Automatic detection of interictal ripples on scalp EEG to evaluate the effect and prognosis of ACTH therapy in patients with infantile spasms

Using scalp EEG to predict seizure recurrence and electrical status epilepticus in children with idiopathic focal epilepsy

PURPOSE

Some individuals with idiopathic focal epilepsy (IFE) experience recurring seizures accompanied by the evolution of electrical status epilepticus during sleep (ESES). Here, we aimed to develop a predictor for the early detection of seizure recurrence with ESES in children with IFE using resting state electroencephalogram (EEG) data.

METHODS

The study group included 15 IFE patients who developed seizure recurrence with ESES. There were 17 children in the control group who did not experience seizure recurrence with ESES during at least 2-year follow-up. We used the degree value of the partial directed coherence (PDC) from the EEG data to predict seizure recurrence with ESES via 6 machine learning (ML) algorithms.

RESULTS

Among the models, the Xgboost Classifier (XGBC) model achieved the highest specificity of 0.90, and a remarkable sensitivity and accuracy of 0.80 and 0.85, respectively. The CATC showed balanced performance with a specificity of 0.85, sensitivity of 0.73, and an accuracy of 0.80, with an AUC equal to 0.78. For both of these models, F4, Fz and T4 were the overlaps of the top 4 features.

CONCLUSIONS

Considering its high classification accuracy, the XGBC model is an effective and quantitative tool for predicting seizure recurrence with ESES evolution in IFE patients. We developed an ML-based tool for predicting the development of IFE using resting state EEG data. This could facilitate the diagnosis and treatment of patients with IFE.

Automatic Detection of Scalp High-Frequency Oscillations Based on Deep Learning

Scalp high-frequency oscillations (sHFOs) are a promising non-invasive biomarker of epilepsy. However, the visual marking of sHFOs is a time-consuming and subjective process, existing automatic detectors based on single-dimensional analysis have difficulty with accurately eliminating artifacts and thus do not provide sufficient reliability to meet clinical needs. Therefore, we propose a high-performance sHFOs detector based on a deep learning algorithm. An initial detection module was designed to extract candidate high-frequency oscillations. Then, one-dimensional (1D) and two-dimensional (2D) deep learning models were designed, respectively. Finally, the weighted voting method is used to combine the outputs of the two model. In experiments, the precision, recall, specificity and F1-score were 83.44%, 83.60%, 96.61% and 83.42%, respectively, on average and the kappa coefficient was 80.02%. In addition, the proposed detector showed a stable performance on multi-centre datasets. Our sHFOs detector demonstrated high robustness and generalisation ability, which indicates its potential applicability as a clinical assistance tool. The proposed sHFOs detector achieves an accurate and robust method via deep learning algorithm.

Computational EEG attributes predict response to therapy for epileptic spasms

OBJECTIVE

We set out to evaluate whether response to treatment for epileptic spasms is associated with specific candidate computational EEG biomarkers, independent of clinical attributes.

METHODS

We identified 50 children with epileptic spasms, with pre- and post-treatment overnight video-EEG. After EEG samples were preprocessed in an automated fashion to remove artifacts, we calculated amplitude, power spectrum, functional connectivity, entropy, and long-range temporal correlations (LRTCs). To evaluate the extent to which each feature is independently associated with response and relapse, we conducted logistic and proportional hazards regression, respectively.

RESULTS

After statistical adjustment for the duration of epileptic spasms prior to treatment, we observed an association between response and stronger baseline and post-treatment LRTCs (P = 0.042 and P = 0.004, respectively), and higher post-treatment entropy (P = 0.003). On an exploratory basis, freedom from relapse was associated with stronger post-treatment LRTCs (P = 0.006) and higher post-treatment entropy (P = 0.044).

CONCLUSION

This study suggests that multiple EEG features-especially LRTCs and entropy-may predict response and relapse.

SIGNIFICANCE

This study represents a step toward a more precise approach to measure and predict response to treatment for epileptic spasms.

A Potential Mechanism of Neurological Impairment in Children With Infantile Spasm: Based on Microanatomic Structure Analysis Employing Voxel-Based Morphometry and Surface-Based Morphometry

BACKGROUND

Infantile epileptic spasms syndrome (IESS) would accompany with severe neurological impairment. Our study aimed to explore the potential mechanism by employing voxel-based and surface-based morphometry to detect brain microwould accompany with severe neurological impairment. Our study aimed to explore the potential mechanism by employing voxel-based and surface-based morphometry to detect brain microanatomic structure alteration.

