Scalp EEG functional connection and brain network in infants with West syndrome

Abstract

The common age-dependent West syndrome can be diagnosed accurately by electroencephalogram (EEG), but its pathogenesis and evolution remain unclear. Existing research mainly aims at the study of West seizure markers in time/frequency domain, while less literature uses a graph-theoretic approach to analyze changes among different brain regions. In this paper, the scalp EEG based functional connectivity (including Correlation, Coherence, Time Frequency Cross Mutual Information, Phase-Locking Value, Phase Lag Index, Weighted Phase Lag Index) and network topology parameters (including Clustering coefficient, Feature path length, Global efficiency, and Local efficiency) are comprehensively studied for the prognostic analysis of the West episode cycle. The scalp EEGs of 15 children with clinically diagnosed string spasticity seizures are used for prospective study, where the signal is divided into pre-seizure, seizure, and post-seizure states in 5 typical brain wave rhythm frequency bands ($\delta$ (1–4 Hz), $\theta$ (4–8 Hz), $\alpha$ (8–13 Hz), $\beta$ (13–30 Hz), and $\gamma$ (30–80 Hz)) for functional connectivity analysis. The study shows that recurrent West seizures weaken connections between brain regions responsible for cognition and intelligence, while brain regions responsible for information synergy and visual reception have greater variability in connectivity during seizures. It is observed that the changes in $\beta$ and $\gamma$ frequency bands of the multiband brain network connectivity patterns calculated by Corr and WPLI can be preliminarily used as judgment of seizure cycle changes in 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 ($\delta$, $\theta$, $\alpha$, $\beta$, $\gamma$) 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 $\delta$ band could be employed to determine West seizure cessation prognosis.

• Through correlation and WPLI obtained from multiband brain network connections, changes in the $\beta$ and $\gamma$ 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.