Corrigendum: Assessing MR-compatibility of somatosensory stimulation devices: a systematic review on testing methodologies

[This corrects the article DOI: 10.3389/fnins.2023.1071749.].

Neurofeedback training of executive function in autism spectrum disorder: distinct effects on brain activity levels and compensatory connectivity changes

BACKGROUND

Deficits in executive function (EF) are consistently reported in autism spectrum disorders (ASD). Tailored cognitive training tools, such as neurofeedback, focused on executive function enhancement might have a significant impact on the daily life functioning of individuals with ASD. We report the first real-time fMRI neurofeedback (rt-fMRI NF) study targeting the left dorsolateral prefrontal cortex (DLPFC) in ASD.

METHODS

Thirteen individuals with autism without intellectual disability and seventeen neurotypical individuals completed a rt-fMRI working memory NF paradigm, consisting of subvocal backward recitation of self-generated numeric sequences. We performed a region-of-interest analysis of the DLPFC, whole-brain comparisons between groups and, DLPFC-based functional connectivity.

RESULTS

The ASD and control groups were able to modulate DLPFC activity in 84% and 98% of the runs. Activity in the target region was persistently lower in the ASD group, particularly in runs without neurofeedback. Moreover, the ASD group showed lower activity in premotor/motor areas during pre-neurofeedback run than controls, but not in transfer runs, where it was seemingly balanced by higher connectivity between the DLPFC and the motor cortex. Group comparison in the transfer run also showed significant differences in DLPFC-based connectivity between groups, including higher connectivity with areas integrated into the multidemand network (MDN) and the visual cortex.

CONCLUSIONS

Neurofeedback seems to induce a higher between-group similarity of the whole-brain activity levels (including the target ROI) which might be promoted by changes in connectivity between the DLPFC and both high and low-level areas, including motor, visual and MDN regions.

Functional and structural connectivity success predictors of real-time fMRI neurofeedback targeting DLPFC: Contributions from central executive, salience, and default mode networks

Real-time functional magnetic resonance imaging (rt-fMRI) neurofeedback (NF), a training method for the self-regulation of brain activity, has shown promising results as a neurorehabilitation tool, depending on the ability of the patient to succeed in neuromodulation. This study explores connectivity-based structural and functional success predictors in an NF n-back working memory paradigm targeting the dorsolateral prefrontal cortex (DLPFC). We established as the NF success metric the linear trend on the ability to modulate the target region during NF runs and performed a linear regression model considering structural and functional connectivity (intrinsic and seed-based) metrics. We found a positive correlation between NF success and the default mode network (DMN) intrinsic functional connectivity and a negative correlation with the DLPFC-precuneus connectivity during the 2-back condition, indicating that success is associated with larger uncoupling between DMN and the executive network. Regarding structural connectivity, the salience network emerges as the main contributor to success. Both functional and structural classification models showed good performance with 77% and 86% accuracy, respectively. Dynamic switching between DMN, salience network and central executive network seems to be the key for neurofeedback success, independently indicated by functional connectivity on the localizer run and structural connectivity data.

Design and Implementation of a Collaborative Clinical Practice and Research Documentation System Using SNOMED-CT and HL7-CDA in the Context of a Pediatric Neurodevelopmental Unit

This paper introduces a prototype for clinical research documentation using the structured information model HL7 CDA and clinical terminology (SNOMED CT). The proposed solution was integrated with the current electronic health record system (EHR-S) and aimed to implement interoperability and structure information, and to create a collaborative platform between clinical and research teams. The framework also aims to overcome the limitations imposed by classical documentation strategies in real-time healthcare encounters that may require fast access to complex information. The solution was developed in the pediatric hospital (HP) of the University Hospital Center of Coimbra (CHUC), a national reference for neurodevelopmental disorders, particularly for autism spectrum disorder (ASD), which is very demanding in terms of longitudinal and cross-sectional data throughput. The platform uses a three-layer approach to reduce components’ dependencies and facilitate maintenance, scalability, and security. The system was validated in a real-life context of the neurodevelopmental and autism unit (UNDA) in the HP and assessed based on the functionalities model of EHR-S (EHR-S FM) regarding their successful implementation and comparison with state-of-the-art alternative platforms. A global approach to the clinical history of neurodevelopmental disorders was worked out, providing transparent healthcare data coding and structuring while preserving information quality. Thus, the platform enabled the development of user-defined structured templates and the creation of structured documents with standardized clinical terminology that can be used in many healthcare contexts. Moreover, storing structured data associated with healthcare encounters supports a longitudinal view of the patient’s healthcare data and health status over time, which is critical in routine and pediatric research contexts. Additionally, it enables queries on population statistics that are key to supporting the definition of local and global policies, whose importance was recently emphasized by the COVID pandemic.