METHODS

The IESS group had 21 males and 13 females (mean age: 17.7 ± 15.6 months), whereas the healthy controls group had 22 males and 10 females (mean age: 29.4 ± 18.7 months). High-resolution 3D T1WI was performed. Computational Anatomy Toolbox implemented in Statistical Parametric Mapping 12 was used to measure the gray matter and white matter volume, and the cortical thickness separately. Independent sample t test was used to assess between-group differences. IESS group was assessed using the Bayley Scales of Infant Development.

RESULTS

The IESS group showed a significantly decreased volume of gray matter in right middle temporal gyrus, inferior temporal gyrus, superior temporal gyrus, right fusiform, and bilateral precuneus (P < 0.001). There were no significant between-group differences with respect to white matter volume or cortical thickness (P > 0.001). The results of Bayley Scales of Infant Development showed that the Mental Development Index (MDI) and Psychomotor Development Index scores of children with IESS were almost concentrated in the range of <70. MDI score showed a positive correlation with gray matter reduction area in IESS group.

CONCLUSION

Children with IESS had impaired cognitive and delayed motor development. And the decreased gray matter in the right temporal lobe, fusiform, and bilateral precuneus could be the potential anatomic basis for impaired function, such as hearing, visual, and language.

Crucial involvement of fast waves and Delta band in the brain network attributes of infantile epileptic spasms syndrome

Objective

This study aims to describe the characteristics of the brain network attributes in children diagnosed with Infantile Epileptic Spasms Syndrome (IESS) and to determine the influence exerted by adrenocorticotrophic hormone (ACTH) or methylprednisolone (MP) on network attributes.

Methods

In this retrospective cohort study, we recruited 19 infants diagnosed with IESS and 10 healthy subjects as the control from the Pediatric Neurology Department at the Third Affiliated Hospital of Zhengzhou University between October 2019 and December 2020. The first thirty-minute processed electroencephalograms (EEGs) were clipped and filtered into EEG frequency bands (2 s each). A comparative assessment was conducted between the IESS group and the controls as well as the pre- and post-treatment in the IESS group. Mutual information values for each EEG channel were collected and compared including characteristic path length (CPL), node degree (ND), clustering coefficient (CC), and betweenness centrality (BC), based on graph theory.

Results

Comparing the control group, in the IESS group, there was an increase in CPL of the Delta band, and a decrease in ND and CC of the Delta band during the waking period, contrary to those during the sleeping period (P < 0.05), a decreased in CPL of the fast waves and an increase in ND and CC (P < 0.05) in the sleep-wake cycle, and a decrease in ND and CC of the Theta band in the waking phase. Post-treatment compared with the pre-treatment, during the waking ictal phase, there was a noted decrease in CPL in the Delta band and fast waves, while an increase was observed in ND and CC (P < 0.05).

Conclusions

The Delta band and fast waves are crucial components of the network attributes in IESS.

Significance

This investigation provides a precise characterization of the brain network in children afflicted with IESS, and lays the groundwork for predicting the prognosis using graph theory.

[Application of brain functional connectivity and nonlinear dynamic analysis in brain function assessment for infants with controlled infantile spasm]

OBJECTIVES

To investigate the role of brain functional connectivity and nonlinear dynamic analysis in brain function assessment for infants with controlled infantile spasm (IS).

METHODS

A retrospective analysis was performed on 14 children with controlled IS (IS group) who were admitted to the Department of Neurology, Anhui Provincial Children's Hospital, from January 2019 to January 2023. Twelve healthy children, matched for sex and age, were enrolled as the control group. Electroencephalogram (EEG) data were analyzed for both groups to compare the features of brain network, and nonlinear dynamic indicators were calculated, including approximate entropy, sample entropy, permutation entropy, and permutation Lempel-Ziv complexity.

RESULTS

Brain functional connectivity showed that compared with the control group, the IS group had an increase in the strength of functional connectivity, and there was a significant difference between the two groups in the connection strength between the Fp2 and F8 channels (P<0.05). The network stability analysis showed that the IS group had a significantly higher network stability than the control group at different time windows (P<0.05). The nonlinear dynamic analysis showed that compared with the control group, the IS group had a significantly lower sample entropy of Fz electrode (P<0.05).

CONCLUSIONS

Abnormalities in brain network and sample entropy may be observed in some children with controlled IS, and it is suggested that quantitative EEG analysis parameters can serve as neurological biomarkers for evaluating brain function in children with IS.

