-based complex Morlet wavelet transform (SE-CMWT) is adopted to acquire channels of interest (CoIs) by calculating the average power of EoIs on each channel. Subsequently, a MRO-DC approach has been proposed to detect HFOs of CoIs. Finally, the concentrations of the detected HFOs are used to localize SOZs better. The proposed approach is implemented in the Xilinx working platform by instigating Verilog code. The result of the proposed approach showed that the MRO-DC had achieved maximum specificity and sensitivity on Field Programmable Gate Array (FPGA). The maximum performance of 99.4% specificity, 98.2% sensitivity, and 0.575 ns computational time is obtained while testing the FPGA to localize the epileptic SOZs.
This work presents the world’s first neural signal processor for seizure prediction, which includes a preprocessing unit, a feature extractor, a reconfigurable support vector machine (SVM) kernel, and a postprocessing unit. Seizure prediction performance is enhanced by on-chip training for model adaptation. Design optimization is applied across the layers of abstraction to minimize the area and energy. The area of the feature extractor is reduced by 28% with an approximated energy operator (AEO). The proposed scaling-based Newton–Raphson (NR) divider reduces the required number of iterations for division by 62.5%. For alternating direction method of multipliers (ADMM)-based SVM training, the computational complexity is reduced by up to 99.9% through pointer-based matrix multiplication. By leveraging the LDL decomposition, 80% multiplications for updating weights are saved. The chip achieves a seizure prediction performance with a 92.0% sensitivity and a 0.57/h false alarm rate (FAR). The training latency is 8.44 ms with a power dissipation of 2.31 mW at 6.05 MHz. Compared with an ARM Cortex-M3 microcontroller, this work achieves a 124
$\times$
higher area efficiency and a 299
$\times$
higher energy efficiency. The chip also supports seizure detection and achieves a sensitivity of 98.6% and an FAR of 0.18/h, exceeding the state-of-the-art designs.
Electroencephalogram (EEG) is a recording of the electrical movement of the brain from the scalp. For sleep disorder analysis the EEG test is done while the subject is sleeping. In this paper discuss about recording of brain signal (EEG) and how these signal play major role to finding in different brain diseases. EEG data can be different when subjects are asleep or when exhausted or when some sort of action takes place. When the patient is awake standard EEG test can be taken, but it may not demonstrate any unusual electrical action. During the sleep, brainwave patterns alter and may show more unusual electrical action.
Epilepsy is a disorder of the brain denoted by frequent seizures. The symptoms of seizure include confusion, abnormal staring, and rapid, sudden, and uncontrollable hand movements. Epileptic seizure detection methods involve neurological exams, blood tests, neuropsychological tests, and neuroimaging modalities. Among these, neuroimaging modalities have received considerable attention from specialist physicians. One method to facilitate the accurate and fast diagnosis of epileptic seizures is to employ computer-aided diagnosis systems (CADS) based on deep learning (DL) and neuroimaging modalities. This paper has studied a comprehensive overview of DL methods employed for epileptic seizures detection and prediction using neuroimaging modalities. First, DL-based CADS for epileptic seizures detection and prediction using neuroimaging modalities are discussed. Also, descriptions of various datasets, preprocessing algorithms, and DL models which have been used for epileptic seizures detection and prediction have been included. Then, research on rehabilitation tools has been presented, which contains brain-computer interface (BCI), cloud computing, internet of things (IoT), hardware implementation of DL techniques on field-programmable gate array (FPGA), etc. In the discussion section, a comparison has been carried out between research on epileptic seizure detection and prediction. The challenges in epileptic seizures detection and prediction using neuroimaging modalities and DL models have been described. In addition, possible directions for future works in this field, specifically for solving challenges in datasets, DL, rehabilitation, and hardware models, have been proposed. The final section is dedicated to the conclusion which summarizes the significant findings of the paper.