Neurofeedback-dependent influence of the ventral striatum using a working memory paradigm targeting the dorsolateral prefrontal cortex

Executive functions and motivation have been established as key aspects for neurofeedback success. However, task-specific influence of cognitive strategies is scarcely explored. In this study, we test the ability to modulate the dorsolateral prefrontal cortex, a strong candidate for clinical application of neurofeedback in several disorders with dysexecutive syndrome, and investigate how feedback contributes to better performance in a single session. Participants of both neurofeedback (n = 17) and sham-control (n = 10) groups were able to modulate DLPFC in most runs (with or without feedback) while performing a working memory imagery task. However, activity in the target area was higher and more sustained in the active group when receiving feedback. Furthermore, we found increased activity in the nucleus accumbens in the active group, compared with a predominantly negative response along the block in participants receiving sham feedback. Moreover, they acknowledged the non-contingency between imagery and feedback, reflecting the impact on motivation. This study reinforces DLPFC as a robust target for neurofeedback clinical implementations and enhances the critical influence of the ventral striatum, both poised to achieve success in the self-regulation of brain activity.

Assessing MR-compatibility of somatosensory stimulation devices: A systematic review on testing methodologies

Functional magnetic resonance imaging (fMRI) has been extensively used as a tool to map the brain processes related to somatosensory stimulation. This mapping includes the localization of task-related brain activation and the characterization of brain activity dynamics and neural circuitries related to the processing of somatosensory information. However, the magnetic resonance (MR) environment presents unique challenges regarding participant and equipment safety and compatibility. This study aims to systematically review and analyze the state-of-the-art methodologies to assess the safety and compatibility of somatosensory stimulation devices in the MR environment. A literature search, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement guidelines, was performed in PubMed, Scopus, and Web of Science to find original research on the development and testing of devices for somatosensory stimulation in the MR environment. Nineteen records that complied with the inclusion and eligibility criteria were considered. The findings are discussed in the context of the existing international standards available for the safety and compatibility assessment of devices intended to be used in the MR environment. In sum, the results provided evidence for a lack of uniformity in the applied testing methodologies, as well as an in-depth presentation of the testing methodologies and results. Lastly, we suggest an assessment methodology (safety, compatibility, performance, and user acceptability) that can be applied to devices intended to be used in the MR environment.

Systematic review registration

https://www.crd.york.ac.uk/prospero/, identifier CRD42021257838.

Training the social brain: Clinical and neural effects of an 8-week real-time functional magnetic resonance imaging neurofeedback Phase IIa Clinical Trial in Autism

LAY ABSTRACT

Neurofeedback is an emerging therapeutic approach in neuropsychiatric disorders. Its potential application in autism spectrum disorder remains to be tested. Here, we demonstrate the feasibility of real-time functional magnetic resonance imaging volitional neurofeedback in targeting social brain regions in autism spectrum disorder. In this clinical trial, autism spectrum disorder patients were enrolled in a program with five training sessions of neurofeedback. Participants were able to control their own brain activity in this social brain region, with positive clinical and neural effects. Larger, controlled, and blinded clinical studies will be required to confirm the benefits.

Directly Exploring the Neural Correlates of Feedback-Related Reward Saliency and Valence During Real-Time fMRI-Based Neurofeedback

Introduction: The potential therapeutic efficacy of real-time fMRI Neurofeedback has received increasing attention in a variety of psychological and neurological disorders and as a tool to probe cognition. Despite its growing popularity, the success rate varies significantly, and the underlying neural mechanisms are still a matter of debate. The question whether an individually tailored framework positively influences neurofeedback success remains largely unexplored.

Methods: To address this question, participants were trained to modulate the activity of a target brain region, the visual motion area hMT+/V5, based on the performance of three imagery tasks with increasing complexity: imagery of a static dot, imagery of a moving dot with two and with four opposite directions. Participants received auditory feedback in the form of vocalizations with either negative, neutral or positive valence. The modulation thresholds were defined for each participant according to the maximum BOLD signal change of their target region during the localizer run.

