Brain network dynamics codify heterogeneity in seizure evolution

Abstract

Dynamic functional brain connectivity facilitates adaptive cognition and behaviour. Abnormal alterations within such connectivity could result in disrupted functions observed across various neurological conditions. As one of the most common neurological disorders, epilepsy is defined by the seemingly random occurrence of spontaneous seizures. A central but unresolved question concerns the mechanisms by which extraordinarily diverse propagation dynamics of seizures emerge. Here, we applied a graph-theoretical approach to assess dynamic reconfigurations in the functional brain connectivity before, during and after seizures that display heterogeneous propagation patterns despite sharing similar cortical onsets. We computed time-varying functional brain connectivity networks from human intracranial recordings of 67 seizures (across 14 patients) that had a focal origin-49 of these focal seizures remained focal and 18 underwent a bilateral spread (focal to bilateral tonic-clonic seizures). We utilized functional connectivity networks estimated from interictal periods across patients as control. Our results characterize network features that quantify the underlying functional dynamics associated with the observed heterogeneity of seizure propagation across these two types of focal seizures. Decoding these network features demonstrate that bilateral propagation of seizure activity is an outcome of the imbalance of global integration and segregation in the brain prior to seizure onset. We show that there exist intrinsic network signatures preceding seizure onset that are associated with the extent to which an impending seizure will propagate throughout the brain (i.e. staying within one hemisphere versus spreading transcallosally). Additionally, these features characterize an increase in segregation and a decrease in excitability within the brain network (i.e. high modularity and low spectral radius). Importantly, seizure-type-specific differences in these features emerge several minutes prior to seizure onset, suggesting the potential utility of such measures in intervention strategies. Finally, our results reveal network characteristics after the onset that are unique to the propagation mechanisms of two most common focal seizure subtypes, indicative of distinct reconfiguration processes that may assist termination of each seizure type. Together, our findings provide insights into the relationship between the temporal evolution of seizure activity and the underlying functional connectivity dynamics. These results offer exciting avenues where graph-theoretical measures could potentially guide personalized clinical interventions for epilepsy and other neurological disorders in which extensive heterogeneity is observed across subtypes as well as across and within individual patients.

Keywords: brain networks; epilepsy; functional connectivity; seizure propagation.

Graphical abstract

Figure 1

Emergence of distinct seizure propagation patterns in a single patient. (A) During a clinical monitoring procedure to identify a seizure onset zone of patients with medication-refractory (drug-resistant) epilepsy, intracranial recording electrodes are implanted. (B) Intracranial activity during two sample seizures recorded from a single patient, which exhibit distinct propagation dynamics. On the left, the seizure activity originates from a few electrodes and persists in the localized area within a single hemisphere (i.e. focal seizure that remains focal). On the right, the seizure activity originates from a few electrodes but diffuses bilaterally to involve electrodes in both hemispheres. This type of seizure is known as focal to bilateral tonic-clonic seizure or focal seizure with bilateral spread. Despite their similarly focal origin, these seizure types induce drastically differential clinical manifestations such that focal to bilateral tonic-clonic seizures are associated with more severe cognitive and behavioural deficits. We hypothesize that such heterogeneity in seizure dynamics emerges from distinct and measurable temporal alterations in the functional brain connectivity networks.

Figure 2

Schematic of graph-theoretical analysis of functional brain dynamics. (A) Locations of implanted intracranial electrodes of a sample patient. (B) We use electrocorticography (ECoG) time-series data from all intracranial electrodes from each patient recorded during a clinical monitoring procedure to locate the seizure onset zone. We estimate the instantaneous functional connectivity of the underlying brain network by computing pairwise correlations of ECoG data across electrodes in a sliding-window manner. The magnitudes of these correlations (restricted between 0 and 1) reflect the strength of connections between each pair of electrodes and are represented by a weighted adjacency or connectivity matrix (see Materials and methods). (C) To investigate time-varying changes in the functional brain connectivity during temporal evolution of each seizure type, we compute a series of connectivity matrices over time and use these as bases to construct functional connectivity networks. (D) A schematic of sample constructed networks, consisting of nodes (electrodes) and edges (connection strength). To quantify alterations within these complex networks over time, we evaluate changes of a series of graph-theoretical attributes which describe globally and locally defined properties of the constructed networks.