Manual inspection of Electroencephalography (EEG) signals to detect epileptic seizures is time-consuming and prone to inter-rater variability. Moreover, EEG signals are contaminated with different noise sources, e.g., patient movement during seizures, making the accurate identification of seizure activities challenging. In a Multi-View seizure detection system, since seizures do not uniformly affect the brain, some views likely play a more significant role in detecting seizures and should therefore be assigned a higher weight in the concatenation step. To address this dynamic weight assignment issue and also create a more interpretable model, in this work, we propose a fusion attentive deep multi-view network (fAttNet). The fAttNet combines temporal multi-channel EEG signals, wavelet packet decomposition (WPD), and hand-engineered features as three key views. We also propose an artifact rejection approach to remove unwanted signals not originating from the brain. Experimental results on the Temple University Hospital (TUH) seizure database demonstrate that the proposed method has increased performance over the state-of-the-art methods, raising accuracy, and F1-score from 0.82 to 0.86, and 0.78 to 0.81, respectively. More importantly, the proposed method is interpretable for medical professionals, assisting clinicians in identifying the regions of the brain involved in the seizures.
Epilepsy is a common brain disorder that causes patients to face multiple seizures in a single day. Around 65 million people are affected by epilepsy worldwide. Patients with focal epilepsy can be treated with surgery, whereas generalized epileptic seizures can be managed with medications. It has been noted that in more than 30% of cases, these medications fail to control epileptic seizures, resulting in accidents and limiting the patient’s life. Predicting epileptic seizures in such patients prior to the commencement of an oncoming seizure is critical so that the seizure can be treated with preventive medicines before it occurs. Electroencephalogram (EEG) signals of patients recorded to observe brain electrical activity during a seizure can be quite helpful in predicting seizures. Researchers have proposed methods that use machine and/or deep learning techniques to predict epileptic seizures using scalp EEG signals; however, prediction of seizures with increased accuracy is still a challenge. Therefore, we propose a three-step approach. It includes preprocessing of scalp EEG signals with PREP pipeline, which is a more sophisticated alternative to basic notch filtering. This method uses a regression-based technique to further enhance the SNR, with a combination of handcrafted, i.e., statistical features such as temporal mean, variance, and skewness, and automated features using CNN, followed by classification of interictal state and preictal state segments using LSTM to predict seizures. We train and validate our proposed technique on the CHB-MIT scalp EEG dataset and achieve accuracy of 94%, sensitivity of 93.8%, and 91.2% specificity. The proposed technique achieves better sensitivity and specificity than existing methods.
Neuroscience and cognitive neuroscience is one of the most fascinating fields of science
in the last decade. The complexities of the brain and the systems based on it have
attracted the attention of researchers in various sciences such as computer science,
mathematics, psychology and engineering. Due to the high levels of complexity of this
science, various models have been proposed to analyze the behavior of the brain from
the surface of a single neuron or a network of neurons that increase scientists insight
into understanding the brain and its function. At the same time, many brain diseases
are the result of some functional or destructive problems in which there is still no suitable treatment for many of these diseases. Abnormal synchronization of a network
of neurons is one of the causes of seizures in epilepsy or severe tremors in Parkinson’s.
In this dissertation, Meshless methods have been developed to simulate neuronal models, coupling and neuronal synchronization systems, as well as a definite and stochastic
control system for neuronal synchronization and its efficiency has been demonstrated.
Also, from the technological point of view and the product of this dissertation, a package based on MATLAB software has been implemented to develop control models in
order to (de)synchronize a network of neural oscillators, which researchers in medical sciences and dynamic systems can evaluate the accuracy of their defined controls
(without the need for professional coding), so that these new control mechanisms can
be used in the treatment of neurological diseases such as epilepsy and Parkinson’s
Epileptic seizure was possibly the main motivation for measuring the electrical potentials of the brain which led to the invention of electroencephalogram (EEG) systems and analysis of EEG patterns. EEG, magnetoencephalograms, and functional magnetic resonance imaging are the major neuroimaging modalities used for seizure detection. This chapter focuses on the use of brain waves mainly EEG for seizure detection, and discusses chaotic behaviour of the EEGs. The phase‐slope index metric used for the brain connectivity measure, identifies increases in the spatiotemporal interactions between channels that clearly distinguish seizure from interictal activity. A seizure prediction system can detect seizures prior to their occurrence and allow clinicians to provide timely treatment for patients with epilepsy. Interictal epileptiform discharges originate from deep within the brain or hippocampus and in many cases are descriptive of the epileptic condition. They are the most reliable biomarkers and are widely used in clinical evaluations.