Results: We found that 4 out of 10 participants were able to modulate brain activity in this region-of-interest during neurofeedback training. This rate of success (40%) is consistent with the neurofeedback literature. Whole-brain analysis revealed the recruitment of specific cortical regions involved in cognitive control, reward monitoring, and feedback processing during neurofeedback training. Individually tailored feedback thresholds did not correlate with the success level. We found region-dependent neuromodulation profiles associated with task complexity and feedback valence.

Discussion: Findings support the strategic role of task complexity and feedback valence on the modulation of the network nodes involved in monitoring and feedback control, key variables in neurofeedback frameworks optimization. Considering the elaborate design, the small sample size here tested (N = 10) impairs external validity in comparison to our previous studies. Future work will address this limitation. Ultimately, our results contribute to the discussion of individually tailored solutions, and justify further investigation concerning volitional control over brain activity.

BCIAUT-P300: A Multi-Session and Multi-Subject Benchmark Dataset on Autism for P300-Based Brain-Computer-Interfaces

There is a lack of multi-session P300 datasets for Brain-Computer Interfaces (BCI). Publicly available datasets are usually limited by small number of participants with few BCI sessions. In this sense, the lack of large, comprehensive datasets with various individuals and multiple sessions has limited advances in the development of more effective data processing and analysis methods for BCI systems. This is particularly evident to explore the feasibility of deep learning methods that require large datasets. Here we present the BCIAUT-P300 dataset, containing 15 autism spectrum disorder individuals undergoing 7 sessions of P300-based BCI joint-attention training, for a total of 105 sessions. The dataset was used for the 2019 IFMBE Scientific Challenge organized during MEDICON 2019 where, in two phases, teams from all over the world tried to achieve the best possible object-detection accuracy based on the P300 signals. This paper presents the characteristics of the dataset and the approaches followed by the 9 finalist teams during the competition. The winner obtained an average accuracy of 92.3% with a convolutional neural network based on EEGNet. The dataset is now publicly released and stands as a benchmark for future P300-based BCI algorithms based on multiple session data.

Volitional Modulation of the Left DLPFC Neural Activity Based on a Pain Empathy Paradigm-A Potential Novel Therapeutic Target for Pain

The ability to perceive and feel another person' pain as if it were one's own pain, e.g., pain empathy, is related to brain activity in the "pain-matrix" network. A non-core region of this network in Dorsolateral Prefrontal Cortex (DLPFC) has been suggested as a modulator of the attentional-cognitive dimensions of pain processing in the context of pain empathy. We conducted a neurofeedback experiment using real-time functional magnetic resonance imaging (rt-fMRI-NF) to investigate the association between activity in the left DLPFC (our neurofeedback target area) and the perspective assumed by the participant ("first-person"/"Self" or "third-person"/"Other" perspective of a pain-inducing stimulus), based on a customized pain empathy task. Our main goals were to assess the participants' ability to volitionally modulate activity in their own DLPFC through an imagery task of pain empathy and to investigate into which extent this ability depends on feedback. Our results demonstrate participants' ability to significantly modulate brain activity of the neurofeedback target area for the "first-person"/"Self" and "third-person"/"Other" perspectives. Results of both perspectives show that the participants were able to modulate (with statistical significance) the activity already in the first run of the session, in spite of being naïve to the task and even in the absence of feedback information. Moreover, they improved modulation throughout the session, particularly in the "Self" perspective. These results provide new insights on the role of DLPFC in pain and pain empathy mechanisms and validate the proposed protocol, paving the way for future interventional studies in clinical populations with empathic deficits.

The boundaries of state-space Granger causality analysis applied to BOLD simulated data: A comparative modelling and simulation approach

BACKGROUND

The analysis of connectivity has become a fundamental tool in human neuroscience. Granger Causality Mapping is a data-driven method that uses Granger Causality (GC) to assess the existence and direction of influence between signals, based on temporal precedence of information. More recently, a theory of Granger causality has been developed for state-space (SS-GC) processes, but little is known about its statistical validation and application on functional magnetic resonance imaging (fMRI) data.