Prognostic interictal electroencephalographic biomarkers and models to assess antiseizure medication efficacy for clinical practice: A scoping review

Antiseizure medication (ASM) is the primary treatment for epilepsy. In clinical practice, methods to assess ASM efficacy (predict seizure freedom or seizure reduction), during any phase of the drug treatment lifecycle, are limited. This scoping review identifies and appraises prognostic electroencephalographic (EEG) biomarkers and prognostic models that use EEG features, which are associated with seizure outcomes following ASM initiation, dose adjustment, or withdrawal. We also aim to summarize the population and context in which these biomarkers and models were identified and described, to understand how they could be used in clinical practice. Between January 2021 and October 2022, four databases, references, and citations were systematically searched for ASM studies investigating changes to interictal EEG or prognostic models using EEG features and seizure outcomes. Study bias was appraised using modified Quality in Prognosis Studies criteria. Results were synthesized into a qualitative review. Of 875 studies identified, 93 were included. Biomarkers identified were classed as qualitative (visually identified by wave morphology) or quantitative. Qualitative biomarkers include identifying hypsarrhythmia, centrotemporal spikes, interictal epileptiform discharges (IED), classifying the EEG as normal/abnormal/epileptiform, and photoparoxysmal response. Quantitative biomarkers were statistics applied to IED, high-frequency activity, frequency band power, current source density estimates, pairwise statistical interdependence between EEG channels, and measures of complexity. Prognostic models using EEG features were Cox proportional hazards models and machine learning models. There is promise that some quantitative EEG biomarkers could be used to assess ASM efficacy, but further research is required. There is insufficient evidence to conclude any specific biomarker can be used for a particular population or context to prognosticate ASM efficacy. We identified a potential battery of prognostic EEG biomarkers, which could be combined with prognostic models to assess ASM efficacy. However, many confounders need to be addressed for translation into clinical practice.

Quantitative pretreatment EEG predicts efficacy of ACTH therapy in infantile epileptic spasms syndrome

OBJECTIVE

This study aimed to determine the correlation between outcomes following adrenocorticotrophic hormone (ACTH) therapy and measurements of relative power spectrum (rPS), weighted phase lag index (wPLI), and graph theoretical analysis on pretreatment electroencephalography (EEG) in infants with non-lesional infantile epileptic spasms syndrome (IESS).

METHODS

Twenty-eight patients with non-lesional IESS were enrolled. Outcomes were classified based on seizure recurrence following ACTH therapy: seizure-free (F, n = 21) and seizure-recurrence (R, n = 7) groups. The rPS, wPLI, clustering coefficient, and betweenness centrality were calculated on pretreatment EEG and were statistically analyzed to determine the correlation with outcomes following ACTH therapy.

RESULTS

The rPS value was significantly higher in the delta frequency band in group R than in group F (p < 0.001). The wPLI values were significantly higher in the delta, theta, and alpha frequency bands in group R than in group F (p = 0.007, <0.001, and <0.001, respectively). The clustering coefficient in the delta frequency band was significantly lower in group R than in group F (p < 0.001).

CONCLUSIONS

Our findings demonstrate the significant differences in power and functional connectivity between outcome groups.

SIGNIFICANCE

This study may contribute to an early prediction of ACTH therapy outcomes and thus help in the development of appropriate treatment strategies.

Integration of multiscale entropy and BASED scale of electroencephalography after adrenocorticotropic hormone therapy predict relapse of infantile spasms

BACKGROUND

Even though adrenocorticotropic hormone (ACTH) demonstrated powerful efficacy in the initially successful treatment of infantile spasms (IS), nearly half of patients have experienced a relapse. We sought to investigate whether features of electroencephalogram (EEG) predict relapse in those IS patients without structural brain abnormalities.

METHODS

We retrospectively reviewed data from children with IS who achieved initial response after ACTH treatment, along with EEG recorded within the last two days of treatment. The recurrence of epileptic spasms following treatment was tracked for 12 months. Subjects were categorized as either non-relapse or relapse groups. General clinical and EEG recordings were collected, burden of amplitudes and epileptiform discharges (BASED) score and multiscale entropy (MSE) were carefully explored for cross-group comparisons.

RESULTS

Forty-one patients were enrolled in the study, of which 26 (63.4%) experienced a relapse. The BASED score was significantly higher in the relapse group. MSE in the non-relapse group was significantly lower than the relapse group in the γ band but higher in the lower frequency range (δ, θ, α). Sensitivity and specificity were 85.71% and 92.31%, respectively, when combining MSE in the δ/γ frequency of the occipital region, plus BASED score were used to distinguish relapse from non-relapse groups.

CONCLUSIONS

BASED score and MSE of EEG after ACTH treatment could be used to predict relapse for IS patients without brain structural abnormalities. Patients with BASED score ≥ 3, MSE increased in higher frequency, and decreased in lower frequency had a high risk of relapse.