We describe a novel method of adaptively controlling epileptic seizure-like events in hippocampal brain slices using electric fields. Extracellular neuronal activity is continuously recorded during field application through differential extracellular recording techniques, and the applied electric field strength is continuously updated using a computer-controlled proportional feedback algorithm. This approach appears capable of sustained amelioration of seizure events in this preparation when used with negative feedback. Seizures can be induced or enhanced by using fields of opposite polarity through positive feedback. In negative feedback mode, such findings may offer a novel technology for seizure control. In positive feedback mode, adaptively applied electric fields may offer a more physiological means of neural modulation for prosthetic purposes than previously possible.
We review some of the history and early work in the area of synchronization in chaotic systems. We start with our own discovery of the phenomenon, but go on to establish the historical timeline of this topic back to the earliest known paper. The topic of synchronization of chaotic systems has always been intriguing, since chaotic systems are known to resist synchronization because of their positive Lyapunov exponents. The convergence of the two systems to identical trajectories is a surprise. We show how people originally thought about this process and how the concept of synchronization changed over the years to a more geometric view using synchronization manifolds. We also show that building synchronizing systems leads naturally to engineering more complex systems whose constituents are chaotic, but which can be tuned to output various chaotic signals. We finally end up at a topic that is still in very active exploration today and that is synchronization of dynamical systems in networks of oscillators.
We evaluate the capability of nonlinear time series analysis to extract features from brain electrical activity (EEG) predictive of epileptic seizures. Time-resolved analysis of the EEG recorded in 16 patients from within the seizure-generating area of the brain indicate marked changes in nonlinear characteristics for up to several minutes prior to seizures as compared to other states or recording sites. If interpreted as a loss of complexity in brain electrical activity these changes could reflect the hypothesized continuous increase of synchronization between pathologically discharging neurons and allow one to study seizure-generating mechanisms in humans.
Recent reports have suggested that chaos control techniques may be useful for electrically manipulating epileptiform bursting behavior in neuronal ensembles. Because the dynamics of spontaneous in vitro bursting had not been well determined previously, analysis of this behavior in the rat hippocampus was performed. Epileptiform bursting was induced in transverse rat hippocampal slices using three experimental methods. Slices were bathed in artificial cerebrospinal fluid containing: (1) elevated potassium ([K+]o=10.5 mM), (2) zero magnesium, or (3) the GABAA-receptor antagonists bicuculline (20 μM) and picrotoxin (250 μM). The existence of chaos and determinism was assessed using two different analytical techniques: unstable periodic orbit (UPO) analysis and a new technique for estimating Lyapunov exponents. Significance of these results was assessed by comparing the calculations for each experiment with corresponding randomized surrogate data. UPOs of multiple periods were highly prevalent in experiments from all three epilepsy models: 73% of all experiments contained at least one statistically significant period-1 or period-2 orbit. However, the expansion rate analysis did not provide any evidence of determinism in the data. This suggests that the system may be globally stochastic but contains local pockets of determinism. Thus, manipulation of bursting behavior using chaos control algorithms may yet hold promise for reverting or preventing epileptic seizures. © 2001 Biomedical Engineering Society.
PAC01: 8719Nn, 8719Xx, 0545Gg, 8717Nn, 8719La
There has been a heated debate about whether chaos theory can be applied to the dynamics of the human brain. While it is obvious that nonlinear mechanisms are crucial in neural systems, there has been strong criticism of attempts to identify at strange attractors in brain signals and to measure their fractal dimensions, Lyapunov exponents, etc. Conventional methods analyzing brain dynamics are largely based on linear models and on Fourier spectra. Regardless of the existence of strange attractors in brain activity, the neurosciences should benefit greatly from alternative methods that have been developed in recent years for the analysis of nonlinear and chaotic behavior.