NEW METHOD

We explored different multivariate computational frameworks to define the optimal combination for GC estimation. We hypothesized a new heuristic, combining SS-GC with a distinct statistical validation technique, Time Reversed Testing, validating it on synthetic data. We test its performance with a number of experimental parameters, including block structure, sampling frequency, noise and system mean pairwise correlation, using a statistical framework of binary classification.

RESULTS

We found that SS-GC with time reversed testing outperforms other frameworks. The results validate the application of SS-GC to generative models. When estimating reliable causal relations, SS-GC returns promising results, especially when considering synthetic data with a high impact of noise and sampling rate.

CONCLUSIONS

In this study, we empirically explored the boundaries of SS-GC with time reversed testing, a data-driven causality analysis framework with potential applicability to fMRI data.

Targeting dynamic facial processing mechanisms in superior temporal sulcus using a novel fMRI neurofeedback target

The superior temporal sulcus (STS) encompasses a complex set of regions involved in a wide range of cognitive functions. To understand its functional properties, neuromodulation approaches such brain stimulation or neurofeedback can be used. We investigated whether the posterior STS (pSTS), a core region in the face perception and imagery network, could be specifically identified based on the presence of dynamic facial expressions (and not just on simple motion or static face signals), and probed with neurofeedback. Recognition of facial expressions is critically impaired in autism spectrum disorder, making this region a relevant target for future clinical neurofeedback studies. We used a stringent localizer approach based on the contrast of dynamic facial expressions against static neutral faces plus moving dots. The target region had to be specifically responsive to dynamic facial expressions instead of mere motion and/or the presence of a static face. The localizer was successful in selecting this region across subjects. Neurofeedback was then performed, using this region as a target, with two novel feedback rules (mean or derivative-based, using visual or auditory interfaces). Our results provide evidence that a facial expression-selective cluster in pSTS can be identified and may represent a suitable target for neurofeedback approaches, aiming at social and emotional cognition. These findings highlight the presence of a highly selective region in STS encoding dynamic aspects of facial expressions. Future studies should elucidate its role as a mechanistic target for neurofeedback strategies in clinical disorders of social cognition such as autism.

Correlated alpha activity with the facial expression processing network in a simultaneous EEG-fMRI experiment

The relationship between EEG and fMRI data is poorly covered in the literature. Extensive work has been conducted in resting-state and epileptic activity, highlighting a negative correlation between the alpha power band of the EEG and the BOLD activity in the default-mode-network. The identification of an appropriate task-specific relationship between fMRI and EEG data for predefined regions-of-interest, would allow the transfer of interventional paradigms (such as BOLD-based neurofeedback sessions) from fMRI to EEG, enhancing its application range by lowering its costs and improving its flexibility. In this study, we present an analysis of the correlation between task-specific alpha band fluctuations and BOLD activity in the facial expressions processing network. We characterized the network ROIs through a stringent localizer and identified two clusters on the scalp (one frontal, one parietal-occipital) with marked alpha fluctuations, related to the task. We then check whether such power variations throughout the time correlate with the BOLD activity in the network. Our results show statistically significant negative correlations between the alpha power in both clusters and for all the ROIs of the network. The correlation levels have still not met the requirements for transferring the protocol to an EEG setup, but they pave the way towards a better understand on how frontal and parietal-occipital alpha relates to the activity of the facial expressions processing network.

A Realistic Seizure Prediction Study Based on Multiclass SVM

A patient-specific algorithm, for epileptic seizure prediction, based on multiclass support-vector machines (SVM) and using multi-channel high-dimensional feature sets, is presented. The feature sets, combined with multiclass classification and post-processing schemes aim at the generation of alarms and reduced influence of false positives. This study considers 216 patients from the European Epilepsy Database, and includes 185 patients with scalp EEG recordings and 31 with intracranial data. The strategy was tested over a total of 16,729.80[Formula: see text]h of inter-ictal data, including 1206 seizures. We found an overall sensitivity of 38.47% and a false positive rate per hour of 0.20. The performance of the method achieved statistical significance in 24 patients (11% of the patients). Despite the encouraging results previously reported in specific datasets, the prospective demonstration on long-term EEG recording has been limited. Our study presents a prospective analysis of a large heterogeneous, multicentric dataset. The statistical framework based on conservative assumptions, reflects a realistic approach compared to constrained datasets, and/or in-sample evaluations. The improvement of these results, with the definition of an appropriate set of features able to improve the distinction between the pre-ictal and nonpre-ictal states, hence minimizing the effect of confounding variables, remains a key aspect.

