Recursive independent component analysis (ICA)-decomposition of ictal EEG to select the best ictal component for EEG source imaging

Ictal EEG Source Imaging With Supplemental Electrodes

INTRODUCTION

Noninvasive brain imaging tests play a major role in guiding decision-making and the usage of invasive, costly intracranial electroencephalogram (ICEEG) in the presurgical epilepsy evaluation. This study prospectively examined the concordance in localization between ictal EEG source imaging (ESI) and ICEEG as a reference standard.

METHODS

Between August 2014 and April 2019, patients during video monitoring with scalp EEG were screened for those with intractable focal epilepsy believed to be amenable to surgical treatment. Additional 10-10 electrodes (total = 31-38 per patient, "31+") were placed over suspected regions of seizure onset in 104 patients. Of 42 patients requiring ICEEG, 30 (mean age 30, range 19-59) had sufficiently localized subsequent intracranial studies to allow comparison of localization between tests. ESI was performed using realistic forward boundary element models used in dipole and distributed source analyses.

RESULTS

At least partial sublobar concordance between ESI and ICEEG solutions was obtained in 97% of cases, with 73% achieving complete agreement. Median Euclidean distances between ESI and ICEEG solutions ranged from 25 to 30 mm (dipole) and 23 to 38 mm (distributed source). The latter was significantly more accurate with 31+ compared with 21 electrodes (P < 0.01). A difference of ≤25 mm was present in two thirds of the cases. No significant difference was found between dipole and distributed source analyses.

CONCLUSIONS

A practical method of ictal ESI (nonuniform placement of 31-38 electrodes) yields high accuracy for seizure localization in epilepsy surgery candidates. These results support routine clinical application of ESI in the presurgical evaluation.

Estimation of seizure onset zone from ictal scalp EEG using independent component analysis in extratemporal lobe epilepsy

OBJECTIVE

The purpose of this study is to localize the seizure onset zone of patients suffering from drug-resistant epilepsy. During the last two decades, multiple studies proposed the use of Independent Component Analysis (ICA) to analyze ictal electroencephalogram (EEG) recordings. This study aims at evaluating ICA potential with quantitative measurements. In particular, we address the challenging step where the components extracted by ICA of an ictal nature must be selected.

APPROACH

We considered a cohort of 10 patients suffering from extratemporal lobe epilepsy who were rendered seizure-free after surgery. Different sets of pre-processing parameters were compared and component features were explored to help distinguish ictal components from others. Quantitative measurements were implemented to determine whether some of the components returned by ICA were located within the resection zone and thus likely to be ictal. Finally, an assistance to the component selection was proposed based on the implemented features.

MAIN RESULTS

For every seizure, at least one component returned by ICA was localized within the resection zone, with the optimal pre-processing parameters. Three features were found to distinguish components localized within the resection zone: the dispersion of their active brain sources, the ictal rhythm power and the contribution to the EEG variance. Using the implemented component selection assistance based on the features, the probability that the first proposed component yields an accurate estimation reaches 51.43% (without assistance: 24.74%). The accuracy reaches 80% when considering the best result within the first five components.

SIGNIFICANCE

This study confirms the utility of ICA for ictal EEG analysis in extratemporal lobe epilepsy, and suggests relevant features to analyze the components returned by ICA. A component selection assistance is proposed to guide clinicians in their choice for ictal components.

Linear distributed inverse solutions for interictal EEG source localisation

OBJECTIVE

To compare the spatial accuracy of 6 linear distributed inverse solutions for EEG source localisation of interictal epileptic discharges: Minimum Norm, Weighted Minimum Norm, Low-Resolution Electromagnetic Tomography (LORETA), Local Autoregressive Average (LAURA), Standardised LORETA, and Exact LORETA.

METHODS

Spatial accuracy was assessed clinically by retrospectively comparing the maximum source of averaged interictal discharges to the resected brain area in 30 patients with successful epilepsy surgery, based on 204-channel EEG. Additionally, localisation errors of the inverse solutions were assessed in computer simulations, with different levels of noise added to the signal in both sensor space and source space.

RESULTS

In the clinical evaluations, the source maximum was located inside the resected brain area in 50-57% of patients when using LORETA or LAURA, while all other inverse solutions performed significantly worse (17-30%; corrected p < 0.01). In the simulation studies, when noise levels exceeded 10%, LORETA and LAURA had substantially smaller localisation errors than the other inverse solutions.

CONCLUSIONS

LORETA and LAURA provided the highest spatial accuracy both in clinical and simulated data, alongside with a comparably high robustness towards noise.

SIGNIFICANCE

Among the different linear inverse solution algorithms tested, LORETA and LAURA might be preferred for interictal EEG source localisation.

