Scalp EEG functional connection and brain network in infants with West syndrome☆
Introduction
West syndrome, known as Infantile spasms, is the most common age-dependent epileptic encephalopathy. West syndrome usually occurs in infancy and early childhood (Cui, Cao et al., 2022, Feng et al., 2022, Pavone et al., 2020). The typical clinical manifestation of West is a sudden contractions of the trunk and limbs, followed by a brief tonic contractions. The recurrent epilepsy can cause children to suffer from various neurological diseases, such as mental movement disorders, developmental delay or decline (Pavone, Striano, Falsaperla, Pavone, & Ruggieri, 2014), and may evolve into more serious epilepsy diseases, such as Lennox–Gastaut syndrome. The cause of West is complicated, the pathogenesis and evolution process are not clear so far, and accurate and effective treatment cannot be performed clinically (Janicot, Shao, & Stafstrom, 2020). Only few children have normal cognitive and motor development, but severe and frequent seizures will permanently damage the cognitive, learning, and language functional areas of the brain in most patients (Capone, Goyal, Ares, & Lannigan, 2006). Therefore, exploring the effects of infantile spastic seizures on functional brain areas is critical, that can promote clinical diagnosis and research on pathogenesis.
Scalp Electroencephalography (EEG) has been widely used in West syndrome analysis for its high practicability and low cost. Chu et al. (2021) proposed multi-scale entropy as an objective biomarker of abnormal EEG in infantile spasms, and analyzed whether it is related to treatment response and seizure outcome before and after treatment. Bernardo, Nariai, Hussain, Sankar, and Wu (2020) proved that the phase coupling of high-frequency oscillation and slow-wave activity can be used as a potential marker for West prognostic treatment. Similarly, McCrimmon et al. (2021) used the scalp high-frequency oscillations to accurately identify West subjects from healthy. Although, the characteristics of EEGs are very useful for clinical diagnosis of West as biomarkers (Cao et al., 2020, Hu, Cao, Lai, Liu et al., 2020, Hussein et al., 2021, Raghu et al., 2020, Wen et al., 2021, Xu et al., 2021), these objective patterns are usually not conducive to understand the underlying neural mechanism of West.
Graph theory has been recently applied in the brain imaging, electrophysiology and epilepsy analysis for its ability in quantifying the characteristics of complex system topologies (Bernhardt et al., 2015, Stacey et al., 2020). Guo et al. (2021) achieved the detection of high frequency oscillations for stereoelectroencephalography (SEEG) in epilepsy using the hypergraph learning. Li, Sohrabpour, Jiang, and He (2021) proposed a novel approach to model the temporal–spatial–spectral dynamics of cross-frequency coupling (CFC) networks, where graph measures were used to characterize the high-frequency and low-frequency hubs in treatment outcome of epilepsy. Lin et al. (2021) used the spatial pattern of network (SPN) features of resting-state scalp EEGs from the functional and effective EEG networks to achieve medically controlled epilepsy classification. Rosch, Baldeweg, Moeller, and Baier (2018) designed a dynamic tracking model, revealing the correlation between early epileptic encephalopathy and the dynamic characteristics of the brain network. In general, the quantitative analysis of brain network can help to understand the neurobiological mechanism for clinical diagnosis.
For West syndrome, only a few research on graph networks have been reported in the past. For instance, Hu, Mower, Shrey and Lopour (2020) used linear cross-correlation coefficients to construct functional connection network to analyze the influence of interictal epileptiform discharges in West syndrome. It is found that the connection strength of brain network in interictal epileptiform discharges is enhanced. Shrey et al. (2018) studied the changes in brain network of West syndrome, and concluded that changes in network connectivity and stability are related to the treatment response. Meanwhile, pre-treatment connectivity is very beneficial for the prediction of short-term treatment response. But the brain network functional connection of existing studies is generally evaluated using simple linear correlation coefficient. Moreover, existing studies mainly analyzed the prognostic control of West with a small number of patients. The changes in brain network connectivity patterns during the West seizure cycle are not well studied.