Automatic BASED scoring on scalp EEG in children with infantile spasms using convolutional neural network

In recent years, the Burden of Amplitudes and Epileptiform Discharges (BASED) score has been used as a reliable, accurate, and feasible electroencephalogram (EEG) grading scale for infantile spasms. However, manual EEG annotation is, in general, very time-consuming, and BASED scoring is no exception. Convolutional neural networks (CNNs) have proven their great potential in many EEG classification problems. However, very few research studies have focused on the use of CNNs for BASED scoring, a challenging but vital task in the diagnosis and treatment of infantile spasms. This study proposes an automatic BASED scoring framework using EEG and a deep CNN. The feasibility of using CNN for automatic BASED scoring was investigated in 36 patients with infantile spasms by annotating their long-term EEG data with four levels of the BASED score (scores 5, 4, 3, and ≤2). In the validation set, the accuracy was 96.9% by applying a multi-layer CNN to classify the EEG data as a 4-label problem. The extensive experiments have demonstrated that our proposed approach offers high accuracy and, hence, is an important step toward an automatic BASED scoring algorithm. To the best of our knowledge, this is the first attempt to use a CNN to construct a BASED-based scoring model.

High-Frequency Oscillations and Epileptogenic Network

Epilepsy is a network disease caused by aberrant neocortical large-scale connectivity spanning regions on the scale of several centimeters. High-frequency oscillations, characterized by the 80-600 Hz signals in electroencephalography, have been proven to be a promising biomarker of epilepsy that can be used in assessing the severity and susceptibility of epilepsy as well as the location of the epileptogenic zone. However, the presence of a high-frequency oscillation network remains a topic of debate as high-frequency oscillations have been previously thought to be incapable of propagation, and the relationship between high-frequency oscillations and the epileptogenic network has rarely been discussed. Some recent studies reported that high-frequency oscillations may behave like networks that are closely relevant to the epileptogenic network. Pathological highfrequency oscillations are network-driven phenomena and elucidate epileptogenic network development; high-frequency oscillations show different characteristics coincident with the epileptogenic network dynamics, and cross-frequency coupling between high-frequency oscillations and other signals may mediate the generation and propagation of abnormal discharges across the network.

EEG biomarkers for the diagnosis and treatment of infantile spasms

Early diagnosis and treatment are critical for young children with infantile spasms (IS), as this maximizes the possibility of the best possible child-specific outcome. However, there are major barriers to achieving this, including high rates of misdiagnosis or failure to recognize the seizures, medication failure, and relapse. There are currently no validated tools to aid clinicians in assessing objective diagnostic criteria, predicting or measuring medication response, or predicting the likelihood of relapse. However, the pivotal role of EEG in the clinical management of IS has prompted many recent studies of potential EEG biomarkers of the disease. These include both visual EEG biomarkers based on human visual interpretation of the EEG and computational EEG biomarkers in which computers calculate quantitative features of the EEG. Here, we review the literature on both types of biomarkers, organized based on the application (diagnosis, treatment response, prediction, etc.). Visual biomarkers include the assessment of hypsarrhythmia, epileptiform discharges, fast oscillations, and the Burden of AmplitudeS and Epileptiform Discharges (BASED) score. Computational markers include EEG amplitude and power spectrum, entropy, functional connectivity, high frequency oscillations (HFOs), long-range temporal correlations, and phase-amplitude coupling. We also introduce each of the computational measures and provide representative examples. Finally, we highlight remaining gaps in the literature, describe practical guidelines for future biomarker discovery and validation studies, and discuss remaining roadblocks to clinical implementation, with the goal of facilitating future work in this critical area.

Assessing Risk for Relapse among Children with Infantile Spasms Using the Based Score after ACTH Treatment: A Retrospective Study

Introduction

Even though adrenocorticotropic hormone (ACTH) demonstrated powerful efficacy in the initially successful treatment of infantile spasms (IS), nearly one-half of patients whose spasms were once suppressed experienced relapse. There is currently no validated method for the prediction of the risk of relapse. The Burden of Amplitudes and Epileptiform Discharges (BASED) score is an electroencephalogram (EEG) grading scale for children with infantile spasms. We sought to determine whether an association exists between the BASED score after ACTH treatment and relapse after initial response with ACTH.

Methods

Children with IS who achieved initial response after ACTH treatment were selected as the study subjects. Those who experienced relapse within 12 months after ACTH treatment were categorized as the relapse group, and those who did not were categorized as the non-relapse group. Their general clinical data and EEG data (using BASED scoring) after ACTH treatment were collected, and compared between groups. Cox proportional hazards models were fit to determine factors associated with relapse.