Improving surgical outcome with electric source imaging and high field magnetic resonance imaging

While most patients with focal epilepsy present with clear structural abnormalities on standard, 1.5 or 3 T MRI, some patients are MRI-negative. For those, quantitative MRI techniques, such as volumetry, voxel-based morphometry, and relaxation time measurements can aid in finding the epileptogenic focus. High-field MRI, just recently approved for clinical use by the FDA, increases the resolution and, in several publications, was shown to improve the detection of focal cortical dysplasias and mild cortical malformations. For those cases without any tissue abnormality in neuroimaging, even at 7 T, scalp EEG alone is insufficient to delimitate the epileptogenic zone. They may benefit from the use of high-density EEG, in which the increased number of electrodes helps improve spatial sampling. The spatial resolution of even low-density EEG can benefit from electric source imaging techniques, which map the source of the recorded abnormal activity, such as interictal epileptiform discharges, focal slowing, and ictal rhythm. These EEG techniques help localize the irritative, functional deficit, and seizure-onset zone, to better estimate the epileptogenic zone. Combining those technologies allows several drug-resistant cases to be submitted to surgery, increasing the odds of seizure freedom and providing a must needed hope for patients with epilepsy.

Prediction of protein-protein interaction sites through eXtreme gradient boosting with kernel principal component analysis

Predicting protein-protein interaction sites (PPI sites) can provide important clues for understanding biological activity. Using machine learning to predict PPI sites can mitigate the cost of running expensive and time-consuming biological experiments. Here we propose PPISP-XGBoost, a novel PPI sites prediction method based on eXtreme gradient boosting (XGBoost). First, the characteristic information of protein is extracted through the pseudo-position specific scoring matrix (PsePSSM), pseudo-amino acid composition (PseAAC), hydropathy index and solvent accessible surface area (ASA) under the sliding window. Next, these raw features are preprocessed to obtain more optimal representations in order to achieve better prediction. In particular, the synthetic minority oversampling technique (SMOTE) is used to circumvent class imbalance, and the kernel principal component analysis (KPCA) is applied to remove redundant characteristics. Finally, these optimal features are fed to the XGBoost classifier to identify PPI sites. Using PPISP-XGBoost, the prediction accuracy on the training dataset Dset186 reaches 85.4%, and the accuracy on the independent validation datasets Dtestset72, PDBtestset164, Dset_448 and Dset_355 reaches 85.3%, 83.9%, 85.8% and 85.4%, respectively, which all show an increase in accuracy against existing PPI sites prediction methods. These results demonstrate that the PPISP-XGBoost method can further enhance the prediction of PPI sites.

Ictal EEG source localization in focal epilepsy: Review and future perspectives

Electroencephalographic (EEG) source imaging localizes the generators of neural activity in the brain. During presurgical epilepsy evaluation, EEG source imaging of interictal epileptiform discharges is an established tool to estimate the irritative zone. However, the origin of interictal activity can be partly or fully discordant with the origin of seizures. Therefore, source imaging based on ictal EEG data to determine the seizure onset zone can provide precious clinical information. In this descriptive review, we address the importance of localizing the seizure onset zone based on noninvasive EEG recordings as a complementary analysis that might reduce the burden of the presurgical evaluation. We identify three major challenges (low signal-to-noise ratio of the ictal EEG data, spread of ictal activity in the brain, and validation of the developed methods) and discuss practical solutions. We provide an extensive overview of the existing clinical studies to illustrate the potential clinical utility of EEG-based localization of the seizure onset zone. Finally, we conclude with future perspectives and the needs for translating ictal EEG source imaging into clinical practice.

EEG Source Imaging (ESI) utility in clinical practice

Epilepsy surgery is an important treatment modality for medically refractory focal epilepsy. The outcome of surgery usually depends on the localization accuracy of the epileptogenic zone (EZ) during pre-surgical evaluation. Good localization can be achieved with various electrophysiological and neuroimaging approaches. However, each approach has its own merits and limitations. Electroencephalography (EEG) Source Imaging (ESI) is an emerging model-based computational technique to localize cortical sources of electrical activity within the brain volume, three-dimensionally. ESI based pre-surgical evaluation gives an overall clinical yield of 73-91%, depending on choice of head model, inverse solution and EEG electrode density. It is a cost effective, non-invasive method which provides valuable additional information in presurgical evaluation due to its high localizing value specifically in MRI-negative cases, extra or basal temporal lobe epilepsy, multifocal lesions such as tuberous sclerosis or cases with multiple hypotheses. Unfortunately, less than 1% of surgical centers in developing countries use this method as a part of pre-surgical evaluation. This review promotes ESI as a useful clinical tool especially for patients with lesion-negative MRI to determine EZ cost-effectively with high accuracy under the optimized conditions.