Figure 3

Clustering coefficient and characteristic path length track diffusivity of seizure activity. Focal to bilateral tonic-clonic seizures (n = 18) display simultaneous increase in the clustering coefficient (CC) and decrease in the characteristic path length (PL) than focal seizures that remain localized within one hemisphere (n = 49). (A) Averages of CC associated with each seizure type are plotted separately for preictal, ictal (during seizure) and postictal periods. CC of interictal (seizure-free) networks is also plotted as a baseline. (B) PL is plotted in the same manner. (C) CC of focal to bilateral tonic-clonic seizures is higher than that of focal seizures that remain focal, 2.25–10 min after seizure onset. (D) PL of focal to bilateral tonic-clonic seizures is lower than that of focal seizures that remain focal, 2–10 min after seizure onset. Statistical comparisons of network measures as a function of seizure types were computed through a bootstrapping procedure where the underlying data distribution of each network measure was resampled at the level of individual seizures to established 95% confidence intervals (CIs). For (A) and (B), significance of changes in the averages of CC and PL across time windows were evaluated by two-way analysis of variances (repeated measures) with a series of post hoc t-tests (two-tailed) for significant main effects. For (C) and (D), error bars indicate 95% CIs across individual seizures in each condition and solid bars show resampled P < 0.05.

Figure 4

Various features of functional connectivity networks display distinct temporal changes as a function of seizure propagation dynamics. Left panels illustrate a series of graph-theoretical measures computed from networks of focal seizures that remain localized (n = 49) and from networks of focal to bilateral tonic-clonic seizures (n = 18). The time-varying differences observed in each of these features as a function of seizure types are plotted in the corresponding right panels. (A) The density of focal to bilateral tonic-clonic seizures is higher than that of focal seizures that remain focal, 1.75–10 min after seizure onset. (B) The assortativity, a measure of network robustness, is lower for focal to bilateral tonic-clonic seizures relative to focal seizures that remain focal, 7.5–9.25 min after seizure onset. (C) The modularity, which captures efficient network integration and global segregation, is higher for focal to bilateral tonic-clonic seizures when compared with focal seizures that remain focal during temporal windows between 14.75–5.75 min before seizure onset and 0.75–1.50 min after the onset. (D) The spectral radius, which relates to the global spread of synchronization in a network, is also higher for focal to bilateral tonic-clonic seizures when compared with focal seizures that remain focal during temporal windows between 14.75–3.75 min before seizure onset and 0.75–1.50 min after the onset. (E) The synchronizability, which estimates the propensity of information to diffuse in a network, shows an increasing trend post-seizure onset for unconstrained seizure dynamics. However, no statistically significant differences were observed in synchronizability across seizure types. Statistical comparisons of network measures as a function of seizure types were computed through a bootstrapping procedure where the underlying data distribution of each network measure was resampled at the level of individual seizures to established 95% confidence intervals (CIs). Error bars indicate 95% CIs across individual seizures in each condition and solid bars show resampled P < 0.05.

Figure 5

Summary of graph-theoretical attributes probed across seizure types. The network features investigated can be categorized into two groups based on the temporal windows at which differential changes in these features emerge as a function of seizure propagation patterns. The time windows where such differences are observed are plotted separately for each of the network measures (resampled P < 0.05). Global features, i.e. the modularity and spectral radius, primarily capture network alterations that occur prior to and shortly after seizure onset. In contrast, the density, assortativity, clustering coefficient and characteristic path length characterize post-onset network reconfigurations induced by different types of propagation dynamics.

Figure 6

Distinct patterns of network properties across seizure types at a single-seizure level. Modularity A and spectral radius B extracted from networks associated with three seizures that share similar onset regions recorded from a sample patient. Seizures 1 and 3 are categorized by an epileptologist as focal seizures that remain focal (sample recordings of Seizure 1 is illustrated in Fig. 1B, left), whereas Seizure 2 is categorized as a focal to bilateral tonic-clonic seizure (sample recordings of Seizure 2 is also illustrated in Fig. 1B, right). Seizures 1 and 3 exhibit similar patterns of modularity and spectral radius overtime, and these temporal dynamics differ from those of Seizure 2. The grey-shaded region highlights the robust signatures associated with the bilateral spread of a focal seizure (i.e. increase in modularity and decrease in spectral radius) and is followed by network changes which are likely seizure and/or patient specific.