As Cabrera-Valdivia et al 1 suggest, ecstatic experiences during spontaneous epileptic seizures are exceedingly rare. Intensely pleasurable experiences are, however, common in self induced seizures, as in the patient whom they describe. Of those 5% of people with epilepsy who are photosensitive some 30% can be shown by prolonged EEG …
This second edition of Seizures and Epilepsy is completely revised, due to tremendous advances in the understanding of the fundamental neuronal mechanisms underlying epileptic phenomena, as well as current diagnosis and treatment, which have been heavily influenced over the past several decades by seminal neuroscientific developments, particularly the introduction of molecular neurobiology, genetics, and modern neuroimaging. This resource covers a broad range of both basic and clinical epileptology.
Preliminary reports have suggested that chronic, intermittent stimulation of the vagus nerve (VNS) is an alternative treatment for patients with medically refractory seizures. We performed a multicenter, randomized, controlled trial to evaluate the efficacy and safety of adjunctive VNS in patients with poorly controlled partial seizures. An implanted, programmable pacemaker-like device was connected to two stimulating electrodes wrapped around the left vagus nerve. One hundred fourteen patients were randomized to receive 14 weeks of high-level stimulation (presumed therapeutic dose) or low-level stimulation (presumed subtherapeutic dose) using a blinded, parallel study design. Seizure frequency was compared with a 12-week baseline. Mean reduction in seizure frequency was 24.5% for the “high” stimulation group versus 6.1% for the “low” stimulation group (p = 0.01). Thirty-one percent of patients receiving high stimulation had a seizure frequency reduction of >50%, versus 13% of patients in the low group (p = 0.02). Treatment emergent side effects were largely limited to a transient hoarseness occurring during the stimulation train. One patient with no previous history of cardiac disease experienced a myocardial infarction during the third month of vagal stimulation. VNS may be an effective alternative treatment for patients who have failed antiepileptic drug therapy and are not optimal candidates for epilepsy surgery.
Bifurcation mechanisms involved in the generation of action potentials (spikes) by neurons are reviewed here. We show how the type of bifurcation determines the neuro-computational properties of the cells. For example, when the rest state is near a saddle-node bifurcation, the cell can fire all-or-none spikes with an arbitrary low frequency, it has a well-defined threshold manifold, and it acts as an integrator; i.e. the higher the frequency of incoming pulses, the sooner it fires. In contrast, when the rest state is near an Andronov-Hopf bifurcation, the cell fires in a certain frequency range, its spikes are not all-or-none, it does not have a well-defined threshold manifold, it can fire in response to an inhibitory pulse, and it acts as a resonator; i.e. it responds preferentially to a certain (resonant) frequency of the input. Increasing the input frequency may actually delay or terminate its firing. We also describe the phenomenon of neural bursting, and we use geometric bifurcation theory to extend the existing classification of bursters, including many new types. We discuss how the type of burster defines its neuro-computational properties, and we show that different bursters can interact, synchronize and process information differently.
A number of seizure events, typical of everyday clinical practice, were tested for possible chaos. After a careful elimination of spurious effects, evidence of chaos was found in two seizure events. This was confirmed by direct examination of exponential separation of initially nearby states in low-dimensional trajectory recoveries. The resulting Lyapunov exponent calculation provides a clear indication of chaos in such events, the definitive conclusion of which cannot be made from dimension calculations alone. The weakness of these techniques, and strategies for their treatment, are emphasized.
Electroconvulsive therapy, once vilified, is slowly receiving greater interest and use in the treatment of mental illness.
Proceedings of the Fifth International Workshop on Mathematical Methods in Scattering Theory and Biomedical Technology
Corfu, Greece, 18 – 19 October 2001
Edited by: Dimitrios Fotiadis ( University of Ioannina, Greece), Christos Massalas ( University of Ioannina, Greece)
The complex chaotic, unstable, noisy and nonlinear dynamics of the brain requires alternative approaches to identification and simulation of brain activity. These approaches differ from the universally accepted stochastic simulation of random processes with a given distribution. In this report we discuss the possibility of using a global optimization approach to the reconstruction of brain dynamics under the assumption that the diagnostic information comes in the form of a nonlinear time series. We consider a method for global reconstruction of nonlinear models for systems where all the necessary variables have not been observed. This technique can be applied to systems with one or several such hidden variables, and can be used to reconstruct maps or differential equations of brain dynamics. The quadratic programing approach to reconstruction of dynamical process is considered. We propose the possibility of the global reconstruction of the Fokker – Planck equation for a multi–variable distribution function which re ects the complexity of the considered brain. Finally, we demonstrate an application of the reconstructing technique to the analysis of a complex noisy system.