Control of Brain Activity in hMT+/V5 at Three Response Levels Using fMRI-Based Neurofeedback/BCI

A major challenge in brain-computer interface (BCI) research is to increase the number of command classes and levels of control. BCI studies often use binary control level approaches (level 0 and 1 of brain activation for each class of control). Different classes may often be achieved but not different levels of activation for the same class. The increase in the number of levels of control in BCI applications may allow for larger efficiency in neurofeedback applications. In this work we test the hypothesis whether more than two modulation levels can be achieved in a single brain region, the hMT+/V5 complex. Participants performed three distinct imagery tasks during neurofeedback training: imagery of a stationary dot, imagery of a dot with two opposing motions in the vertical axis and imagery of a dot with four opposing motions in vertical or horizontal axes (imagery of 2 or 4 motion directions). The larger the number of motion alternations, the higher the expected hMT+/V5 response. A substantial number (17 of 20) of participants achieved successful binary level of control and 12 were able to reach even 3 significant levels of control within the same session, confirming the whole group effects at the individual level. With this simple approach we suggest that it is possible to design a parametric system of control based on activity modulation of a specific brain region with at least 3 different levels. Furthermore, we show that particular imagery task instructions, based on different number of motion alternations, provide feasible achievement of different control levels in BCI and/or neurofeedback applications.

Feature analysis for correlation studies of simultaneous EEG-fMRI data: A proof of concept for neurofeedback approaches

The identification and interpretation of facial expressions is an important feature of social cognition. This characteristic is often impaired in various neurodevelopmental disorders. Recent therapeutic approaches to intervene in social communication impairments include neurofeedback (NF). In this study, we present a NF real-time functional Magnetic Resonance Imaging (rt-fMRI), combined with electroencephalography (EEG) to train social communication skills. In this sense, we defined the right Superior Temporal Sulcus as our target region-of-interest. To analyze the correlation between the fMRI regions of interest and the EEG data, we transposed the sources located at the nearest cortical location to the target region. We extracted a set of 75 features from EEG segments and performed a correlation analysis with the brain activations extracted from rt-fMRI in the right pSTS region. The finding of significant correlations of simultaneously measured signals in distinct modalities (EEG and fMRI) is promising. Future studies should address whether the observed correlation levels between local brain activity and scalp measures are sufficient to implement NF approaches.

Epileptic seizure predictors based on computational intelligence techniques: a comparative study with 278 patients

The ability of computational intelligence methods to predict epileptic seizures is evaluated in long-term EEG recordings of 278 patients suffering from pharmaco-resistant partial epilepsy, also known as refractory epilepsy. This extensive study in seizure prediction considers the 278 patients from the European Epilepsy Database, collected in three epilepsy centres: Hôpital Pitié-là-Salpêtrière, Paris, France; Universitätsklinikum Freiburg, Germany; Centro Hospitalar e Universitário de Coimbra, Portugal. For a considerable number of patients it was possible to find a patient specific predictor with an acceptable performance, as for example predictors that anticipate at least half of the seizures with a rate of false alarms of no more than 1 in 6 h (0.15 h⁻¹). We observed that the epileptic focus localization, data sampling frequency, testing duration, number of seizures in testing, type of machine learning, and preictal time influence significantly the prediction performance. The results allow to face optimistically the feasibility of a patient specific prospective alarming system, based on machine learning techniques by considering the combination of several univariate (single-channel) electroencephalogram features. We envisage that this work will serve as benchmark data that will be of valuable importance for future studies based on the European Epilepsy Database.

Output regularization of SVM seizure predictors: Kalman Filter versus the "Firing Power" method

Two methods for output regularization of support vector machines (SVMs) classifiers were applied for seizure prediction in 10 patients with long-term annotated data. The output of the classifiers were regularized by two methods: one based on the Kalman Filter (KF) and other based on a measure called the "Firing Power" (FP). The FP is a quantification of the rate of the classification in the preictal class in a past time window. In order to enable the application of the KF, the classification problem was subdivided in a two two-class problem, and the real-valued output of SVMs was considered. The results point that the FP method raise less false alarms than the KF approach. However, the KF approach presents an higher sensitivity, but the high number of false alarms turns their applicability negligible in some situations.

