Abstracts

Rationale: Two seemingly contradictory theories describe the source of ictal traveling waves. The first posits that the source arises from a slowly advancing ictal wavefront characterized by collapsing inhibitory restraint (Smith et al., Nature Communications, 2016). The second is that the source of traveling waves arises from a fixed location or network (Martinet et al., Nature Communications, 2017). Because these theories were developed using independent microelectrode array recordings and methods from different institutions, the different theories may reflect patient variability or method sensitivity.

Methods: In this work, we combine the methods and microelectrode array recordings (n=11 patients, n=31 seizures) from the previous studies to analyze the source of ictal traveling waves and understand the factors that led to the different observations. To do so, we validate both existing methods on an in silico dataset before applying each method to the full cohort of seizures. We then introduce a third novel method that does not assume linearity of the wave propagation patterns, as in the previous methods. We illustrate how our conclusions can arise in vivo using a mean-field computational model that has previously been used to describe seizure dynamics and other neurological states.

Results: We show that both of the previously established methods to characterize ictal traveling waves perform similarly, and that seizures show evidence of both theoretical paradigms. We detect an ictal wavefront in 17/31 seizures. When present, application of a novel third method shows the pattern of ictal wavefront propagation is correlated with that of traveling waves immediately before ictal wavefront passage, and anticorrelated after ictal wavefront passage. However, both correlations appear only transiently (n=17 seizures, t in [-10, 20] s surrounding ictal wavefront passage). Moreover, we identify multiple shifts in traveling wave directions, suggesting that neither a fixed network nor an ictal wavefront acts alone to produce traveling waves (Figure 1). Using a computational model, we illustrate how an ictal wavefront interacts with a fixed source, and show that this model reproduces features of the traveling wave dynamics observed in vivo.

Conclusions: We show that the existing controversy around the source of ictal traveling waves does not result from differences in methods or patient populations. Instead, we find that ictal traveling waves are generated by interactions between both types of sources — an ictal wavefront and fixed network — proposed in previous theories.

Figure 1: The direction of fast traveling waves during human seizures. (A) Example fast traveling wave directions from three patients, computed using two methods. The ictal wavefront passes over the microelectrode array at direction 0 degrees at time 0 s. (B) Angular distribution of traveling wave directions before (left) and after (right) ictal wavefront passage. (C) Mean traveling wave direction before and after ictal wavefront passage for each patient and seizure.

Funding: Please list any funding that was received in support of this abstract.: National Institute of Neurological Disorders and Stroke (R01-NS110669).