Epilepsy is a common neurological disorder characterized by recurrent seizures, most of which appear to occur spontaneously. Our research, employing novel signal processing techniques based on the theory of nonlinear dynamics, led us to the hypothesis that seizures represent a spatiotemporal state transition in a complex chaotic system. Through the analysis of long-term intracranial EEG recordings obtained in patients with medically intractable seizures, we discovered that seizures were preceded by a preictal transition that evolves over tens of minutes. This transition is followed by a seizure. Following the seizure, the spatiotemporal dynamics appear to be reset. The study of this process has been hampered by its complexity and variability. A major problem was that the transitions involve a subset of brain sites that vary from seizure to seizure, even in the same patient. However, by combining dynamical analytic techniques with a powerful global optimization algorithm for selecting critical electrode sites, we have been able to elucidate important dynamical characteristics underlying human epilepsy. We illustrate the use of these approaches in confirming our hypothesis regarding postictal resetting of the preictal transition by the seizure. It is anticipated that these observations will lead to a better understanding of the physiological processes involved. From a practical perspective, this study indicates that it may be possible to develop novel therapeutic approaches involving carefully timed interventions and reset the preictal transition of the brain well prior to the onset of the seizure.
Epilepsy is a dynamical disorder of the brain. The existence of an epileptogenic focus that progressively entrains normal brain areas results in transitions of the brain from chaotic to less chaotic spatiotemporal states, the epileptic seizures. Measures of chaos of the electrical activity (EEG) of the brain (e.g., Lyapunov exponents) can be used to quantify this entrainment. Using optimization theory, in particular quadratic integer programming, we were able to select the brain sites that are entrained to each other and subsequently follow their entrainment over long time periods in one patient with 24 seizures. This procedure, which is applied to epilepsy research for the first time, resulted in the identification of the epileptogenic focus and showed the possibility of prediction of epileptic seizures well in advance of their occurrence.
Epilepsy is one of the most common disorders of the nervous system, second only to strokes. We have shown in the past that progressive entrainment between an epileptogenic focus and normal brain areas results in transitions of the brain from chaotic to less chaotic spatiotemporal states, the well-known epileptic seizures. The entrainment between two brain sites can be quantified by the T-index between measures of chaos (e.g., Lyapunov exponents) estimated from the brain electrical activity (EEG) at these sites. Recently, by applying optimization theory, and in particular quadratic zero-one programming, selecting the most entrained brain sites 10 minutes before seizures and subsequently tracing their entrainment backward in time over at most 2 hours, we have shown that over 90% of the seizures in five patients with multiple seizures were predictable. In this communication we show that the above procedure, applied to measures of angular frequency in the state space (average rate of phase change of state) estimated from EEG data per recording brain site over time in one of our patients with 24 recorded seizures, produces very similar results about the predictability of the epileptic seizures (87·5%). This finding implies an interrelation of the phase and chaos entrainment in the epileptic brain and may be used to refine procedures tot long-term prediction of epileptic seizures as well as to generate a model of the disorder within the framework of dynamical nonlinear systems.
1. Introduction; 2. Some helpful tools; 3. Visualization of the
pendulum's dynamics; 4. Toward an understanding of chaos; 5. The
characterization of chaotic attractors; 6. Experimental
characterization, prediction, and modification of chaotic states; 7.
Chaos broadly applied; Further reading; Appendix A. Numerical
integration - Runge-Kutta method; Appendix B. Computer program
listings; References; Index.