In this paper, we aim to analyze the scalp EEG functional connectivity and network topology in the prognosis of West seizure cycle, and also to explore the pathogenesis of West syndrome. The analysis is conducted on scalp EEGs of 19 channels, where the electrode leads are assigned as network nodes and the features on 5 sub-bands of EEGs (, , , , ) are extracted for analysis (Cao, Chen et al., 2021, Cao, Hu et al., 2021, Cui, Hu et al., 2022, Hu, Cao, Lai, Wang et al., 2020, Wang et al., 2021). The connection matrices and weighted networks based on correlation, coherence, Time Frequency Cross Mutual Information (TFCMI), Phase-Locking Value (PLV), Phase Lag Index (PLI), Weighted Phase Lag Index (WPLI) are constructed from the constituent components of EEGs (amplitude, frequency, phase angle). A two-tailed t-test is adopted to detect the intensity of brain area connectivity patterns in the West seizure cycle. To assess the efficiency of brain networks in processing global and local information during West seizure cycles, 4 network topological parameters, Clustering coefficient (C), Global efficiency (Ge), Local efficiency (Le), Feature path length (Pl), are derived. The stage differences in local and global efficiency of brain network in 5 sub-bands during the seizure cycle are observed. The study on 15 West syndrome children shows that:
- •
During the seizure cycle, there is asymmetry in connectivity between the left and right brain regions.
- •
Significant changes in the strength of connectivity between parietal and occipital regions, frontopolar and temporal lobe regions can be observed.
- •
Changes in brain area connectivity patterns and network parameters in band could be employed to determine West seizure cessation prognosis.
- •
Through correlation and WPLI obtained from multiband brain network connections, changes in the and frequency bands could be effectively applied for West seizure prediction.
Fig. 1 shows the overall framework for functional connection and brain network analysis on West syndrome in this paper. There mainly include 4 parts for the analysis: (A) Data collection for West syndrome subjects, (B) EEG preprocessing and artifact removal, (C) Representative correlation feature extraction, and (D) Functional connectivity analysis and result evaluation.
Section snippets
Subject identification
The purpose of this study is to establish different brain network patterns to analyze the changes in brain areas before and after a series of spasms. The data of West patients are selected from standard pediatric clinical EEGs, labeled by clinical neurophysiologists with expertise in pediatric EEG. Table 1 details the participant information. The average age of the patients is 11 months, the male to female ratio was 2:3, and all have at least one spastic seizure. The main basis for patient
Methodology
To overcome the inherent limitations of EEGs, such as low spatial resolution, the time, space, and frequency domain characteristics are fully exploited to construct the functional network based on 5 typical brain rhythms for West seizure cycle analysis. Different brain rhythm information will be first extracted. The connections used to construct the EEG functional network are partially based on the signal phase angles from Hilbert transform.
Results and discussions
The study aims to determine whether the network feature differences could be applied for judging cluster spasm seizure prognosis and cessation prognosis, and the analysis is carried out using clinical EEGs from 15 patients diagnosed with West syndrome. Here, 19 electrodes are selected as brain network nodes. To clearly show the connectivity between brain regions, we used a two-tailed t-detection method to count the association matrix and find out the functional brain connectivity patterns.
Conclusions
In this paper, the scalp EEG based functional connectivity and network topology parameters have been analyzed for the prognosis of childhood West’s seizure cycle. The medical statistics and anatomical principles have been combined to explore the brain mechanisms of West’s seizures, which are very valuable for clinical research implications. The study is carried out on the EEG data of 15 children suffered from West syndrome, and the analysis reveals that (1) there exists asymmetry property in
Ethical standards
This study has been approved by the Children’s Hospital, Zhejiang University School of Medicine and registered in Chinese Clinical Trail Registry (ChiCTR1900028804). All patients gave their informed consent prior to their inclusion in the study.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
References (47)
Oscillatory interactions between sensorimotor cortex and the periphery
Current Opinion in Neurobiology
(2007)- et al.