Results

A total of 64 children with IS were enrolled in the study, of which 37 (57.8%) experienced a relapse, and the median duration after ACTH treatment was 3 (1.5, 6) months. The BASED score was significantly higher in the relapse group than in the non-relapse group. Cox modeling demonstrated that BASED score was independently associated with relapse. The patients with a score greater than or equal to 3 showed a high rate (89.3%) of relapse. The relapse group had stronger, more stable EEG functional networks than the non-relapse group, and there were obvious correlations between BASED score and functional connectivity.

Conclusion

This study suggests the BASED score after ACTH treatment has potential value as a predictor for relapse after initial response. Children with IS who have a BASED score greater than or equal to 3 after the initial response of ACTH carry a high risk of relapse within 1 year.

Supplementary Information

The online version contains supplementary material available at 10.1007/s40120-022-00347-7.

Automatic Detection of High-Frequency Oscillations Based on an End-to-End Bi-Branch Neural Network and Clinical Cross-Validation

Accurate identification of high-frequency oscillation (HFO) is an important prerequisite for precise localization of epileptic foci and good prognosis of drug-refractory epilepsy. Exploring a high-performance automatic detection method for HFOs can effectively help clinicians reduce the error rate and reduce manpower. Due to the limited analysis perspective and simple model design, it is difficult to meet the requirements of clinical application by the existing methods. Therefore, an end-to-end bi-branch fusion model is proposed to automatically detect HFOs. With the filtered band-pass signal (signal branch) and time-frequency image (TFpic branch) as the input of the model, two backbone networks for deep feature extraction are established, respectively. Specifically, a hybrid model based on ResNet1d and long short-term memory (LSTM) is designed for signal branch, which can focus on both the features in time and space dimension, while a ResNet2d with a Convolutional Block Attention Module (CBAM) is constructed for TFpic branch, by which more attention is paid to useful information of TF images. Then the outputs of two branches are fused to realize end-to-end automatic identification of HFOs. Our method is verified on 5 patients with intractable epilepsy. In intravalidation, the proposed method obtained high sensitivity of 94.62%, specificity of 92.7%, and F1-score of 93.33%, and in cross-validation, our method achieved high sensitivity of 92.00%, specificity of 88.26%, and F1-score of 89.11% on average. The results show that the proposed method outperforms the existing detection paradigms of either single signal or single time-frequency diagram strategy. In addition, the average kappa coefficient of visual analysis and automatic detection results is 0.795. The method shows strong generalization ability and high degree of consistency with the gold standard meanwhile. Therefore, it has great potential to be a clinical assistant tool.

Scalp-HFO indexes are biomarkers for the lateralization and localization of the epileptogenic zone in preoperative assessment

During the noninvasive evaluation phase for refractory epilepsy, the localization of the epileptogenic zone (EZ) is essential for the surgical protocols. Confirmation of laterality is required when the preoperative evaluation limits the EZ to bilateral anterior temporal lobes or bilateral frontal lobes. High-frequency oscillations (HFOs) are considered to be promising biological markers for the EZ. However, a large number of studies on HFOs stem from intracranial research. There were few quantitative measures for scalp HFOs, so we proposed a new method to quantify and analyze scalp HFOs. This method was called the "scalp-HFO index" (HI) and calculated in both the EZ and non-EZ. The calculation was based on the numbers and spectral power of scalp HFOs automatically detected. We labeled the brain lobes involved in the EZ as regions of interest (ROIs). The HIs based on the ripple numbers (n-HI) and spectral power (s-HI) were significantly higher in the ROI than in the contra-ROI (P = 0.012, P = 0.003), indicating that HIs contributed to the lateralization of EZ. The sensitivity and specificity of n-HI for the localization of the EZ were 90% and 79.58%, respectively, suggesting that n-HI was valuable in localizing the EZ. HI may contribute to the implantation strategy of invasive electrodes. However, few scalp HFOs were recorded when the EZ was located in the medial cortex region.NEW & NOTEWORTHY We proposed the scalp-high-frequency oscillation (HFO) index (HI) as a quantitative assessment method for scalp HFOs to locate the epileptogenic zone (EZ). Our results showed that the HI in regions of interest (ROIs) was significantly higher than in contra-ROIs. Sensitivity and specificity of HI based on ripple rates (n-HI) for EZ localization were 90% and 79.58%, respectively. If the n-HI of the brain region was >1.35, it was more likely to be an epileptogenic region. Clinical application of HIs as an indicator may facilitate localization of the EZ.