Epileptic seizure prediction based on a bivariate spectral power methodology

The spectral power of 5 frequently considered frequency bands (Alpha, Beta, Gamma, Theta and Delta) for 6 EEG channels is computed and then all the possible pairwise combinations among the 30 features set, are used to create a 435 dimensional feature space. Two new feature selection methods are introduced to choose the best candidate features among those and to reduce the dimensionality of this feature space. The selected features are then fed to Support Vector Machines (SVMs) that classify the cerebral state in preictal and non-preictal classes. The outputs of the SVM are regularized using a method that accounts for the classification dynamics of the preictal class, also known as "Firing Power" method. The results obtained using our feature selection approaches are compared with the ones obtained using minimum Redundancy Maximum Relevance (mRMR) feature selection method. The results in a group of 12 patients of the EPILEPSIAE database, containing 46 seizures and 787 hours multichannel recording for out-of-sample data, indicate the efficiency of the bivariate approach as well as the two new feature selection methods. The best results presented sensitivity of 76.09% (35 of 46 seizures predicted) and a false prediction rate of 0.15(-1).

Space time frequency (STF) code tensor for the characterization of the epileptic preictal stage

We evaluate the ability of multiway models to characterize the epileptic preictal period. The understanding of the characteristics of the period prior to the seizure onset is a decisive step towards the development of seizure prediction frameworks. Multiway models of EEG segments already demonstrated that hidden structures may be unveiled using tensor decomposition techniques. We propose a novel approach using a multiway model, Parallel Factor Analysis (PARAFAC), to identify spatial, temporal and spectral signatures of the preictal period. The results obtained, from a dataset of 4 patients, with a total of 30 seizures, suggest that a common structure may be involved in seizure generation. Furthermore, the spatial signature may be related to the ictal onset region and that determined frequency sub-bands may be more relevant in preictal stages.

EPILAB: a software package for studies on the prediction of epileptic seizures

A Matlab®-based software package, EPILAB, was developed for supporting researchers in performing studies on the prediction of epileptic seizures. It provides an intuitive and convenient graphical user interface. Fundamental concepts that are crucial for epileptic seizure prediction studies were implemented. This includes, for example, the development and statistical validation of prediction methodologies in long-term continuous recordings. Seizure prediction is usually based on electroencephalography (EEG) and electrocardiography (ECG) signals. EPILAB is able to process both EEG and ECG data stored in different formats. More than 35 time and frequency domain measures (features) can be extracted based on univariate and multivariate data analysis. These features can be post-processed and used for prediction purposes. The predictions may be conducted based on optimized thresholds or by applying classifications methods such as artificial neural networks, cellular neuronal networks, and support vector machines. EPILAB proved to be an efficient tool for seizure prediction, and aims to be a way to communicate, evaluate, and compare results and data among the seizure prediction community.

Optimized feature subsets for epileptic seizure prediction studies

The reduction of the number of EEG features to give as inputs to epilepsy seizure predictors is a needed step towards the development of a transportable device for real-time warning. This paper presents a comparative study of three feature selection methods, based on Support Vector Machines. Minimum-Redundancy Maximum-Relevance, Recursive Feature Elimination, Genetic Algorithms, show that, for three patients of the European Database on Epilepsy, the most important univariate features are related to spectral information and statistical moments.

A computational environment for long-term multi-feature and multi-algorithm seizure prediction

The daily life of epilepsy patients is constrained by the possibility of occurrence of seizures. Until now, seizures cannot be predicted with sufficient sensitivity and specificity. Most of the seizure prediction studies have been focused on a small number of patients, and frequently assuming unrealistic hypothesis. This paper adopts the view that for an appropriate development of reliable predictors one should consider long-term recordings and several features and algorithms integrated in one software tool. A computational environment, based on Matlab (®), is presented, aiming to be an innovative tool for seizure prediction. It results from the need of a powerful and flexible tool for long-term EEG/ECG analysis by multiple features and algorithms. After being extracted, features can be subjected to several reduction and selection methods, and then used for prediction. The predictions can be conducted based on optimized thresholds or by applying computational intelligence methods. One important aspect is the integrated evaluation of the seizure prediction characteristic of the developed predictors.