Interictal scalp fast ripple occurrence and high frequency oscillation slow wave coupling in epileptic spasms
Clinical Neurophysiology
(2020) - et al.
Network analysis for a network disorder: the emerging role of graph theory in the study of epilepsy
Epilepsy & Behavior
(2015) - et al.
Electroencephalography complexity in infantile spasms and its association with treatment response
Clinical Neurophysiology
(2021) - et al.
EEG-defined subtypes of children with attention-deficit/hyperactivity disorder
Clinical Neurophysiology
(2001) - et al.
Directed clustering in weighted networks: A new perspective
Chaos, Solitons & Fractals
(2018) - et al.
Deep feature fusion based childhood epilepsy syndrome classification from electroencephalogram
Neural Networks
(2022) - et al.
3D residual-attention-deep-network-based childhood epilepsy syndrome classification
Knowledge-Based Systems
(2022) - et al.
Effect of interictal epileptiform discharges on EEG-based functional connectivity networks
Clinical Neurophysiology
(2020) Gamma, fast, and ultrafast waves of the brain: their relationships with epilepsy and behavior
Epilepsy & Behavior
(2008)
Semi-dilated convolutional neural networks for epileptic seizure prediction
Neural Networks
Reorganization of functional connectivity during the motor task using EEG time–frequency cross mutual information analysis
Clinical Neurophysiology
Infantile spasms syndrome, west syndrome and related phenotypes: what we know in 2013
Brain and Development
EEG based multi-class seizure type classification using convolutional neural network and transfer learning
Neural Networks
Complex network measures of brain connectivity: uses and interpretations
Neuroimage
Strength and stability of EEG functional connectivity predict treatment response in infants with epileptic spasms
Clinical Neurophysiology
Emerging roles of network analysis for epilepsy
Epilepsy Research
An improved index of phase-synchronization for electrophysiological data in the presence of volume-conduction, noise and sample-size bias
Neuroimage
Slower alpha rhythm associates with poorer seizure control in epilepsy
Annals of Clinical and Translational Neurology
Brain network connectivity assessed using graph theory in frontotemporal dementia
Neurology
Unsupervised eye blink artifact detection from EEG with Gaussian mixture model
IEEE Journal of Biomedical and Health Informatics
Epileptic classification with deep transfer learning based feature fusion algorithm
IEEE Transactions on Cognitive and Developmental Systems
Epileptic signal classification with deep EEG features by stacked CNNs
IEEE Transactions on Cognitive and Developmental Systems
Cited by (16)
EEG-based functional connectivity analysis of brain abnormalities: A systematic review study
2024, Informatics in Medicine UnlockedDementia classification using a graph neural network on imaging of effective brain connectivity
2024, Computers in Biology and MedicineUsing chemosensory-induced EEG signals to identify patients with de novo Parkinson's disease
2024, Biomedical Signal Processing and ControlMeasuring multivariate phase synchronization with symbolization and permutation
2023, Neural NetworksSeizure detection algorithm based on improved functional brain network structure feature extraction
2023, Biomedical Signal Processing and ControlCitation Excerpt :Nevertheless, many children have other types of epilepsy later in their lives. WS presents epileptic spasms, and the patients’ EEG has peak rhythm disorder characteristics [82,83]. The EEG dataset used in this section is WS patient data from the Children’s Hospital, Zhejiang University School of Medicine.
Graph-Theory-Based Multilevel Cortical Functional Connectivity Developmental Analysis
2024, IEEE Transactions on Cognitive and Developmental Systems
- ☆
This work was supported by the National Key Research and Development Program of China (2021YFE0100100, 2021YFE0205400), the National Natural Science Foundation of China (U1909209, 62003119), the Key Research and Development Program of Zhejiang Province (2020C03038), the Natural Science Key Foundation of Zhejiang Province (LZ22F030002), the Open Research Projects of Zhejiang Lab (2021MC0AB04), and the Zhejiang Provincial Natural Science Foundation (LBY21H090002, LBY21H090001).