The problems encountered in the study of three-dimensional Hamiltonian systems by means of the Poincare cross-sections are reviewed. A new method to overcome these problems is proposed. In order to visualize the four-dimensional “space” of section we introduce the use of color and rotation. We apply this method to the case of a family of simple periodic orbits in a three-dimensional potential and we describe the differences in the orbital behavior between regions close to stable and unstable periodic orbits. We outline the differences between the transition from stability to simple instability and the transition from stability to complex instability. We study the changes in the structure of the 4D “spaces” of section, which occur when the family becomes complex unstable after a DU→Δ or a S→Δ transition. We conclude that the orbital behavior after the transition depends on the orbital behavior before it.
Stochastic resonance is by now a well studied phenomenon whereby certain nonlinear systems, subject to weak input signals, have the property that the presentation of stochastic forcing, or “noise,” can enhance the coherence of the output. Since its introduction in 1981, this curious phenomenon has been the object of much study, yet a number of questions remain. In addition to offering an update to recent reviews, we hope here to set the stage with a brief tutorial, raise some questions and then to offer a speculative look towards the future.
This paper deals with the role of neural-network based prediction for the modeling of nonlinear dynamical systems. We show experimentally that the backpropagation learning rule to train neural networks and the prediction error, so widely utilized in teaching and comparing nonlinear predictors, do not consistently indicate that the neural network based model has indeed captured the dynamics of the system that produced the time series. Frequently, but not always, the neural network when used as an autonomous system in a feedback configuration was able to generate a time series that has dynamical invariants similar to the original time series. We show that the estimation of the dynamical invariants (correlation dimension, largest Lyapunov exponent) of the predicted and original time series are an appropriate tool to validate the predictive model.
Epilepsy is one of the most common disorders of the nervous system. The progressive entrainment between an epileptogenic focus and normal brain areas results to transitions of the brain from chaotic to less chaotic spatiotemporal states, the epileptic seizures. The entrainment between two brain sites can be quantified by the T-index from the measures of chaos (e.g., Lyapunov exponents) of the electrical activity (EEG) of the brain. By applying the optimization theory, in particular quadratic zero-one programming, we were able to select the most entrained brain sites 10 minutes before seizures and subsequently follow their entrainment over 2 hours before seizures. In five patients with 3–24 seizures, we found that over 90% of the seizures are predictable by the optimal selection of electrode sites. This procedure, which is applied to epilepsy research for the first time, shows the possibility of prediction of epileptic seizures well in advance (19.8 to 42.9 minutes) of their occurrence.
Open nonlinear systems are capable of self-organization in space and time. This realization constitutes a major breakthrough of modern science, and is currently at the origin of explosive developments in chemistry, physics and biology. Observations and numerical computations of nonlinear systems surprise us by their inexhaustible and sometimes nonintuitive variety of structures with different shapes and functions. But as well as variety one finds on closer inspection that nonlinear phenomena share universal aspects of pattern formation in time and space. These similarities make it possible to bridge the gap between inanimate and living matter at various levels of complexity, in both theory and experiment. This book is an account of different approaches to the study of this pattern formation. The universality of kinetic, thermodynamic and dimensional approaches is documented through their application to purely mathematical, physical and chemical systems, as well as to systems in nature: biochemical, cellular, multicellular, physiological, neurophysiological, ecological and economic systems. Hints given throughout the book allow the reader to discover how to make use of the principles and methods in different fields of research, including those not treated explicitly in the book.
Methods of controlling systems that behave chaotically are described. The use of chaos to stabilize lasers, electric circuits, and rabbit hearts and to send secret messages is examined.
Recently, interest has turned to the mathematical concept of chaos as an explanation for a variety of complex processes in nature. Chaotic systems, among other characteristics, can produce what appears to be random output. Another property of chaotic systems is that they may exhibit abrupt intermittent transitions between highly ordered and disordered states. Because of this property, it is hypothesized that epilepsy may be an example of chaos. In this review, some of the basic concepts of nonlinear dynamics and chaos are illustrated. Mathematical techniques developed to study the properties of nonlinear dynamical systems are outlined. Finally, the results of applying these techniques to the study of human epilepsy are discussed. The application of these powerful and novel mathematical techniques to analysis of the electroencephalogram has provided new insights into the epileptogenic process and may have considerable utility in the diagnosis and treatment of epilepsy.
